[ 3 / biz / cgl / ck / diy / fa / g / ic / jp / lit / sci / tg / vr ] [ index / top / reports / report a bug ] [ 4plebs / archived.moe / rbt ]

2017/01/28: An issue regarding the front page of /jp/ has been fixed. Also, thanks to all who contacted us about sponsorship.

/diy/ - Do-It-Yourself

View post   

[ Toggle deleted replies ]
File: 121 KB, 1280x1280, 1555095905606.png [View same] [iqdb] [saucenao] [google] [report]
1615963 No.1615963 [Reply] [Original] [archived.moe]

benched thread: >>1610590

0. Electrics ≠ electronics. Appliances/mains stuff to /qtddtot/ or /sqt/. PC assembly to >>>/g/.
1. Do your own homework. Search web first. Re-read all documentation/datasheets related to your components/circuits. THEN ask.
2. Pics > 1000 words. Post relevant schematic/picture/sketch/9001.5 hours in MS Paint with all part numbers/values/etc. when asking for help. Focus/lighting counts.
3. Read posts fully. Solve more problems than you create.
4. /ohm/ is an anonymous, non-smoking general.

>I'm new to electronics, where to get started?
It is an art/science of applying principles to requirements. Find problem, learn principles, design and verify solution, build, test, post results, repeat

>Project ideas:

>Principles (by increasing skill level):
Mims III, Getting Started in Electronics
Platt, Make: Electronics
Geier, How to Diagnose & Fix Everything Electronic
Kybett & Boysen, All New Electronics Self-Teaching Guide
Scherz & Monk, Practical Electronics for Inventors
Horowitz and Hill, The Art of Electronics

>Design/verification tools:
NI Multisim
iCircuit for Macs
KiCAD (pcb layout software, v5+ recommended)

Mouser, Digi-Key, Arrow, Newark, LCSC (global)
RS Components (Europe)
eBay/AliExpress sellers, especially good for component assortments/sample kits (caveat emptor)
Local independent electronics distributors

>Related YouTube channels:

>Li+/LiPo batteries
Read this exemplary resource first: https://www.robotshop.com/media/files/pdf/hyperion-g5-50c-3s-1100mah-lipo-battery-User-Guide.pdf
>I have junk, what do?
Take it to the recycler.

>> No.1615993


fugg how did I not consider that thanks

if the voltage applied at the JP connector goes below 3.2V the LED turns off
it works on the bread board but I wasn't sure about the base current

good thank you

>> No.1616012

I know basic circuit theory but HF scares me. If I make a PCB and the signal for an external antenna has to propagate through, say, an inch of traces on the board before reaching the antenna connector, will that inch increase the effective length of the antenna and throw off the resonant frequency?

>> No.1616019
File: 55 KB, 818x522, 1546451976438.png [View same] [iqdb] [saucenao] [google] [report]

the antenna doesn't actually start until the ground and RF conductors diverge. if you implement that trace as a transmission line and match the impedance of that trace to the nominal system impedance, the trace will simply transmit all power into the antenna and not radiate anything. Pic related, transmission line calculator included with KiCAD, having calculated the correct width of the trace to create a 75 ohm microstrip transmission line on a typical substrate with 1 oz. copper (but you'll probably want to use a 50 ohm system impedance by default). there are many other such calculators. JLCPCB has one with presets for their board substrates
you should take a little detour into transmission line theory next

>> No.1616020

Kudos for posting a real circuit diagram!
>below 3.2V the LED turns off
Ok, thanks.
I do not recognise the pinout of IC1, what type is it?
I cannot see any hysteresis, are the (+) and (-) inputs correct?
Can you explain the expected function of Q1?

>> No.1616031

>you should take a little detour into transmission line theory next
Do you know of a good introduction to this? I don't plan to get deep into HF, I just want to add things like LoRa to my projects

>> No.1616036

>the antenna doesn't actually start until the ground and RF conductors diverge

wat? How are they connected together? PCBs are a flat plane. You can't make twisted pair traces. Is it just two adjacent traces or something that are capacetively coupled? Do you use vias to alternate the sides of the board or something?

>> No.1616048
File: 96 KB, 850x510, 1550371054874.png [View same] [iqdb] [saucenao] [google] [report]

not that poster, but consider it a comparator with open-drain output. their output sense tends to be inverted compared to that expected from op-amps

the hams are usually the best source for elementary radio knowledge
if you don't want to get deep in the weeds, then it might be enough to look up the datasheet and all the app notes of your LoRa module/IC and follow their guidelines and recommendations

>You can't make twisted pair traces
you don't need to. you just need to keep the transmitted power from flying out of the conductor
>Is it just two adjacent traces or something that are capacetively coupled
also inductively, but yes
>Do you use vias
no, in fact you should avoid them on the signal conductor because of the impedance discontinuity they introduce into the line, where the transmitted power can be radiated from the line and thereby lost

>> No.1616066

im interessed in building my own portable digital music player. Everything i found is heavly based on raspberry pi and ready to use DAC. Closest thing i found is Pipod. Is there any more dedicated way to operate audio files? Im new into electronics and im trying to find resources about this

>> No.1616072
File: 7 KB, 264x191, 741.png [View same] [iqdb] [saucenao] [google] [report]

>output sense tends to be inverted
I know waffling when I see it.
Open drain has no base drive for Q2.
It's more likely a 741.

>> No.1616078

does that make sense in the context of the circuit, though? OP claims he's using a comparator and that the circuit works well excluding the output transistor. I think he just grabbed the first symbol in EAGLE that looked like a triangle with + and - ports on it
>Open drain has no base drive for Q2.
there's R6+R9, just barely enough base drive to get a few mA through the LED. OP should add a ~10k pullup to the output so that the hysteresis resistor and the transistor are more strongly driven when the output isn't sinking

>> No.1616087

Ok, let's leave it at that.

>> No.1616093
File: 1.21 MB, 3200x1800, IMG_20190523.jpg [View same] [iqdb] [saucenao] [google] [report]

/ohm/ & /3dpg/ in harmony

>> No.1616094

oh IC1 is just a normal op amp comparator
I'm using a siemens tba222
There's a hysteresis, I can't explain it well but I got the circuit idea from here:
Q1 is a mosfet to prevent short circuits when the battery is connected in reverse

>> No.1616095 [DELETED] 

>~10k pullup to the output
good idea, i'll try that

>> No.1616106

>OP should add a ~10k pullup to the output
that didn't raise the current by much according to falstad
I tried lowering the LED resistor to 300, that turned up the base current to 50 uA which seems to work well

>> No.1616113

opamp ≠ comparator. if it's a TBA222 then it's an opamp and the hysteresis feedback should go to the + terminal. nomenclature is important

>> No.1616120

Does it make that much of a difference though? In the datasheet it says under applications "opamp and comparator"

>> No.1616136

if the output pin isn't at Vcc or Vee 99% of the time, then it should be an op-amp.

>> No.1616157
File: 5 KB, 754x42, 1555543545414.png [View same] [iqdb] [saucenao] [google] [report]

but do you know why 50uA?
likely: the dropper resistor for the LED is still in the way of the emitter, keeping your transistor from behaving like a switch. what you had created, more or less, is a switched current source that turns on just enough to keep the emitter one diode drop below the base. Ie = (Vb - 0.65)/Re and Ic = Ie-Ib so Ic = Ie*(1-(1/beta))
somewhat likely: the opamp is running well below its design voltage and can't push its transistors hard enough

often, you can use an op-amp in place of a comparator, but the opamp will generally be slower to respond and the outputs may not swing as tightly to the rails as you might like
anyway, the TBA222 is a terrible op-amp for this application. its input resistance is on the same order as your voltage dividers, it's only guaranteed to work with input voltages 3V or more inside the power rails, and it isn't even rated to run on any less than an 8V supply. you're very lucky that it works and is even close to accurate

>> No.1616159

trying to figure out how/what the fuck current and voltage actually are

I can't find any satisfactory analogy that makes sense, so I'm falling back to math

is there a version of ohm's law that treats voltage as a function of current, or vice versa - current as a function of voltage?

>> No.1616164
File: 144 KB, 550x799, tba221-2.png [View same] [iqdb] [saucenao] [google] [report]

>siemens tba222
It's a 30 year old mil version of the 741. The minimum (single) supply voltage is 8V and the corresponding output voltage is only a few volts. 5V supply is not a good idea. The data sheet is at http://www.ceipsa.com/datasheet/TBA222.pdf

>> No.1616177

>I can't find any satisfactory analogy that makes sense, so I'm falling back to math
Afaik the only analogy that is successfully taught in school is the hydraulic analogy, i.e. voltage is pressure and current is flow, and it works for many simple cases. The analogy falls apart with magnetism but for the purpose of getting students to understand it's pretty good.

If you really want to get it, study these: http://www.feynmanlectures.caltech.edu/II_toc.html

>is there a version of ohm's law that treats voltage as a function of current, or vice versa - current as a function of voltage?
Rearrange the terms of V=IR. Or were you thinking of something else.

>> No.1616196

>Rearrange the terms of V=IR. Or were you thinking of something else.

I wanted to find the integrals and derivatives of voltage as a function of current, but I'm not realizing that as ohm's law is linear, it won't be very informative

that, and practically every circuit has a different I/V curve so just doing the integral/derivative of ohm's law is pretty much patently useless

>> No.1616200

>not realizing
now realizing

>> No.1616210

these feynman lectures are fantastic by the way, thank you

>> No.1616217

>it won't be very informative
not for a resistor. inductors and capacitors? absolutely.

>> No.1616231

Build actual circuits and stop fucking around with the math bullshit. You're muddying the waters more that way and you'll have a harder time untangling it later.

Build shit. Measure voltages, current, probe waveforms etc. Adjust different parameters of your circuit and see how it affects your output. You will gradually build intuition and once you have a base then you dive into the math behind it.

>> No.1616248

>I wanted to find the integrals and derivatives of voltage as a function of current
Take MIT 6002x, free online course. You'll learn the appropriate functions for inductors and capacitors, and you'll learn how those devices behave with a step input and AC input. It's a good foundation for understanding electronics.
I suggest taking this course in parallel with studying circuit simulations and real life dabbling with circuits like the other anon suggests.

>> No.1616257

How about trying to derive ohm's and kirchoff's laws from maxwell's equations? Plus the operation of an inductor and capacitor. That's about as fundamental as you can get. You can also derive EM waves if you so desire.

>> No.1616320

Here's one:
Some guy swallowed an entire smartphone, and the doctors were worried that the battery wouldn't be terribly good for him. You know what surgery they gave him? Liposuction!

>> No.1616321


>> No.1616322

Yeah I figured that the base current was dependent on the emitter current, thanks.
I just double checked and it's actually a tba221. Where does it say in the datasheet that it can only run ov atleast 8V?
But for my final circuit I'll probably use something else

>> No.1616325
File: 12 KB, 588x70, 1550161442154.png [View same] [iqdb] [saucenao] [google] [report]

yeah, an MCP6001 would be great on a 5V supply
Pic related. ±4 split = 8 single

>> No.1616326

Deriving the basic circuit laws from Maxwell's is a very useful exercise for sure, especially when you're working with anything prone to interference.

>> No.1616328

I've actually never done so, though I have done the wave, capacitor, and inductor equations. I'm assuming they're not terribly difficult though, I might give them a go.

>> No.1616336

I wish there was a datasheet search browser extension, so you could just select a part number, right-click > search datasheet > [list of datasheet sites]. Can't be too hard, can it? There's a reverse image search extension that basically does this already but with different sites and to images instead of text.

>> No.1616345
File: 8 KB, 432x183, NX2301P.png [View same] [iqdb] [saucenao] [google] [report]

The most fatal error is Q1. You used the wrong symbol (n-channel instead of p-channel) and you forgot the drain-source diode which will do exactly what you wanted to prevent in case of reversed battery polarity: it pulls the input down to -Vbat and kills the amp/comp (which also has the wrong input polarity if you want a hysteresis action). I'd suggest to scrap the entire circuit and start from scratch. It's perfectly normal, that's how you learn.

>> No.1616346

>±4 split = 8 single
oh I didn't know it works like that, thanks

>> No.1616349

that's right, I just downloaded the symbol from somewhere because I was too lazy to make one myself
I'll change those

>> No.1616376

> practically every circuit has a different I/V curve so just doing the integral/derivative of ohm's law is pretty much patently useless
Ohm's law describes the current in a conductor, where voltage and current are proportional, and their ratio is a constant termed "resistance". It doesn't apply to anything else.

There are mathematical models for semicondutors (Shockley diode equation, Ebers–Moll model for a BJT), but they aren't used that much. Practical circuits tend to be designed in such a way that simplified models can be used.

>> No.1616395
File: 320 KB, 1200x1200, 26805[1].jpg [View same] [iqdb] [saucenao] [google] [report]

Lawnmover guy here, so i found the exact switch used and it seems the white thing is nothing more than a 0.1uf cap.
Will that alone be enough to work as a snubber, or are there some extra parts inside of that switch?
Basically i want to replace that switch with my own ordinary switch, and need to know if connecting that white cap in parallel with the new switch is enough to retain the snubber functionality.

>> No.1616422

You don't need a physical diode. All MOSFETs have an internal body diode that exists due to the nature of their construction. All spice models for MOSFETs will model the body diode even if the symbol omits it. You don't need to add any external diodes.

>> No.1616478

Can you not see that the body diode _is_ the problem?

>> No.1616532

I'm trying to repair an old 80s 1Ghz service generator / AM/FM modulator signal generator device model SSI-2000. It's made by "Spectrum Specialties Inc", which was a tiny california company bought out by Wavetek a million years ago.

I've started digging into it (it was designed by a madman, but pretty neat inside, all the digital control is 74xx TTL) It'd be a lot faster if I could find a service manual for it. Anyone have leads besides:

Furious Googling

>> No.1616543

elektrotanya comes to mind

>> No.1616579

Nice thanks anon, didn't know about that site. No hits but lots of good stuff there.

>> No.1616626
File: 497 KB, 500x350, beacon.gif [View same] [iqdb] [saucenao] [google] [report]

I recently built a circuit for a model lighthouse that has a twilight switch and a flasher for a 2 W light bulb. It is powered solely by small AC adapter so power consumption isn't an issue.

I'd like to build a miniature flashing beacon (something like pic related) with a super bright 5 mm white LED, solar panel and some kind of power storage.

I already have a few small solar panels that are rated at 12 V/20 mA. I was thinking to use supercapacitors as power storage for this, but how should I estimate required capacity?

At the moment it seems I would need:
- step-down converter (panel 4..12V --> 2.5V) for charging the capacitor
- step-up converter (2.5 V --> 3.5 V) for driving the LED
- low frequency oscillator/flasher
- comparator to turn the flasher on when the panel voltage drops near zero.

All of these would need to consume as little quiescent current as possible. Any suggestions for ICs/capacitors suitable for this kind of device?

>> No.1616629

Honestly it would probably be way cheaper / faster to just get a quality solar pathway light and replace the LED with a flashing one.

>> No.1616630

Sorry, forgot to mention the current candidates:

- step-down converter (panel 4..12V --> 2.5V) for charging the capacitor

- step-up converter (2.5 V --> 3.5 V) for driving the LED

- low frequency oscillator/flasher
>HEF4093 NAND oscillator, supply current: 0.25 μA

- comparator to turn the flasher on when the panel voltage drops near zero.
>MAX921, supply current: 4 μA

>> No.1616638

Too much converting. I would use lower voltage PV cells and explore cheap garden light circuits for the flashing.

>> No.1616655

yep. ground is just a chosen origin point

just the cap

a plain Schottky would have been plenty

the 12V solar cell is definitely too much for this application. a lower-voltage, lower-current cell would fit better, especially if you NiMH battery instead of supercaps, which, outside of a memory protection or other micropower application, are only good for minutes. easy to charge, easy to care for, can be set to your choice of voltage

>> No.1616702
File: 19 KB, 500x327, wiring-outlet-to-outlet-diagram.png [View same] [iqdb] [saucenao] [google] [report]

Pic related.
So I'm redoing some outlets and 12/2 wires in basement.
Using wall mounted outlet box. Box doesn't have enough room to fit the wire nuts
Do I have to wire up the grounds using wire nuts like shown or can I just use the existing ground screw and get both of the wires in it?
Or do I have to use wire nuts and now I have to go to the store like I was wanting to avoid doing?

>> No.1616703

/ohm/ RULE 0
sparky go home

>> No.1616714

My bad, just saw electrics general

>> No.1616716

>plain Schottky
yes, any diode would do.

>> No.1616729

>2-wire cable
looks like 3 wires to me

>> No.1616732

Is there a good place online to learn PCB design? I'm willing to spend a few dollars if Udemy or something similar is good.

Basically I joined a team project, and they need someone who knows PCB design, which nobody does, but I'm an EEng student and willing to learn so I volunteered.

>> No.1616752

bench supplies strive to minimize their output capacitance to improve current limiting. without a big output capacitor, how else do they maintain voltage regulation during large load steps? i can't seem to figure it out other than "they don't" and maybe that parasitic Ls help more than i think.

do you want to learn the fundamental design considerations or how to use a specific tool? i really couldn't find anything good when i learned it about 2 years ago. i honestly think your best bet is just to pirate altium and wing it for this one project. pcb software typically has a tedious workflow of design components -> create schematic -> route pcb with schematic aids, but you can bypass that and just manually create traces and component footprints as long as you're careful to get things right. when you're done use oshpark's website to confirm that your output files (gerber files) are correct. add as many debug/bodge supporting features as you can possibly fit. triple check everything. i nearly never make a first rev pcb without issues, including a single day turnaround that really needed to be right. if you have the space to solder on extra shit and cut/reveal traces with an exacto then you won't need to worry.

>> No.1616754

>place online to learn PCB design?

never heard of any such place. like oral sex, you learn by doing.

>> No.1616774

Do you enjoy being stupid?

>> No.1616777

No I'm just ignorant to how houses are wired up. But I see a ground, neutral, and live wire going into and out of the "2-wire cable".

>> No.1616779
File: 114 KB, 730x678, two conductors - with ground.png [View same] [iqdb] [saucenao] [google] [report]

>I see a ground, neutral, and live wire going into and out of the "2-wire cable".

>> No.1616782

If that's the nomenclature than sure I guess, but you can't show me a cable with three distinct copper channels and tell me "that's two conductors, the ground isn't a conductor".

>> No.1616788

industry nomenclature is always nonsense. it could just as easily be 000-wire cable.

>> No.1616808

Fundamental design I suppose, as I don't even know where to start. I'll do what you said, thanks for the advice.

I'll get right to it then.

>> No.1616811

what's your project? you might be able to just shove an arduino and some breakouts in a box

>> No.1616812

Get this book: "the circuit designer's companion." It's practical, not bogged down with theoretical math.

>> No.1616816

For solar panels, you ideally want a maximum power point tracking (MPPT) converter. This varies the current drawn from the panel to maximise power (the optimal current draw increases with light intensity).

You should avoid the step-up converter; use higher-voltage capacitors, capacitors in series, or rechargeable cells (preferably lithium, e.g. 18650, as those have enough voltage to drive the LED from a single cell). For efficiency, you want to drive the LED from a buck (step-down) converter, as they naturally act as constant-current sources (without an output capacitor), so you don't need to waste power on a passive current limiter.

>> No.1616832

Power distribution system for some sort of autonomous vehicle that will be sent to some space station (or attached to something sent to space?) in a few years. It's a volunteer based project but with actual stakes, hence why we have people who don't know what the fuck they're doing, coming together to do complicated shit.

Thank you.

>> No.1616846
File: 179 KB, 1028x708, RS-232.png [View same] [iqdb] [saucenao] [google] [report]

My scope has an RS-232 serial port on the back of it. Looking at the RS-232 standard, a low signal is 3-15V below the 0V rail, and a high signal is 3-15V above the 0V rail. If I want to interpret this with a micro, I'll need to convert this into a standard TTL level. I'm pretty sure a standard logic level shifter won't work since it requires a common 0V for a logical low. Do I just use optocouplers? It will only be 4 or so.

>> No.1616849
File: 173 KB, 1040x558, MC145406.png [View same] [iqdb] [saucenao] [google] [report]

Here's the other side, it's being driven by an MC145406 off ±12V rails. Yes this is a CRT oscilloscope with 24 dense schematic pages in its service manual.

>> No.1616850
File: 209 KB, 1062x1375, 1550980501493.png [View same] [iqdb] [saucenao] [google] [report]


>> No.1616856

MAX232 chip is the standard way everybody does it, although if you're only reading incoming datayou can go from the -12v /+12v to 0v/5v with only three resistors . Also the RTS/CTS wires were so infrequently used that a lot of USB-> RS232 converters don't even bother with them.

>> No.1616858

I expect that since all 4 of those wires are shown to be wired up to one of the main microcontrollers, I'll need to both transmit and receive. So I'll see about ordering some of those guys I guess. 70c for 5 isn't exactly a steep investment. I wonder if I'll be able to cram the whole MCU circuit inside the D-25 plug moulding?

>> No.1616866

>I wanted to find the integrals and derivatives of voltage as a function of current
You're looking for the equations that describe the voltage-current-inductance/capacitance relationships; v=Li'(t) and i=Cv'(t).

>> No.1616868

Assuming your circuit is relatively simple (like you don't have to worry about stray inductance, capacitance, or trace length) it's pretty simple.
>get PCB CAD software
>design schematic
>simulate if you're a pussy
>figure out the constraints and configuration of your board manufacturer (they will usually have a list of settings)
>have your software switch to its routing mode
>tell it your constraints (how big/small traces should/could be, etc.)
>draw lines to where the software says it should go, or let the software do it for you if you like spaghetti

>> No.1616890

You can *use* an op-amp as a comparator, you just have to add some positive feedback so it has hysteresis, and otherwise think of it as a comparator. Of course it won't have an open collector (or open drain, as the case may be) output like an actual comparator, so you can't wired-OR the outputs of several together, but a switching diode will take care of that.

>> No.1616906

Comparators don't have to have hysteresis, though it is a good idea. Actually I wonder if there are comparators with schmitt triggers in them that you feed a current or voltage into an extra pin in order to adjust their threshold? Wouldn't be too much of a stretch to have two in a 16-pin IC.

>> No.1616921
File: 52 KB, 825x665, Capture.png [View same] [iqdb] [saucenao] [google] [report]

Anyone have a recommendation for intermediate/advanced book on circuit design, with more practical circuits? I'm thinking something like Sedra-level math breaking down
>pic related
particularly I'm at the point where I've designed a similar circuit, but the particular effects of things like the 56nF feedback cap and output cap(ie why this value) and its relation to the BW/slew of the op-amps, the 'large-to-small signal' effects of the diode, etc. I frequently get caught up in these more-nuanced aspects and it always ends up just wasting time or I'm just throwing in random caps at places

>> No.1616926


worth mentioning the op-amps diode clamp input but also I want to say it wouldn't matter for that particular ckt

>> No.1616942


if you're thinking a circuit is designed the same way a formula is derived, you'd be wrong. it usually starts with a broad idea like ''use a stable voltage reference and feed it into an op-amp that uses a high-power NPN to boost power. then use a small resistor in series with the load to feed another op-amp that shuts down the first op-amp if current gets too high."

then you breadboard it, make changes until it works, then iterate to fix any weirdness. like the 5uF cap which was probably added to soft-start something or delay something.

the weird voltage reference was probably chosen coz they wanted to use something from the same company that made this design.

the 500uF cap could be anything from 22uF to 2000uF. it's not computed, just chosen from whatever you have at hand.

>> No.1616958

Yeah, I get that. In fact, it was when I realized how over-blown the analysis for a fucking two transistor amp in class was that I got disinterested and started reading linear tech app notes/art of elec. But I feel now that I've come full circle and want to be able to at least predict things like feedback stability, effects of impedance. e.g I know this ckt would fail/oscillate if there wasn't much resistance on the Darlington, and that a cap across CB might help, but this is all 'practical' knowledge and I can't apply this to more complex examples.

I'm interested in working in analog design so I know there's got to be a semblance of some small-signal/control theory when you're designing an IC, right?

also I didn't think bout just using whatever value on hand, good point

>> No.1616960

>cap across CB might help
actually I think would only prevent parasitic oscillation but the fact that I'm not sure is my point

>> No.1616963

>I wonder if there are comparators with schmitt triggers in them that you feed a current or voltage into an extra pin in order to adjust their threshold?
It could be done with 3 comparators and few resistors.

>> No.1616967
File: 13 KB, 800x375, Non-Symmetrical-Schmitt-Trigger.png [View same] [iqdb] [saucenao] [google] [report]

>3 comparators and few resistors
Well I meant more just a single adjustment resistor instead of having to have more than one passive outside your IC.
Actually, just having the resistors from the rails to V+ internally and having to add the Vout to Vin+ resistor(s) yourself would mean you could do it with the same pin-count as a normal comparator.
But then it wouldn't be usable as anything but a schmitt trigger with those resistors there, so having an enable pin that turns those resistors on or off via FETs would be needed for it to remain as a general-purpose comparator.

>> No.1616972

Fuck designing circuits on single sided pcbs is suicide fuel

>> No.1616977
File: 865 KB, 748x666, awl.png [View same] [iqdb] [saucenao] [google] [report]

I find it's all about strategically positioning your resistors such that they act as jumpers. But in a highly digital circuit with lots of wires going from one IC straight into another you don't really have this option. Hence why I'm delving into PCB rivets as plated through-holes for diy etched PCBs.
The 0.9x2.5mm rivets were a bit too small in both dimensions (a 1N4004 lead wouldn't fit in them and they didn't show enough length on the other side of the board), so I've ordered some 1.3x3mm ones. Not sure what I'll do about getting a drill bit to fit them though, hopefully a standard HSS 1.5mm will suffice. I've also ordered a nice, sharp, tapered leather awl for riveting them down while keeping the centre open, pic related.

>> No.1616981

How important is to add a tiny cap in dc circuits to remove ac ripple? It only matters in audio and such right?

>> No.1616985

Oh shit the resistor thing is genius. As for the rivets i like to use resistor legs and solder them on both sides or use them as jumpers on the bottom side

>> No.1616986
File: 21 KB, 426x344, 1534237106519.jpg [View same] [iqdb] [saucenao] [google] [report]

I think Maxim and AD/LTC have a few with adjustable hysteresis

>add as many debug/bodge supporting features as you can possibly fit. triple check everything
can confirm, this advice is golden

you mean the 56nF? that's for a 5ms TC on the current limiter

>to remove ac ripple
for this reason? not super important but you should still keep the supply as clean as you can so that your other circuitry (resistor dividers, dc shifters, etc.) doesnt couple any supply noise into audio or other sensitive paths
>to provide power for ICs that might have high switching or output currents
very important

Pic related

>> No.1616988

>Pic related
They come in THT as well

>> No.1616992

Nah man. Finding the ideal layout that doesn't require jumping between sides is pure zen. Besides, unless you're using a shit tier machine you can run traces under components which makes it super easy.

>> No.1616993


>> No.1616994
File: 125 KB, 851x1128, y tho.jpg [View same] [iqdb] [saucenao] [google] [report]


>> No.1616997

Gotta start somewhere, and it's a hell of a lot more forgiving when it comes to etching custom PCBs. I'll move on to big boy microparts eventually, when I get a hot air gun and/or reflow oven.

>> No.1616999

you don't need hot air to solder smd resistors

>> No.1617001

Are those from aliexpress? can you give a link?

>> No.1617006
File: 42 KB, 248x296, bs.png [View same] [iqdb] [saucenao] [google] [report]

I like circuit diagrams because they tell you a story without a wall of prose. This one starts with obvious bullshit (pic). A more interesting part is the unconventional use of the CA3086 to generate a stable reference voltage.

>> No.1617007
File: 10 KB, 423x407, 51749982_10205238476958653_1478417727143018496_n.jpg [View same] [iqdb] [saucenao] [google] [report]

Anyone here always register a new JLCPCB account to get discount??

>> No.1617011

>you mean the 56nF?
I think he was referring to the 5uF on the reference, though I think that's for stability. but thank you for extra info; was that a rough estimation or is it really as simple as the 100k/56n combo?

>> No.1617013

lol why do you say BS? because of the missing b/e connection.
amen though, Intersil had some notes + schematics but Renesas fucked it all up can't find them. I have a bunch of those Q arrays I salvaged from an old tek scope begging to be used, including 3 with a >1GHz cutoff

>> No.1617025
File: 22 KB, 477x294, ca3086.png [View same] [iqdb] [saucenao] [google] [report]

>Why BS?
Ok, let's call it 'material omission'.
Here's the interesting ref stuff, reverse Vbe, tempcomp and all.

>> No.1617033

MAX 921 has a dedicated pin for adjustable hysteresis.

>> No.1617034

Will it have a problem if i test my inverter with low VCE first even lower than VGE???

>> No.1617039
File: 3 KB, 235x208, inverter.png [View same] [iqdb] [saucenao] [google] [report]

No idea, I always test my inverter with a logic signal.

>> No.1617042

>etching custom PCBs
but then you gotta drill 'em
>big boy microparts
1206 (inch) is easy to solder with a 1.5mm screwdriver tip and the usual 0.8mm solder, as are SOICs. go to your local beauty supply and invest in some eyebrow tweezers if you don't want to wait for ali's slow boat. tacky flux is also very helpful. the saving grace of no soldermask is that an excess of solder has somewhere to go

oh, that one. derp

actually (100k||680k)*56n but close enough. I am supposing the delay is there to accommodate charging the output cap

>> No.1617043

>but then you gotta drill 'em
drilling pcbs is comfy

>> No.1617065

There are 3-5 common designs/models on the site and I'd encourage you to look through them, but note that some have a hole in the tip for feeding a thread through.

>invest in some eyebrow tweezers
Yeah I might do that since my chinesium tweezers are over 1mm out of alignment at the tip, but I'd rather have a fine-point hold-down/bench clamp. In any case all my ICs (like 500 of them) are THT so I won't be swapping unless I have a project that needs me to buy a bunch of new components. Not to mention I'm not exactly dealing with high-speed or high-density circuits.
>no soldermask
You underestimate me.

What do you use to drill them? I'm thinking some sort of vertical fixture like a mini drill-press would be nice, since I've already broken one cheap carbide PCB drill (the 1mm one).

>> No.1617078

a small electric grinder

>> No.1617094
File: 41 KB, 501x879, drill_stand.jpg [View same] [iqdb] [saucenao] [google] [report]

on a drill stand

>> No.1617096

I need a bias supply for some ICs on a project I'm working on, does anyone know a relatively simple/sensible way of getting a 1W @15V or thereabouts supply from a high voltage DC rail (say 180~200V)?

>> No.1617098
File: 41 KB, 1030x439, Screenshot_2019-05-25_13-59-59.png [View same] [iqdb] [saucenao] [google] [report]

Noob here trying to understand the schematic for an LED hourglass kit I've purchased.

I've managed to figure out how most of the circuit works, but I can't understand why they chose to use a 4069 hex inverter to make a hacky pulse generator instead of using a 555. Is there something I'm missing here that would mean a 555 wouldn't work? As far as I understand the 555 is designed to properly compare voltage levels of the discharging capacitor, whereas this relies on transitioning between logic high and logic low, purposefully transitioning through the indeterminate region, which I imagine would cause inconsistent behaviour / timing.

Full schematic is here but my screenshot includes everything that's relevant:

>> No.1617099

normal mains adapter runs on DC

>> No.1617100

I think you may be getting confused. RTS (technically RTR) and CTS are commonly used and supported. DTR/DSR and DCD are only needed for modems.

> although if you're only reading incoming data you can go from the -12v /+12v to 0v/5v with only three resistors .

Aside from the actual voltages, the levels are inverted. A "mark" (1/true/high) is -3V to -15V, "space" (0/false/low) is +3V to +15V. If you're using a microcontroller's built-in UART (rather than bit-banging a GPIO pin), you can't fix this in software (by inverting the data) because it applies to the start/stop bits as well as the data bits.

>> No.1617102

because 555s are shit?

>> No.1617105

How so?

>> No.1617106

>all relaxation oscillators are shit

>> No.1617107
File: 66 KB, 614x444, max232.png [View same] [iqdb] [saucenao] [google] [report]

I'll probably end up bit banging anyhow, unless someone happens to have a library out for ardy that works with this type of serial protocol.

High current draw and pin count for what it does, also >>1617106 to a lesser extent.

>> No.1617108
File: 265 KB, 640x640, PBO-1-S15.png [View same] [iqdb] [saucenao] [google] [report]

you're drawing too much current to use a simple regulator. fortunately many ac-dc modules can handle a dc input, and you're right there in the range of rectified ac voltages.

>> No.1617109

>many ac-dc modules can handle a dc input
I'd be as adventurous to say that they all do.

>> No.1617110

the little ones all can, but some bricks with PFC could get badly confused by DC.

>> No.1617112

Oh yeah that exists. Also arguably one could use a capacitive dropper to get LV to its controller with less wasted power, which could be significant in the case of a very low-current converter.

>> No.1617113

>why they chose to use a 4069
Because they had them and because CMOS? This is a standard oscillator arrangement and there is nothing inconsistent. I see no reason to prefer a 555 and its horrible supply spike.

>> No.1617117

no you can do it by hand
but it helps if you etch a small hole in the pads for centering

>> No.1617119

Thanks anon, I think that will do.
I was thinking about something I could implement myself rather than buying a pre-made module, but I can just try to copy the concept I guess.

>> No.1617133
File: 8 KB, 445x286, pwr-to-rpm.png [View same] [iqdb] [saucenao] [google] [report]

>you can do it by hand
You can and I did, but when you have to drill many boards the stand turns out to be very comfortable. Etching away the center comes naturally when you use laser printer plus UV exposure.

My next step was to modify the electronics in the drill from power control to (real) speed control. It's quiet and stable now, no annoying wheee wheee wheee any more..

>> No.1617140

There's indeed not a single bypass cap in the entire circuit.

>> No.1617218

>making my shitty diy pcbs by printing on foil and doing the uv thing and then etching with acid
>hmm i wonder of professional factories make them, i bet it's much better than this amateur trash i'm doing
>look up pcb makjng video
>they print the schematic on foil, then expose it on pcb with uv and then etch it with acid
holy shit i was the pcb factory all along

>> No.1617227
File: 221 KB, 1200x800, 1200px-Locomotive_ChS4-109_2012_G1.jpg [View same] [iqdb] [saucenao] [google] [report]

Is that also how the power/speed is adjusted in electric locomotives?

I'm still not sure what's the difference between the two circuits. High-side/low-side switch?

>> No.1617242

Sure, but where's the fun with that?

I think so too... Multiple dc-dc conversions seems complicated and inefficient.

Yeah, I started to think that 12 V panel might be more suitable for charging a small NiMH pack. Maybe 4..5 cells in series? I wonder if there's any way to build a "simplified" MPPT controller with logic ICs & comparators controlling a step-down switch-mode regulator.

>> No.1617249

if you don't want to use a prebuilt module, there are buck converters that run off mains-level dc. e.g. ST's VIPer06/16/26 ICs

MPPT isn't worth it unless you can undersize your batteries to save hundreds of currency units
you can size your battery pack and solar panel to stand up to a 1/10C trickle charge and not even need a controller

>> No.1617256

I like that picture a lot

>> No.1617259

Rivets are good too tho, but not always that practical.

Resistor legs can be fine, but they can also add unwanted interference/noise etc

>> No.1617271

>what's the difference
Left (org): Motor has adjustable power via phase angle, no feedback, loss of torque, load dependent speed (the whee whee effect). Right (mod): Motor (and its counter EMF) is now part of a feedback loop and acts on the phase angle. This enables adjustable speed and high torque (until you smell the amps if lack of fan speed). It's difficult to stop by hand because the full torque kicks in as soon as you try to slow it down.

>> No.1617339

> I wonder if there's any way to build a "simplified" MPPT controller with logic ICs & comparators controlling a step-down switch-mode regulator.
Solar panels should be followed by a boost converter, as they are constant-current sinks (buck converters are constant-current sources).

You can do all-analog MPPT, but it's a lot easier with a microcontroller. For a start, you need to be able to measure instantaneous power, i.e. voltage times current. Then you need to measure correlation between power and current. For an analog circuit, the first part needs a multiplier, the second can be done with a PLL.

>> No.1617373

>I wonder if there's any way to build a "simplified" MPPT controller with logic ICs & comparators controlling a step-down switch-mode regulator.

>> No.1617578

Not him but those are some nice links there. I'd quite like to learn a few of those subcircuits, they look like they could be handy for things like guitar pedals and audio filters and the like, maybe even radio circuits.

>> No.1617583

Is 0.8mm the idea diameter for pcb component holes? i want to buy one of them .8mm sets from chinks and use it to drill all the holes to fit all the diodes ics and caps, since i can't be asses to change the drills all the time

>> No.1617584

10/10 ingenious use of a CD4053 triple SPDT analog switch as an oscillator and synchronous detector

>> No.1617614

Given your condition I would use 1.0 mm. Dot board also has 1 mm holes. Pin headers fit but diodes like 1N540X do not, they need 1.3 mm.

>> No.1617615
File: 46 KB, 330x330, 23-emilia-clark.w330.h330.jpg [View same] [iqdb] [saucenao] [google] [report]


>> No.1617623
File: 783 KB, 2048x2968, saltonbepper_mics.jpg [View same] [iqdb] [saucenao] [google] [report]

help me /ohm/, you're my only hope

ok, so I'm building a couple lofi but high quality microphones from copper/steel salt & pepper shakers. the salt shaker is appropriately a piezo crystal microphone and requires no real electronics so that one's done. however, the pepper shaker is (also appropriately) a carbon granule microphone made from a carbon telephone microphone. those require a power source (supplied by the phone line). i want to power it with standard microphone +48v phantom power. i need help designing the circuit. i have a few ideas from an EE friend but he's not too sure it'll work and I'm too much of an electronics pleb to know the difference....

>> No.1617630
File: 19 KB, 425x419, 1375773299400.jpg [View same] [iqdb] [saucenao] [google] [report]

Do you guys think the arduino is fast enough to perform signal processing on a sound to isolate the most prominent frequency?

>> No.1617631

you trying to make a feedback eliminator or similar anon?

>> No.1617632
File: 28 KB, 448x750, PolyTune3-large.jpg [View same] [iqdb] [saucenao] [google] [report]

No I want to embed an arduino and screen into my electric guitar and act as a tuner since I'm too poor to afford pic related

>> No.1617634

Fair enough. Also there are probably Flash tuners you can use free online if you plug into the pewter. Come to think of it, there are probably a multitude of free software tuners you download anyway. Not to say that will suit (that wouldn't be the least bit /diy/ after all) but it's an option

>> No.1617639
File: 7 KB, 400x400, tegaki.png [View same] [iqdb] [saucenao] [google] [report]

Something like this? It's basically what I'd use for a HV geiger counter or electret, and I assume the same thing applies.
1/(2π*R1*C1) = f_cutoff = ~20Hz
C0 is just there for symbolic reasons, while R0 needs to be of enough of a value to not sink too much of the signal into the PSU, but not so large that the mic doesn't get enough current. Putting a potentiometer here instead of the resistor wouldn't be a bad idea, especially for testing. You may also have issues with a 48V voltage spike occurring at the output every time you turn it on, so putting a resistor in series with C1 would probably help.
If you had a datasheet for the carbon mic I might be able to scrounge up some actual component values.

>> No.1617641
File: 437 KB, 1662x1290, Screen Shot 2019-05-26 at 8.41.40 pm.png [View same] [iqdb] [saucenao] [google] [report]

ok so this is all over my head but thank you anon! here are the specs from the seller. I'm pretty sure the DC power in phone lines is 6-12 volts so that's what it's designed to handle.

Also, a mic cable is 3-pin XLR (+ audio, - audio, and then ground, with the 48v phantom traveling up the audio pins). I feel like that's one of the parts that complicates things.

>> No.1617644

Ok that's good enough I guess. Problem is this sort of low impedance (compared to geigers and electrets) means you'll be sinking a lot of current at any decent voltage, which might be why a lot of these circuits are using transformers. Also I can't seem to sim the damn thing using a jfet as the variable mic impedance so I'll resort to using paper.
Here's a link that might help a little:

>> No.1617650
File: 17 KB, 534x298, huh.png [View same] [iqdb] [saucenao] [google] [report]

Ok now that I've done some testing with this neat feature, I've found two things. Firstly, the resistor in series with the PSU has to be such a low value that it's probably better to use an inductor, with a cap after it to ground. Unfortunately, the voltage spike issue is much worse with an inductor than with a resistor. Secondly, 48V through 65Ω is a lot of power, so I'd use no more than 12V.

Another important thing to note is what sort of audio input you'll be using. Simple computer mic inputs typically (at least the TRRS ones on laptops and phones) have their own low-voltage (3V or so) high-impedance (3kΩ or so) phantom power designed for use with electret microphones. To bypass this you'd need a capacitor in series with your audio output and possibly a buffer amplifier in there somewhere too if your output impedance is too high.
But a proper audio mixer, and maybe also a USB soundcard, will probably not have this issue.

>> No.1617651

More findings: an LM317-based constant-current supply would work tolerably well instead of an LC filter if your resistor can handle the current at that voltage drop, and a silicon diode or two should handle the startup spike pretty well.

>> No.1617653
File: 73 KB, 1296x752, test 1.png [View same] [iqdb] [saucenao] [google] [report]

Here's my test circuit, with the LC part disconnected:

>> No.1617655

Thanks again anon! Still a bit over my head but this is stuff I can take to my friend to look over and see if we can build this.

To clarify, this mic is designed to be run from a proper microphone preamp. That is, it has a 3-pin XLR, an input impedance of 1.5kΩ-3kΩ, and it provides 48v phantom power to pins 2 and 3, which are also the pins the audio signal comes from. It'll need a transformer but it also needs the 48v to be toned down to ~12v (or less). My brain is even more scrambled now than it was when we started!

Cheers m8, you're a champ

>> No.1617658

Well if you have a 48V supply, you might be able to just use a resistor to drop that voltage, which will approximate enough of a constant-current supply that it should work pretty well. But it will be even more heat for the resistor to drop than the mic itself, so running it off a buck converter or other such LV power supply might be preferable.
>it'll need a transformer
You mean for impedance matching? You might be better off doing that with a bunch of passives that approximate a matching network.

If your brain is a little scrambled, I'd make sure you have a handle on Ohm's and Kirchoff's laws, and probably also on basic capacitor and inductor behaviour in both AC and DC circuits. Then filters and phasors and the like would make a good topic too I suppose, followed by whatever the hell impedance matching is supposed to be.

>> No.1617680

>Ohm's and Kirchoff's laws
If Kirchoff then Om.

>> No.1617738

Hard no

>> No.1617740

the internal adc goes up to 15ksps so you could capture the lower half of the audible range, to ~7khz i guess. that's if you compile your own asm or c though, the arduino abstractions might fuck you no matter what. you also couldn't process the sound in realtime. you'd have to sample a small portion and then spend seconds processing it.

>> No.1617743

Buy the viagra instead, it is compatible with arduino ide and much faster

>> No.1617747

Okay i will order 1mm and 1.3 mm set for fatter components

>> No.1617749

Can you program the arduino to do sinc interpolation on your samples?

>> No.1617751

Why don't you just use your phone with a tuner app?

>> No.1617756

I'm not quite sure what your intended application is but if you are trying to find the lowest partial (fundamental) in a given phrase of music you do realize a simple FFT based peak detector won't necessarily do that. The highest sample in a given set of music isn't always the fundamental. It can often be low frequency bass hits or sharp high frequency peaks which can often be several dB above your average signal level depending on the genre of music.

>> No.1617775
File: 44 KB, 704x481, carbon-xlr.jpg [View same] [iqdb] [saucenao] [google] [report]

You already have the circuit. What you need is a real carbon microphone you can hear when you shake it and an audio transformer with a center tap. It will behave like a normal balanced mic. If it is too loud you can encrease the value of the 6K8 resistor to 10K or more. The expensive carbon mic is from https://www.payphone.com/carbon-xmit.html but any other one salvaged from an ancient POTS phone will do as well.

>> No.1617817
File: 43 KB, 1742x1084, tube_headphone_amp.png [View same] [iqdb] [saucenao] [google] [report]

Where is a good place to put volume pots in an audio (tube) amplifier? Input, output, feedback loop, or somewhere inter-stage? I can't really seem to get good results anywhere due to the pot loading the circuit in unusual ways. Should I add a cathode follower or a more conventional emitter follower somewhere?

I'm using a 100k pot with the tap in parallel with a 33k to create a fake log pot as described by

>> No.1617874

each stage introduces noise that is independent of input level. each stage's input noise adds up. I suggest putting the volume pot toward the output end of the circuit, like on the control grid of VL3
>due to the pot loading the circuit in unusual ways
make sure it's ac coupled, also consider placing a small cap (a few hundreds of pF) across the clockwise and wiper terminals to counteract resistive loading

>> No.1617889

Alright, I'll give that a shot and report back with the results of the sim.

>> No.1617894

I did intended to direct couple that stage but I guess AC coupling it has the benefit of making D1 and R21 unnecessary since now I don't have to worry about arcing between the grid and cathode before the heater warms up.

>> No.1617915
File: 46 KB, 1805x1088, tube_headphone_amp2.png [View same] [iqdb] [saucenao] [google] [report]

Like this? Seems to work reasonably well. I do need to lower R24 from 10k to 3.3k to get roughly the same amount of gain but it's not a big issue. Should work okay.
I do have some distortion for frequencies under 1kHz. I don't know where it starts exactly but probably somewhere between 200-500Hz and gets worse the lower you go. Not something caused by the pot though. I think it's a circuit design issue.

>> No.1617916

That wire bridging C20 to C21 shouldn't be there, oops.

>> No.1617920
File: 66 KB, 1920x1080, tube_headphone_amp_20hz_freq_response.png [View same] [iqdb] [saucenao] [google] [report]

Yeah I can't accept this. No good.

>> No.1617925

How does it run without and volume control?

I notice R28 with R29 and assume that is your volume control - not a good spot for sure, remove it. I would put it between V4 (signal in) and C8. That would be the ideal place.

>> No.1617953

I have the carbon button and the transformer pictured. the circuit was an idea my friend drafted up but he's not 100% confident in it. the thing i really need to work out is powering it safely with 48v phantom power

>> No.1617973

C20 is good but I was thinking more like Pic related. C21 just adds phase delay and brings the possibility of parasitic oscillations

>using a voltage regulator diode to set the bias
that sure does look like crossover distortion to me. try two diodes in series, each thermally coupled to one of Q3/Q4

I don't see anything wrong with it. that circuit will only put about 7mA through the mic element which should be plenty comfy for it

>> No.1617976

Is it possible to have crossover distortion at low frequencies but not at high frequencies because at 1kHz to 20kHz the waveform looks good. Even with FFT 2nd harmonic starts at about -50dB at 1kHz

>> No.1617984 [DELETED] 
File: 76 KB, 589x468, 1531031185532.png [View same] [iqdb] [saucenao] [google] [report]

fuck, didn't pic related

>> No.1617985

Oh, I see. I wasn't sure exactly where that small cap was supposed to go. I thought you meant from the tap to ground.

>> No.1617988
File: 78 KB, 589x468, 1534320199296.png [View same] [iqdb] [saucenao] [google] [report]

fuck, didn't pic related then wrong pic related

>> No.1617991
File: 899 KB, 2560x2840, Clipboard.png [View same] [iqdb] [saucenao] [google] [report]

Alright well I tried using two 1N4149s and two diode connected MJE340s in the circuit to replace the LM385 for the bias for the class AB but it hasn't made much of any impact on the waveform. The two diode connected transistors have shifted the waveform up slightly but the waveshape is still the same.

>> No.1618010
File: 2.37 MB, 4160x3120, IMG_20190527_011913041.jpg [View same] [iqdb] [saucenao] [google] [report]

>be bored
>decide to open up LED mood light to see what's inside
>*notices buldge*
>uwu what's this?

Didn't notice any problems before but I guess I need to replace this.
I know nothing about lipos.
If I just buy another lipo with the correct voltage and wire it up will it be fine?
Is the charge control circuitry on the battery or on the board? If it's on the battery what do I need to look out for when buying a new one, and if it's on the board will it work fine with a battery of a different capacity or do I need to buy one with the same mAh?

>> No.1618015

set more scopes

as long as the replacement is no smaller in capacity, the dimensions should be printed on the package and encoded into the part number
show better picture of board and IC markings but my guess is U3 is the charge controller. that failure could also be pressure- or heat-related, is that adhesive foam block bit for real?

>> No.1618024

also, what's up with those current regulator diodes? their impedance seems really high compared to the pullups

>> No.1618030
File: 2.06 MB, 4160x3120, IMG_20190527_015136939.jpg [View same] [iqdb] [saucenao] [google] [report]

Thanks for the advice
U3 has "5056 1726" printed on it, and I can't find anything from a quick Google search.
Here's some more pics, sorry for the crappy photos.
Also out of interest what's so bad about the foam?

>> No.1618032
File: 2.75 MB, 3120x4160, IMG_20190527_015439152.jpg [View same] [iqdb] [saucenao] [google] [report]


>> No.1618041

It's an active load. Using current sources for loads increases the gain of an amplifier significantly. They usually are implemented as current mirrors though but it's high impedance either way because it's a current source.

>> No.1618045

looks like a TP4056 clone. those can throw off quite a bit of heat, was the battery near it? if so try to move it away
the uneven compression gives me the creeps

>> No.1618052
File: 2.28 MB, 4160x3120, IMG_20190527_022534242.jpg [View same] [iqdb] [saucenao] [google] [report]

The battery goes right underneath the circuit board like this, then that smaller diffuser screws over the top, then the big sphere goes over the whole thing and is water sealed at the bottom, so the whole thing is tightly enclosed with zero airflow - I think you're right about heat being the cause.
I can't really move the battery because it would block the light in any other position, perhaps drilling some airflow holes in the bottom of the plastic base would be a good idea?

>> No.1618062

What is the best way to program on the STM32 Blue Pill? Do I bother trying to install a usb bootloader?

>> No.1618066

depends on how much you care if it's still sealed
but taking the screw-on dome off whenever you charge is a good idea

probably not worth the effort of installing one on chips that don't already have one in system ROM. SWD works great for me

>> No.1618074

Alright yeah I'll figure something out, the bigger dome is attached with screws so it's going to be a lot of effort to keep taking that on and off, I'll probably just drill some holes because I never cared for the waterproofing anyway.
Thanks again for the help, I appreciate it.
One final thing though, where do you buy these flat lipo batteries from? I thought they would be all over ebay because of how common they are in consumer electronics but all that comes up when I search for "4.2v lipo" or "4.2v lithium" are RC batteries and AA shaped batteries. I came across a couple on AliExpress just now, but surely that can't be the only place to get them?

>> No.1618081

i appreciate the concentric trace layout. i hate fuckers that use rectilinear on a circular board.

>> No.1618086

good question, depends where you are because of hazmat shipping restrictions. google shopping may be your best bet

>> No.1618092

Where can i get that same mood light?

>> No.1618102

How does one protect electronics from the sun/rain? I was given a project to setup a car counter with Raspberry Pi and a camera, but I'm not sure on where to get resources to how to protect the thing (as it would be setup on the street).

Any advice?

>> No.1618109

ip-rated enclosures (and connectors), in boxes with materials that are vandal resistant and aren't uv sensitive if it's a long term setup. you can spend anywhere from $10 to $6k on one so shop around. hammond mfg has a good selection that's often on digikey/amazon.

>> No.1618242

This is the exact one I bought 2 years ago, but this listing isn't sold anymore:
It's just a chink light that's rebranded by a bunch of sellers though so hopefully you can use that listing to find one that's the same. I paid £25 for it 2 years ago.

>> No.1618267
File: 17 KB, 718x512, PSU.png [View same] [iqdb] [saucenao] [google] [report]

I have a circuit which I want to run from either mains or batteries. Basically, I want it to be powered by a power supply and charge the batteries when it's plugged into mains, and then run off the battery when it's disconnected.

What's the best way to do this? Is pic related a good idea, or can you just charge the batteries under load and basically use the charger as a power supply? I'm thinking I'll need a buck-boost converter or something to regulate the output from the battery anyway, so I'm not worried about the higher voltage from the charger.

>> No.1618268

>run off the battery

you mean run on batteries

if it ran off the batteries it wouldn't have any power

>> No.1618279

>>1618267 Lead batteries could be used in buffer mode with no problems. Lithium is not, but you still could use charger as main power supply.
Also there is no need in relay. Just get a 15v psu.

>> No.1618280
File: 15 KB, 489x512, PSU-BAT.png [View same] [iqdb] [saucenao] [google] [report]

What about this?

>> No.1618286

The relay isn't there to protect the PSU, it's to avoid using the batteries in buffer mode.
But if that's not a problem then I might as well just completely get rid of the PSU, right?

But I should mention that I'm planning to have two batteries in parallel so that I can hotswap them, which I think will mean either a two-output charger or two separate chargers. If they're charging at different voltages, then I wouldn't be able to tie them together and use them as a power supply, would I?

>> No.1618294
File: 318 KB, 1062x1375, chargin mah lazer.png [View same] [iqdb] [saucenao] [google] [report]

how big of a circuit? devices between the size of a phone and the size of a laptop have some form of power path management. Pic related is just one example of hundreds of off-the-shelf solutions to the problem of directing current flows around a battery-powered system. Microchip has some too, see MCP73871 (resistor-programmed instead of I2C-programmed)

>> No.1618295

If you slap an inductor between a power source and a load, and the power source is say 24V, does that mean that instead of the load getting full on balls deep 24V, it will slowly ramp up from 0V to 24V?

>> No.1618319

Yes. The voltage at the load will be 24*(1-e^(-t/T)) where T=L/R.

But note that while the inductor limits the rate at which the current can increase, it also limits the rate at which it can decrease. If you disconnect the load or the load attempts to reduce its current draw, you'll get a voltage spike, possibly a large one.

>> No.1618347
File: 39 KB, 407x398, magnets.png [View same] [iqdb] [saucenao] [google] [report]

I want to use a TLE5012 magnetic angle sensor.
Common sense tells me I need a diametral magnet for this. Am I correct?

>> No.1618371

If you have some wimpy finger sized inductor, what kind of voltage are we talking here if you rip off the load?

>> No.1618377

Whatever is needed to maintain the current. The inductance determines how rapidly the current will drop off for a given voltage: V=L*dI/dt.

If you have a switch in the path, opening the switch will result in an arc as the current continues to flow. You can get around the issue by putting a flyback diode in parallel with the inductor, or using a snubber network.

>> No.1618394

>Whatever is needed to maintain the current
there must be a limit, surely the inductor can't poop out billion volts

>> No.1618406

>chinks will make you 10 10x10cm pcbs with a silk screen, solder mask and everything like that for two bucks
how in the fuck do they make any profit? i can very barely diy a single pcb without a solder mask and a silkscreen for $2

>> No.1618420

I've used those services from a few different companies. They advertise a super cheap price but then when it comes to checkout it's around $20 not $2, plus an additional $15 in customs and import fees you don't normally have to pay to get something from China.

>> No.1618423

the limit is when something that shouldn't conduct starts to do so anyway. usually the air. hence why you can see sparks in your wall outlet when unplugging a running fan or vacuum or whatever.

>> No.1618425

even after the new tariffs i think i pay under $10 at elecrow for 5 10x10s.

you're also not counting the monetary value of your labor, which makes the disparity really ridiculous.

>> No.1618438

But that would mean that even a tiny battery powere inductor could kill you if you grabbed the leads and removed the load and then it would shove like 0.5A through you

>> No.1618456

> there must be a limit, surely the inductor can't poop out billion volts
The limit is whatever is needed to maintain the current, at least instantaneously. In practice, parasitic capacitance will absorb some of it. As for the rest, the voltage peaks at the point that something which should be an insulator (air, winding insulation, whatever) becomes a conductor.

Most methods of generating high voltages use inductance in one form or another.

>> No.1618459

Yes, that's correct.

Bear in mind that 1) the smaller the inductor, the faster the current drops off, 2) the higher the induced voltage, the faster the current drops off. So if it's a very small inductor, you might only get that 0.5A for a nanosecond.

The total energy stored in an inductor is (1/2)*L*I^2. That energy can't just disappear; when dumped into a resistive load, it's equal to the integral of V*I w.r.t. time.

>> No.1618465

>inductor could kill you
No, skin resistance is too high and the spike is too short. It's the volts that jolts, it's the mills that kills said the navy man.

>> No.1618468

from memory there's also a minimum duration for the lethal current to actually kill you, which makes it even harder for tiny inductances to be fatal.

>> No.1618473


This way. Its easy to install an usb bootloader after you have done it once

>> No.1618474

Flash an ST-Link clone as a "Black Magic Probe" for comfy flashing and debugging.

>> No.1618477
File: 89 KB, 824x680, electric-safety.png [View same] [iqdb] [saucenao] [google] [report]

Yes, image from my eshock folder.

>> No.1618485

>It's the volts that jolts, it's the mills that kills said the navy man.
The thing that annoys me about this (cute) saying is jolts are also from current, and you don't get current without voltage. A "jolt" isn't merely uncomfortable, something that gives you an uncomfortable zap can also kill you.

>> No.1618513

is the viagra worth it? i really like my fartduino nano's, and it's nice that they enjoy 100% library support

>> No.1618514

> it's the mills that kills
Right. But an inductor is a current source. The voltage will be whatever needs to be to maintain the current. High skin resistance makes the current drop off quicker, but it doesn't affect the initial value (which is whatever was already flowing through the inductor).

You can quite easily kill yourself with a large enough inductor and a low-voltage battery.

>> No.1618518

>You can quite easily kill yourself with a large enough inductor and a low-voltage battery.
But there's no ground reference.
You'd have to hold both ends with both arms - why would you ever touch an inductor like that?

>> No.1618584
File: 11 KB, 532x526, selfiecide.jpg [View same] [iqdb] [saucenao] [google] [report]

>why would you ever
Between the lines
The urge to kill

>> No.1618599
File: 2.09 MB, 1664x3744, rat guts.jpg [View same] [iqdb] [saucenao] [google] [report]

Wew, first time holding a soldering iron in 11 years. And first time soldering together anything significant. Went pretty well I think.
What an absolutely relaxing way to spend your time.
Anyway, just built my first real guitar pedal (well, the actual first was an A/B box but that doesn't count cause there's not really any circuitry in it). Was kinda worried about the holes (which I drilled by hand) since with the A/B box the drill fucked off on me and the holes weren't in the right place. Took it slow this time. Centerpunching and first drilling a really small pilot hole with some WD40, then drilling a big enough pilot hole for the step drill, and then carefully drilling the rest of the hole. Came out perfectly.
Now I just need to think of a nice way to finish it. Kinda wanna etch it but I've never done that before and storing corrosive chemicals in my bedroom doesn't sound like the best idea.

>> No.1618643

Okay, this is epic.
E-bike has PWM switching frequency around 15 kHz. Why it makes bars go down on phone (2100 MHz)? How can I make it FCC compliant?

>> No.1618644

>How can I make it FCC compliant?
More shielding

>> No.1618645

You need autism.
I always liked designing complex boards on one layer, while keeping amount of jumpers minimal.

>> No.1618647

Well, should I ground chassis? Controller is inside aluminium box.
Wrap wires in aluminium tape? Put caps across phases?

>> No.1618671

automation and a continuous flow

it's the amps that cramp

that's an engineering judgment. if you're content where you are and can do all the things you want to do without problem, no great reason to move other than shame
that said, in addition to more RAM and more flash, the STM32 line does provide some useful and interesting capabilities once you step outside of the arduino crib. the Cube tool does a lot of the setup drudgery work for you and allows you to change your pin/peripheral setups and MCU selections later, just fill in the BEGIN/END blocks in the generated code and use the proper defines
whereas duino's emphasis is on easy-mode least common denominator microcontrolling, one of ST's approaches to hardware abstraction is to present one pure, task-oriented interface to some particular peripheral function to enable easy migration of code across the entire STM32 line. the other is a helpful set of low-level libraries and macros

isn't it though? powder coat best coat, of course

>bars go down on phone (2100 MHz)?
is it the motor and/or controller, or could it be something else like Doppler shift and/or delay shift? 30cm/ns in free space means a lot of uncertainty at scooter speeds when trying to fit into a TDMA slot
>ground chassis
kek, to what? you don't have a low impedance path to earth if you're on a rubber-wheeled scooter
>inb4 I'll drag a chain
you'll just create a spark gap generator which is worse

>> No.1618682

>or could it be something else like Doppler shift and/or delay shift?
No dude. Cell phones work well in moving vehicles.

>> No.1618690

>is it the motor and/or controller
I don't know. Shit doesn't make sense anyway. I have doubts that MOSFETs used there can into such frequencies (50 ns rise time, 50 ns fall time, that is 10 MHz, and shit would heat as hell). MCU is running on 16MHz.
Arcing can produce such shit, right?
> or could it be something else like Doppler shift and/or delay shift? 30cm/ns in free space means a lot of uncertainty at scooter speeds when trying to fit into a TDMA slot
No. Speed is too low for this, this is not a fucking fighter jet.
>kek, to what? you don't have a low impedance path to earth if you're on a rubber-wheeled scooter
Battery ground to chassis.

>> No.1618693

Square wave harmonics?

>> No.1618694

Maybe. I need an oscilloscope for this though.
Shit, it sucks to do electronicals without oscilloscope... Which one should I get, btw?

>> No.1618696

they don't need to go from full on to full off to create RFI. they just have to be a bit stronger on an inverse square weighted basis than the tower
MOSFETs can ring if there is excess inductance in the gate lead. may need to add/adjust gate resistance (but too much makes the FETs transition slower and run hotter)

Rigol DS1054Z with the hacks, of course

>> No.1618698

>Arcing can produce such shit, right?
Also how are you snubbing the inductive kick from switching mosfets?
Do you have a probe to see what's emitting the most EMI?
As for the cables, wrapping them in shielding tape might help... 3m makes shielding tape out of copper or tinned copper.
Closely coupling the positive and negative current paths for the motor may help.

>> No.1618699

Broken soviet CRT dumpster scope for $10 from the junk shop, of course.

Can those do fourier transforms? It's the only thing I'd particularly want with my new semi-analog scope, and also why I want to hook up the D25 RS232 serial interface so I can process the data in python with my computer. Interfacing directly over USB sounds like a pain, so I've decided to simply write the data onto a micro SD and read that with my computer's SD slot afterwards.
Got a shark card on order along with the USBASP and microSD+logic level shifter module and MAX232.

Would ferrite beads on the battery cables do any good?

>> No.1618703

>MOSFETs can ring if there is excess inductance in the gate lead. may need to add/adjust gate resistance (but too much makes the FETs transition slower and run hotter)
I didn't touch it, I have doubts chinks did something that wrong.
>Rigol DS1054Z with the hacks, of course
Well, too expensive.
Okay. I know what is happening now. Fucking wiring is doing this again I guess.
>Also how are you snubbing the inductive kick from switching mosfets?
No, I guess... https://avdweb.nl/Article_files/Solarbike/Motor-controller/China-BLDC-motor-controller-36V-250W.pdf
(older version of controller, but power circuitry looks identical. Only difference is low voltage power, in my it is done via SMPS of sort)
>Do you have a probe to see what's emitting the most EMI?
No, I don't even scope.
>Broken soviet CRT dumpster scope for $10 from the junk shop, of course.
But what about USB scopes? I have couple laptops that I don't mind frying.

>> No.1618715
File: 325 KB, 1280x960, 1028199223.jpg [View same] [iqdb] [saucenao] [google] [report]

>USB scopes
I've heard mixed results. I have a USB logic analyser and it seems to work pretty well for the $7 I spend on it though, so a USB scope might be worth getting. Look at some online reviews. From what I remember a bottom-line hantek was around $50, but I doubt they'd be high frequency.
I'd start by looking for 2nd hand scopes near you. The soviet dumpster scope thing was just me memeing you (though pic related is being sold local to me), but you can find fairly decent old scopes that are still pretty up-to-date. Nothing with more than 2 channels or above 100MHz in all likelihood, but that's still pretty good for a lot of uses.

>> No.1618722

Too bad I live in non-diy (when it comes to electronics) shithole...

>> No.1618806

Well what's your price range anyhow?

>> No.1618811

Lower=better. But I don't want buy complete junk at same time.

>> No.1618818
File: 44 KB, 592x369, 1546819292449.jpg [View same] [iqdb] [saucenao] [google] [report]

I picked up an old solid state transceiver at hamfest. I hooked it up to 13.8V but it doesn't power on. The ammeter on the power supply slowly climbs from 0 to 15A when the inline fuse blows and something smells hot in the radio. Nothing looks obviously out of order on the boards. Caps aren't visibly leaking. What's a good strategy for finding the faulty component without this thing catching on fire?

>> No.1618846
File: 26 KB, 548x396, Capture.png [View same] [iqdb] [saucenao] [google] [report]

I want to build a micro-power led flasher using lm3909 in cascade with a dark-activated switch at battery's junction. What battery estimation should I look at to power 4-6 flashing leds using this approach. Picrel is the switch circuit

>> No.1618860

try seeing if Vcc shorts to ground, lmao. Start with coils I guess

>> No.1618885

What is this ripple thing i keep hearing about when using bucks that isn't present in linear regulators?

>> No.1618892
File: 56 KB, 681x518, 1531504158482.png [View same] [iqdb] [saucenao] [google] [report]

>What battery estimation should I look at
right in the datasheet, at the bottom of the application schematic

>> No.1618894
File: 105 KB, 699x638, 1551991198121.png [View same] [iqdb] [saucenao] [google] [report]

read harder, third from the bottom

>> No.1618898

I am not a galaxy brain, i don't know what that means ;_;
Decibels represent sound why do they mater in voltage conversion is a mystery to me

>> No.1618903

switching power supplies and unregulated transformer-rectifier power supplies will generate some amount of ripple in their output voltage proportional to load current. in the first case at tens or hundreds of kHz, in the second case at twice the mains frequency
decibels are a logarithmic measure of a ratio. for voltage or current, 20dB = a factor of 10, 40dB = a factor of 100, etc. what once was multiplicative becomes additive. so, for example, 20mV of 120Hz ripple on the input of the LM317 will be attenuated by 65dB to roughly 10µV at the output of the regulator. some other circuit downstream with only a 20dB power supply rejection ratio would couple 1µV of the original 120Hz ripple into its output, which is important for a designer to be aware of, lest it be unacceptably amplified by some subsequent stage

>> No.1618909

>Would ferrite beads on the battery cables do any good?
Not sure actually
I can't recommend the book "electromagnetic compatibility engineering" by Ott highly enough, check it out, even get a pdf and skim it if you don't want to buy.
It's the bible of shielding and noise reduction.

>> No.1618925

i see, that makes much more sense, luckily i don't work with such voltage ripple sensitive circuitry so that shouldn't be an issue for me at all
Also if i understand it correctly, then if you need to remove ripply from say some shitty buck converter you can just slap say the lm317 behind it to considerably flatten then voltage line right?

>> No.1618944

Wrong. You remove ripple with low pass filters, not with regulators.

>> No.1618947

A regular regulator is a pretty effective low pass filter at 120Hz

>> No.1618970

yes, the state of the art in bench power supply design is to do just that

>> No.1618972
File: 21 KB, 611x319, H6YzCie.png [View same] [iqdb] [saucenao] [google] [report]

Question regarding orange-pi zero-plus GPIO labeling. IS the SCK in this picture SCL? It's next to the SDA and has the same prefix.

>> No.1618977

yep, two-wire clock however you want to spell it

>> No.1618978
File: 320 KB, 1477x1108, 20190528_185209.jpg [View same] [iqdb] [saucenao] [google] [report]

Why the smaller one can handle more current???
bigger wire??

>> No.1618983

bigger wire, yes, also different core material with higher inductance factor, also possibly optimistic chink uprating

>> No.1618988

Looks like a staged fake to me.

>> No.1619002

Hello, /ohm/.
I want to buy an oscilloscope for building/troubleshooting/tuning DIY synthesizers and other music devices.
I don't have much space on my desk, so I'm wondering if I can get away with using one of those cheap portable Chink scopes.
Anyone have experience with this type of stuff?

>> No.1619003

Do you have a pc on your desk? Get a USB scope.

>> No.1619005

thicker wire, better ability to radiate and conduct heat outwards, and more heat tolerant components. But 4mH may be pushing it a fair bit, I'd rate it no more than 400µH.

>> No.1619006

Yeah there's usually my PC or at least a laptop right next to where I build stuff.
Can you recommend one?

>> No.1619008


I measured across the DC input and it's not a direct short, it shows around 400M ohms which jumps around a lot. It seems like some sort of feedback loop? with how it goes slowly from 0 current use to a high amount, but nothing on the radio is powering on (lights, etc).

Normally I'd trace through the circuit looking for where it fails, but I can't do that as I think this would catch on fire if I left it plugged in for more than 10 seconds.

>> No.1619014

A thermal camera would be a great method for seeing where the heat is going, but without one perhaps with a bit of anhydrous alcohol of some sort you might be able to see where hot traces/components are going by where it evaporates first in those 10 seconds before death. Just feeling about with a cautious finger after powering it on should help. If it isn't smoking it likely isn't a problem with any windings, and if truly no caps are gone (check the old film caps) it's probably a silicon problem. My bet is that the main antenna driving transistors have shit the bed, since I'm assuming it uses something like a half-bridge or totem pole somewhere along there.

Is there a service manual for that model that includes a circuit diagram? Looks old enough.

>> No.1619020

Hi all, hope you can point me in the right direction.
I'm looking for a transistor that can drive a high voltage (12v), low current load.
I'm having trouble with google, as everyone's trying to drive high current loads there.
I'm looking for something small and cheap that can give a couple milliamps to trick something reading a voltage level, nothing more.
Stuff like the TIP120/IRF510 are totally overkill for this application.
I appreciate the help, I never go over digital logic levels in my own work so I'm kinda clueless here.

>> No.1619024

3904 (NPN) or 3906 (PNP)

>> No.1619026

>high voltage
u wot

Anyhow you're looking for small-signal MOSFETs, so filtering by small transistor packages (SOT-23 or TO-92 depending on orientation) might help greatly.
I'm assuming you want a MOSFET since you'll get a diode drop through an NPN that will be suboptimal for the purpose of measuring voltage, at least without taking a bit of care.

>> No.1619029

It's high voltage FOR ANTS

>> No.1619058

Looks like just what I'm after, cheers!
>high voltage
For digital 12v is high voltage.
>suboptimal for the purpose of measuring voltage
It's more the output that I'm after, I'm emulating a voltage for an old analog display dial that I've been tasked with digitising and modularising the internals of, while not changing the interface.

>> No.1619072


OK I will give that a go, it does seem like the heat is coming from the area around the final output transistors and that's where the power comes into the unit. Thank you sir

>> No.1619075

>For digital 12v is high voltage
No it's not. 4000 series CMOS can do up to 15V no problem and any small signal transistor can switch at least 30V at low current, except perhaps most JFETs. Common 2N3904 has a Max Vceo of 40V, the PN2222 is 30V, and the 2N7000 will do 60V.

>> No.1619115
File: 7 KB, 501x265, LM317-constant-current-power-supply[1].png [View same] [iqdb] [saucenao] [google] [report]

Is this good enough to feed a humble 250mw laser?
The l1 will be the laser and the R1 will be a potentiometer, which i will very slowly start turning from zero up, to find the current at which the laser turns on

>> No.1619116

>from zero up
what is 1.25/0?

>> No.1619121


>> No.1619125

>which i will very slowly start turning from zero up
>to find the current at which the laser turns on
You don't know what you're doing. If you want an explanation, ask.

>> No.1619127
File: 49 KB, 307x284, 317_current.png [View same] [iqdb] [saucenao] [google] [report]


>> No.1619133

>You don't know what you're doing.
Doesn't matter, i am just following a tutorial, which is for a weaker led so i am just checking if lm317 can drive a stronger laser

>> No.1619144

If by zero you mean maximum then yes, more resistance = less current
If you start with the pot all the way down, you will melt the thing

>> No.1619219

not necessary with a battery, when current is increased, the voltage is down so the amperage will never go to infinity

>> No.1619239

here it datasheet

>> No.1619265

Battery or not, current can't go to infinite, and you don't need a lot of current to fry a laser diode. Don't suggest relying on battery's internal resistance for >>1619115 to not fry their laser.

>> No.1619268

Well shit, that has to be a far better core material than what I'm used to. You could calculate the core's magnetic permeability with an online toroid inductance calculator and the core dimensions and number of turns, if that would interest you.
Actually I did so right now, and the relative permeability looks to be around 2000, which isn't actually that fantastical.

Yes, the ESR of the battery will limit current, but only to a value that's likely far larger than you'd want. Same reason people don't drive single LEDs off car batteries.

>> No.1619284

Ground wire choke, open design
What does this mean??

>> No.1619288

I think when it comes to ferrites and current if you want high current lower permeability is better since you can handle higher currents before your core saturates. For higher power switching converters having an effective permeability around 50-250 is usually desirable. You'd usually used gapped cores for this or something like molypermalloy which has a distributed air gap.

>> No.1619363
File: 114 KB, 1000x812, 1556785997296.jpg [View same] [iqdb] [saucenao] [google] [report]


>> No.1619386

Well shit, good to know.
I'm guessing the bag full of burned out 2n2222's was just me doing something stupid then.
Thanks for the tip anon.

>> No.1619400

Could you put an electric field across a PIN diode to turn it into a JFET I wonder? Also I want to see precision manufactured galena/other cat's whisker crystals put in a resonant cavity with a voltage across them because crystal lasers sound like my kind of aesthetic.

is this OC

Sounds like you were putting too much current through the things. You always use a base resistor, right?

>> No.1619425

I will ask here, since /g/ is stupid.
I have a laptop, which has insanely low PWM frequency, like 200 Hz. I can sorta see shit blinking, actually. Especially on low brightness.
After reading a lot of datasheets, I still don't understand how LEDs are driven. Do LCD panels have some smarts (advanced controller, which takes PWM duty-cycle as input and converts it to appropriate PWM or maybe even DC voltage), or dumb transistor, which basically takes signal from motherboard and switches the LEDs. Or it depends on brand?

Idk, I have noticed, that Chi-Mei/Samsung panels make me sick, while LG and AUO panels are OK. So, changing panel to LG will fix the problem, right?

>> No.1619457

it is

>Do LCD panels have some smarts
yes, some panel models have content-adaptive backlight control and provide a second PWM to the backlight to multiply with the backlight intensity signal from the system controller (to save backlight power on dark images). maybe you can disable this feature in BIOS or drivers
>changing panel to LG
if you can find a functional replacement that fits, sure

>> No.1619459

>maybe you can disable this feature in BIOS or drivers
Hm... Lemme check real quick.
>if you can find a functional replacement that fits, sure
Afaik I have eDP 30 pin connector. Can I also increase resolution from 1366x768 to FullHD? I have seen my laptop being sold with better panel, so I suppose that it would work.

>> No.1619463
File: 9 KB, 400x400, tegaki.png [View same] [iqdb] [saucenao] [google] [report]

What's blinking? The monitor backlight? Chances are it's just PWM, it may be possible to change in BIOS. You can also put a decently-sized inductor in series with the PWM wire going to the backlight to turn it into a more linear constant-current, along with a "freewheel" diode in there to cut off the nasties, perhaps a capacitor too. Pic related. Choosing the inductor is a matter of checking how much power the backlight uses and at what current, calculating the energy used by the backlight in a single PWM cycle at maximum, then use that current and energy value to calculate how large an inductor would have to be to store that. With a large enough capacitor there you can probably go a fair bit below that value, but you might also get ringing.
Feel like soldering into your laptop?

>> No.1619477

the backlight controller may already be regulating current because jesus these aren't lolduino kiddies designing these things
thus it is very likely that there is already a constant-current boost converter with PWM chip and inductor supplying the backlight, and I doubt it should take kindly to unwitting fuckery

>> No.1619478

Okay, I didn't find anything in BIOS.
>What's blinking? The monitor backlight
Yes. Frequency is too low. Move head slightly and boom, double image.
>it may be possible to change in BIOS.
I have doubts in this. All LCD datasheets say that PWM input frequency should be somewhere around 200-500 Hz in case of ChiMei screen (they also stated why, waterfall effect), and some panels (AUO) state like 200-20kHz. LG panel states 200-1kHz.
Considering the fact that you can change LCD panels without any issues, I think that different screens have different driving circuits. Some have something stupidly simple, like ChiMei, and probably more complex, which have PWM in kHz range. Or maybe even DC-DC.
>you can also put a decently-sized inductor in series with the PWM wire going to the backlight to turn it into a more linear constant-current, along with a "freewheel" diode in there to cut off the nasties, perhaps a capacitor too. Pic related.
This will not fit inside the lid.
>Feel like soldering into your laptop?
Not really, I don't want to blow up a relatively new laptop.

In any case, ChiMei screen has horrible colors. Like absolutely disgusting. And 1366x768 meme in 2019 is kinda not cool.
So I will replace the screen at some point anyway. Probably TN, but LG with better colors.

>> No.1619483

And yeah, on CCFL monitors I didn't have such issues. As far as I know, they have switching in 20+ kHz range due to inverters and shit.
High voltage is generated on motherboard. LCD is just switching it somehow, or converting (not sure) based on PWM input.
Adding something there will definitely create mustard gas.

>> No.1619486

>>1619463 I have a somewhat similar issue with 24v 6A cc/cv meanwell psu and chink dimmer. There is no noticeable flicker but a high pitched noise when dimmed anywhere between 20 to 80 percent. Will LC filter fix that? How big inductance and capacitance should be?

>> No.1619504

don't try to use a dimmer on a constant current supply. you can mod some Mean Well drivers to accept an external signal for dimming, which apparently means finding the dimming input pads and connecting a pot to them
also draw wiring diagram

>> No.1619516

>You always use a base resistor, right?
I just let PWM do its thing, this was probably my problem.
>"If I switch fast enough at the right duty cycle, nothing much will get through!"
Yes I am that dumb.

>> No.1619518
File: 135 KB, 2357x1221, Screenshot 2019-05-29 at 07.50.40.png [View same] [iqdb] [saucenao] [google] [report]

>don't try to use a dimmer on a constant current supply
Yep, CC psu will loose it's feedback without the load but my psu is CC/CV so it should be fine. I have done it few times on a different projects with no issues.
>don't try to use a dimmer on a constant current supply.
I need wireless dimming here.

>> No.1619519

>CCFL monitors
Those things are bulky and fragile, there's a reason we don't put those in modern laptops.

You'd be better off modding a potentiometer in there in place of the current set resistor that I'm assuming it has, so you can simply change the output current. Assuming you're running it on cc mode. And assuming it doesn't have a dimmer-input pin like the other anon said.

That's bad practice in general, for the logic circuitry as well as for the transistor and load. You could potentially fry an MCU solely from sinking/sourcing too much current to a transistor base. I'd always ensure I have resistors on the base, and something to limit current to the collector (unless it's switching an inductor for a short enough time.

Well it would be best if your dimming frequency was significantly less than that of the PSU's frequency. Anyhow, have you thought about using a digital potentiometer controlled by an ESP or something to vary the current directly?

>> No.1619525
File: 150 KB, 1560x848, Screen Shot 2019-05-29 at 3.05.20 pm.png [View same] [iqdb] [saucenao] [google] [report]

>unless it's switching an inductor for a short enough time.
This was the idea, but I guess I've been pushing my luck too far.
It's only going to a voltmeter, I thought that'd be acting as something of a limiting factor, but I guess not.
I have no fucking clue how whoever built this got it to work, but somehow, it just does.
Meanwhile I gotta build 10 more of the things, and shit's just burning out transistors left and right..

>> No.1619528
File: 352 KB, 1908x1921, Screenshot 2019-05-29 at 09.17.35.png [View same] [iqdb] [saucenao] [google] [report]

>have you thought about using a digital potentiometer controlled by an ESP or something to vary the current directly?
It's an interesting option but i'm looking for a simpler solution. What about LC filter between psu and dimmer?

>> No.1619533

some controller ICs accept PWM and/or analog dimming. a designer could modulate the high logic level of the PWM to perform that multiplication mentioned earlier

sorry, you'll need drivers that dim

>> No.1619542

>MOSFET’s are voltage controlled while Transistors are current controlled
>the fucking things had a MOSFET the whole time

>> No.1619545

Maybe this should go on /qtddtot/ but anyway, how much ripple can I expect from a cheap phone brick style chink "regulated" 12v 2A psu? I don't own a scope to test it, some ripple is fine but it should absolutely not go above 13.8v. Was thinking of buying a Astron linear PSU but that may be overkill for my needs.

>> No.1619546

put a capacitor and speaker on it and have a listen
or, also put a diode and a film capacitor on it as a low-tier peak detector. monitor with high-impedance voltmeter

>> No.1619548
File: 9 KB, 230x225, Multimeter.jpg [View same] [iqdb] [saucenao] [google] [report]

>high-impedance voltmeter
Impedance is irrelevant, resistance is enough.

>> No.1619550

Uh yeah I was wondering about your use of a FET there, since you can't burn that from gate current. But you almost certainly couldn't burn it from just switching a panel meter on and off either, unless you somehow managed to sustain it in the linear region and the power meter takes more current than I'm assuming. FETs are somewhat susceptible to ESD however, is it a 2N7000 or other old FET (old ones lack as much ESD protection)? Are you sure it's in the right orientation and right polarity? Because either of those could easily cause your FETs to short something, though not likely in this particular circuit with a low-current load on it.

You mean like I suggested in >>1619463? The CC driver would need to think that it's driving a constant current (assuming that's an issue) and you'd need to split up this current, so perhaps using a much smaller output filtration cap on the PSU, PWMing a MOSFET that shorts the PSU output to ground, and having an LC filter after this might work. But you said that it was running fine with the CV kicking in whenever you turned the channel off, and the only issue was whining, so that's besides the point.

An LC filter after the normal PWM dimmer won't do any harm, but it's doubtful whether it would actually cut this whine. A backwards LC filter between the PSU and the dimmer might do something as far as stopping the PSU's output current from dropping.

I'm looking at the datasheet now, and can conclude that your best bet would be wirelessly controlling/replacing the cc adjustment potentiometer, or buying the ELG-150-24B or 24AB or 24BE. It seems those ones can have either a DC analog voltage, a PWM signal, or a variable resistor applied to the dim pins in order to change the output current. The 24D2 has a cool sounding "built-in smart timer dimming and programmable function".

Looks like the Arlight uses wifi, so hooking up your own IOT circuit with an ESP8266 and a digital pot would be the cheapest method.

>> No.1619556

I was using transistors, when the design called for FETs.
This is what I get for just looking at a thing, going “that looks like a transistor, lemme grab a 2n2222” and installing it without thinking.
The original had a VN4210a, so totally different circuit, and I am a total brainlet.

>> No.1619563

Ah, yes MOSFETs don't necessarily require a resistor at their gate because they are a form of IGFET; insulated gate field effect transistor. Emphasis on the "insulated gate" part, meaning no current can flow from the gate to the source or drain. A BJT (bipolar junction transistor, i.e. normal NPN or PNP) is as you noted current-mode, and requires a current flowing through the base to the emitter to operate. This base-emitter junction acts like a diode, so by feeding 5V directly to the base of a BJT without any resistor you're basically shorting the 5V through a diode to ground.
So it's important to make a distinction between MOSFETs and BJTs, as both fall under the category of transistors. Also watch out fo JFETs (these do not have an insulated gate, the gate is usually isolated from the current path via a reverse-biased PN junction) and IGBTs (insulated gate bipolar transistor, they switch like a MOSFET but with a BJT-style collector-emitter junction, used for high power). There's a few other oddballs out there (like whatever EEPROM and flash use), but they're pretty rare in common use.

>> No.1619565

Ah, to me Transistor== BJT while transistor=/=MOSFET, from now on I’ll be more specific.
I’ll make sure I avoid JFETs and IGBT’s.

>> No.1619568

JFETs are too specialised (and expensive) to run into normally, but it's somewhat possible that you run into IGBTs when looking for HV MOSFETs. I know I did when designing my geiger counter, and I was looking through my electronics scrap store for their highest voltage FETs. The voltage rating on the IGBTs were fine, but the frequency I was trying to switch them at was apparently too fast, and that thing never got above 50-100V or so. Ordering some IRF840s from ali turned out to be the right decision.

>> No.1619587

>my geiger counter
Does it work? I'd really like to see the circuit diagram.

>> No.1619592
File: 8 KB, 400x400, tegaki.png [View same] [iqdb] [saucenao] [google] [report]

>does it work
Well, the boost converter worked fine until I shorted the cap on one too many screwdriver and it stopped working. This was while the geiger tube was connected (via a large resistor), and I don't think the tube (SBM-20) will have been hurt, but for now it's just sitting half-complete on a breadboard. The actual counting circuit is a bit ambiguous, but last I remember I had one side of the tube grounded, the other side tied to the boost converter output via a 10M resistor, with a high-pass filter coming off the top of the tube. Pic related. The LEDs are to clip any voltage higher than 5V, because zeners are for nerds. Might need them going to the 5V rail too. I know during startup the high-pass will see a high-amplitude voltage spike, which I'm hoping to eliminate. I might feed the signal through one of my opto-isolators after the trigger if it looks like a good idea.
The output of the filter would either go to an amplifier then to a schmitt trigger, or straight to a schmitt trigger, feeding whatever MCU or FET+piezo or whatever I wanted. Probably an ATTiny85 with a small display on it, but arguably I could use a specialised frequency-counter+display driver chip. Like the one that's inside my function generator I guess, and since it has an int/ext switch for the counter, I guess I'll use it for the testing. Now that I have a decent scope (400Vpk max) I can likely get back on this project.

The boost converter itself was a comparator oscillator with the second comparator in the IC being used as feedback to shut off the oscillator when the voltage had been reached.

>> No.1619598
File: 8 KB, 400x400, tegaki.png [View same] [iqdb] [saucenao] [google] [report]

you're using the standard GM tube filter that i've seen in at least two places, but i wonder if a bjt would work better

>> No.1619609

>i wonder if a bjt would work better
Yeah I guess so, perhaps even a darlington. All the designs you can find are split into high-side or low-side sensing. Some low-side circuits use a smaller value resistor at the bottom side (such that if the tube shorts the voltage at it still won't be large as the rails) to act as a sort of current shunt, but the signal to noise ratio is pretty bad this way.
Just using a straight transistor like that will be somewhat susceptible to boost-converter noise and will be harder to tune, but the simplicity of it might make it worthwhile pursuing. I've certainly noticed that these ebay geiger counter circuit kits are more expensive than they should be.

Anyone know a good frequency counter and display driver chip? I'm looking at the ICM7216. But I'm not sure whether it would be any good at geiger counting, and they're not exactly the cheapest ICs. I can see why most of the frequency counter kits I was browsing through used PICs. Though these pics aren't that cheap either. Without going for one of those one-time-programmable 2c MCUs, what's the cheapest MCUs people here go for?
Doing the boost converter feedback and even oscillation with the MCU might also serve to save on parts, but that would require constant use of its ADC. Still, it might decrease the ripple thanks to doing variable duty-cycle in software.

>> No.1619620

that is literally just two bjts superglued together

>> No.1619641

I'm looking for a portable soldering iron, not battery powered, still something that plugs in.

I have a great desktop soldering station, but I find myself in a lot of situations where I need to solder something together while away from the house. Is there a decent iron that is easily portable?

>> No.1619674
File: 41 KB, 900x675, 230-V-25-50-W.jpg [View same] [iqdb] [saucenao] [google] [report]

My on-the-road iron: 25/50 W (red button) 6€

>> No.1619676

I hate breadboarding shit.

Designing the circuit in a simulator is so comfy and neat, but then you go ahead and replace it in the real world with wires and components and it's just so gross and messy.

>> No.1619678
File: 63 KB, 800x800, bla.jpg [View same] [iqdb] [saucenao] [google] [report]

10 bucks buys you one with digital temperature control these days

>> No.1619682
File: 390 KB, 2080x1560, IMG_20190529_180643.jpg [View same] [iqdb] [saucenao] [google] [report]

>>1619550 Thanks anon.
>>1619641 Here is mine chink T12 handle with tiny analog controller. Works great with 18v bosch battery and 19v laptop charger.

>> No.1619694


those guys are so funny, aren't they? i am laughing so hard right now!

anyway on a more serious note, get TS100

>> No.1619736

It should be simple and robust.
A tool, not a toy.

>> No.1619770

Is there a youtube channel that makes and explains projects using microcontrollers that aren't arduinos?

>> No.1619776

People who don't use Arduinos can read and research shit on their own - so they are less likely to watch youtube videos about it.

>> No.1619788

Looks like something right up my road. I'll dig through amazon and see what I can find, thanks!

Looks great and I would, but I only need it for simple stuff like reattaching a wire or really simple stuff. No need to drop $60+ for that. I just didn't want a radio-shack style fire starter.

>> No.1619791

You just need to look for specific microcontrollers that aren't under arduino support.
Here is a set of videos for using the PIC 8-bit line:

>> No.1619797

that's cool thanks

>> No.1619801

So I decided to mess with oscilators. I've made pic related and noticed two things:
>I've used no capacitors and manhattan style building, the system oscillates arround 15Mhz. If there are no capacitors, where does the stray capacitance comes from? The BE junction?
>the output inverter produces a sinewave, not a square wave. Why is that? Too little gain?

>> No.1619807

>I've made pic related
hmm, I see...

>> No.1619820
File: 69 KB, 800x551, T2h0L8Xt0aXXXXXXXX-549039712.jpg [View same] [iqdb] [saucenao] [google] [report]

>>1619788 It looks like this kit is becoming obsolete but you still could find one.

>> No.1619826

What sort of technology car keys use? they are tiny, have huge range and battery in them lasts 20 years even if you use them every day

>> No.1619848
File: 23 KB, 685x464, Untitled.png [View same] [iqdb] [saucenao] [google] [report]


>> No.1619864

>I've used no capacitors

Those symbols with 100pF by them are capacitors...

>> No.1619886

In the real life circuit. Without capacitors the simulator will go bananas.

>> No.1619911

God this schematic is shit. Is this supposed to be a buffered phase shift oscillator? I can't tell because everything is a haphazard mess and the feedback appears to be in the wrong place.

>> No.1619918

>simulator will go bananas
Failstad, nuff said.
Look up 'ring oscillator'.

>> No.1619927

Most of the capacitance is in the B-E junction, there'll probably also be a few pF in the wiring. The transistor's datasheet will tell you the actual capacitance values for the transistor.

>> No.1619932
File: 84 KB, 1873x721, miller killers.png [View same] [iqdb] [saucenao] [google] [report]

You know there are ways to mitigate the miller capacitance right?

>> No.1619933

the real world is gross and messy, yes

a. look up the Miller effect
b. not enough gain in the buffer, try a long-tailed pair

why don't you crack one open and find out?
or google

>if I don't tell the simulator about my capacitance, it should Just Work

>> No.1619939

>it should Just Work
it should use its model

>> No.1619940
File: 149 KB, 823x168, 3RC-s.png [View same] [iqdb] [saucenao] [google] [report]

better version

>> No.1619972

That's an odd circuit for reading the temperature, just a low-pass filter and a voltage divider fed into an op-amp without any feedback. Or at least, no direct negative feedback from the op-amp's output. It feeds a schmitt trigger which feeds the switching transistor, which sort of gives feedback by changing the iron's temperature.
But that first link you posted is not that, its 8-pin IC is a single MOSFET. No clue if they have any temperature feedback at all.
I'd go for one of the cheap digital T12 OLED modules without a case instead, but surprisingly they're like $12 each. There are also similar 7-segment boards for only $8, which are probably a better deal.

>> No.1619986

Yes I know. I just want to know from WHERE does the capacitance comes, seeing the system needs a delay to work.
>simple java script simulator that doesn't even have cursors needs to simulate semi conductor physics
what. I don't know much about electronics but I know about modelling and systems. You should strike a balance where you get the job done with just enough complexity to make it work. Going crazy on details on your model without a corresponding gain in precision, computation or information you get from it is stupid. You don't need to simulate the non-linear, dynamic gate source capacitance in falstad, you just put a 1nf cap there to get the job done.

>> No.1619992

Ring oscilator. Yeah it's a time delay/phase shift wibbly wobbler

>> No.1620014

>You don't need to simulate the non-linear, dynamic gate source capacitance in falstad
this. Ebers-Moll can lick my Gummel-Poon

>> No.1620045
File: 11 KB, 649x304, Screenshot_2019-05-30_03-19-57.png [View same] [iqdb] [saucenao] [google] [report]

I have those drl lights for daytime driving whose schematic (drawn from the pcb layout, I left the labels as they were) I have in picrel. I'm pretty sure I have drawn it correctly. The problem is I'm confused by that upper right transistor. Why is it even there?

>> No.1620048

okay, how are these connected to the rest of the electrical system?

>> No.1620050

>Those things are bulky and fragile, there's a reason we don't put those in modern laptops.
They weren't big. They were just power hungry.

Btw, can I replace CCFL with LEDs. I found LCD monitor in trash, but 45W for 20 inches aint impressive.

>> No.1620052

Unfortunately I can't check that now, but if they were connected straight to a car battery it wouldn't make any sense
Those lights come in pairs, left and right. I have an idea that maybe they are connected in series, so the gnd of the first one is supply of the second one, hence those strange labels, but I can't really check right now.

>> No.1620067

It's quite possible that the transistors you drew have a different pinout. The only other reason you'd have a transistor shorting the power rails of LEDs is if they're shorting the output to a constant-current supply, since shorting it will cause the voltage to drop a bunch and the power to drop to near zero.
But A: those LEDs already have current limiting resistors, and B: you said that they're directly across the battery.
So I can only assume that the transistor is not in the pictured orientation. It could also be a different 3-pin component, like a linear regulator (which could arguably be true if they're using it as a current sink).
Though it might be more likely that you misread the traces.

>> No.1620070

>Though it might be more likely that you misread the traces.
That's always a possibility, but I checked them a few times even with a backlight just to make sure.
Both of those 3-pin smd elements had the same labels - J3Y. And google says that's a npn. But I guess it might be possible for it to be something else, or even both of them to be different?

>> No.1620073

that's a mechanical problem first. happy soldering

that sounds right, where the top Q steals current from the top LEDs and passes it to some LEDs or other load on GND1 when CTR is low, and passes all current through both sets of LEDs when CTR is high. what it looks like depends on the characteristics of the load on GND1 (or +12VGND2?)

>> No.1620075

It's a fairly generic marker that could probably refer to multiple different transistors, which might have different pinouts. I'd turn it on and probe about with a DMM, while looking for 0.6V diode/transistor-junction drops, with the CTR both on and off.

>> No.1620113

possible but highly unlikely. it's probably just an S8050 with the typical pinout for SOT-23 discrete npn (clockwise from lower left: base emitter collector)

>> No.1620142
File: 2.89 MB, 4656x3492, IMG_20190510_210144.jpg [View same] [iqdb] [saucenao] [google] [report]

I'm building a 'house' battery box for the back of my truck.
Plan is to wire the house battery in parallel with the starter battery, with a voltage sensing relay inline on the hot side so that the battery charges off my alternator, then disconnects when my truck is off.

I have two questions:
>What type of quick disconnect can I put on the side of my battery box that will handle the amperage of my alternator?
>If I have some 12v accessories connected to my house battery, say a laptop charger, while my alternator is running will that fry anything?

>> No.1620151

>asks for connectors to handle a rating
>doesn't give us the rating
you have failed. anyway look into forklift battery connectors, lots of sizes and options
>will that fry anything
does it fry when your truck is recharging its own battery? the automotive electrical environment is pretty hostile all by itself and devices are built to withstand it. an extra battery box being switched in while the engine is running is about like the aircon compressor clutch kicking in, that is, not much

>> No.1620154

What sort of resistor do i need if i need really precise resistance? say i need a range of 1 - 50ohm and need to be able to reliably set something like 30.5 ohm

>> No.1620155

Does anyone play with hall effect sensors a lot. Im looking at creating a type of position sensor. The catch is theres some rotations involved. Does the sensors output relate roughly to


>> No.1620162

>>asks for connectors to handle a rating
>>doesn't give us the rating
My current alternator is 65A but I might upgrade to something larger if charging 2-3 batteries kills the current one, like 120A.
At I guess 12v?

I don't know how alternators work pls no bully ;_;

>> No.1620165

>look into forklift battery connectors
Oh I was actually already looking at these but hoping to find one that was like, a waterproof socket, so I could mount it out of the side of the battery box

>> No.1620166

Why not make a test rig? For a simple case where the magnet is constantly pointing at the hall sensor through all angles and radii, you're probably right, though I think it may be etc/r^3 instead since it's not a monopole.
But if the magnet itself rotates to point elsewhere, you'd have to take into account the spacial flux density equation of the magnet itself. You may be able to approximate it to be a simple dipole if you're dealing with long distances, but since hall-sensors aren't exactly good to those distances I imagine you'll have the sensor no more than 5 magnet radii away, meaning you may have to all that nasty math.
A 3-axis magnetometer may help your calibration process if you are dealing with this extra degree of freedom.

>> No.1620172

>That's an odd circuit for reading the temperature
It's simple af. First opamp is a comparator, second schmitt trigger next bjt driver and P-fet. No special ICs.
>No clue if they have any temperature feedback at all.
Opamp is cheap and there is no cold junction compensation but it still regulates the temperature. ±10 degrees is not a big deal.
>cheap digital T12 OLED modules
Are they using proper compensated thermocouple amplifier ICs?

>> No.1620173
File: 190 KB, 396x280, filename.png [View same] [iqdb] [saucenao] [google] [report]

>Opamp is cheap
But there isn't an op-amp on the linked PCB (616dev). It just has some passives, a 7805, an 8-pin MOSFET IC, and a SOT package of some variety. Pic related.
>Are they using proper compensated thermocouple amplifier ICs
Pretty sure. I know my T12 station has a very similar module inside and it has an ambient temperature readout, which can be either from the thermistor in the iron's handle, or from the sensor on the PCB.

>> No.1620175
File: 93 KB, 750x439, TB2sFWFaFXXXXXkXXXXXXXXXXXX-549039712.jpg [View same] [iqdb] [saucenao] [google] [report]

>>1620173 U sure?

>> No.1620178
File: 259 KB, 750x750, TB2BkWdaz3nyKJjSZFHXXaTCpXa_!!204161235.jpg [View same] [iqdb] [saucenao] [google] [report]

Shit never mind, looks like it has more parts on the back. They look pretty neat, I might be tempted to get one.

>> No.1620181

I was torn between 1/r^2 and 1/r^3 but according to https://sensing.honeywell.com/hallbook.pdf its inverse square if its head on for unipolar sensors. So basically if the sensor only reacts to north pole its treated as a monopole. But I havent seen an equivalent graph/model for a linear sensor.

>> No.1620185

At long distances I'm pretty sure it will approximate to /r^3, but that's an artefact of the magnet's field, not of anything about the sensor.
So what orientation is your magnet going to be? Head on?

>> No.1620193

It would be hard to describe without an image. But it rotates while moving closer. Which is why I included the cos(theta) part. It never rotates more than 90 degrees though. Theres some particular constraints so Im going to have a very particular gauss curve. Id just rather have an equation to use as my guide instead of plotting and finding a best fit gauss because the former works better for electronics in idea cases.
Though I suppose it wouldnt be too hard to create a test and move it back and forth just to make sure it obeys the rules I think it does.

>> No.1620203
File: 9 KB, 400x400, tegaki.png [View same] [iqdb] [saucenao] [google] [report]

You'll have to draw some sort of diagram because I'm getting confused which rotation you're referring to. The cos(theta) component would be for what direction the hall sensor is pointing to if the magnetic field is coming in a straight line at it from the magnet. The more complicated equation f(phi) would come into play if the magnet itself were not pointing at the hall sensor constantly, picrel.

>> No.1620205

fixed resistors? they're sold by tolerance. there are 25-turn pots but I can't vouch for their repeatability or linearity
what's your project?

if there's a lid, you could make the quick-connection inside the box. easier to weatherize one side of the lid-box joint with a flexible seal than the whole thing

>> No.1620243

I'm new to eagle and I can't figure out how to make a board that has metric measurements.
I tried switching units but there's no way I can shape the board exactly how I want to because there's no dimension line...

>> No.1620280
File: 53 KB, 283x932, Annotation 2019-05-30 211420.jpg [View same] [iqdb] [saucenao] [google] [report]

Use KiCad instead

>> No.1620364

Are RF oscilators just black magic? No matter how sound is the math, the inductor I spin always sucks ass, the ones that even oscilate are more distorted than who knows what.
Is there a idiot-baby proof topology for me to get my sea legs? I'm aiming for 50Mhz but at this point anything that works will make me happy

pls help

>> No.1620397

My favorite is the emitter-coupled oscillator that has no frequency-dependent components apart from the parallel LC resonator, just two transistors and one resistor to adjust the loop gain. It always works, needs only 1.5V and is self-limiting to about 1Vpp. With a reasonable Q inductor it produces 1MHz on a few microamps.

>> No.1620398

Okay, update.
I found broken (literally) LCD in recycler, and I looked on PCB. It was from AUO and it had inductor and couple capacitors suspiciously close to back-light LEDs.
So, enginiggers in samsung/chi-mei are fucking assholes, for not putting proper driver. 1 fucking cent on inductor and capacitor. Pathetic.
I guess I can increase PWM frequency by adding simple 6-pin micro and some pajeet code,which will take 200 Hz PWM and convert it to 1 kHz PWM.

>> No.1620457
File: 3.37 MB, 3120x4160, IMG_20190530_215022163~2.jpg [View same] [iqdb] [saucenao] [google] [report]

Could someone please explain to me the purpose of the 100k and 470 resistors at the top of this circuit?
It's a section of a house alarm circuit from make:electronics.
I think the 470 might be there to prevent a short from the 555 output to ground, but what's the purpose of the 100k? Is it just there to increase efficiency by limiting the current?

>> No.1620469

The output of T1 is not connected to the emitter resistor of Q2 but to its base resistor.

>> No.1620475

Yeah I realized that. So its more like
But theres only 1 degree of freedom so Im still playing with that part of the equation.

>> No.1620508

I like 900M series tips. I don't like hot-swappable T12-like thing.

>> No.1620515

But the tip will be in much worse thermal contact with the element and thermocouple, since the tip is separate. You get better performance by having the tip permanently bonded to the element and thermocouple, such as in T12 and that one 3.5mm weller tip that the TS80 uses.
Though I'd personally lean further towards a mini T12 station instead of a TS100; they sell variants without a mains power supply in them which would work off a battery for portable use.

>> No.1620519

The 100kΩ is for current limiting so Q2's base-emitter junction doesn't fry. The current is limited to around about 50µA through the base, which corresponds to a maximum current of about 5mA through the collector.
The 470Ω is more unconventional, as it forces some linearity on Q2's operation. Q1 is also operating in common-collector mode, which looks fucking bonkers, so I conclude that the author of the circuit just doesn't know what they're doing.

>> No.1620618

How do I read a SIM card without a reader?

>> No.1620621

you google
/ohm/ RULE 1

>> No.1620632
File: 72 KB, 500x410, 1550357969612.png [View same] [iqdb] [saucenao] [google] [report]

Is investing $300+ on a J-Link worth it?

>> No.1620652

yes, i whipped one up after painstakingly working out a xtal bjt ckt and it was a great 20MHz oscillator that took a few minutes

>> No.1620682
File: 37 KB, 492x565, eco-test.jpg [View same] [iqdb] [saucenao] [google] [report]

The ECO also works well at the low frequency end. I tested an inductor that my LCR meter claimed to have 5H. The oscillator experiment arrived at 6H. Supply 1.5V, scope 1:10 probe directly across the LC circuit.

>> No.1620686

When you can get fake ones for $10 nah.

>> No.1620691

that's a big inductor

>> No.1620693

it would be more exciting if you had a ±2% or better inductor as a performance standard

>> No.1620702

for you

>> No.1620704

It's the HV winding of a ferrite core transformer used e.g. for photo flash or GM tubes.

>> No.1620707

Ah, so really fine wire. I made my GM-tube boost converter using a 470u without any issue. Does a photo-flash circuit run closer to a boost converter, or a car ignition coil? i.e. is it pulsed multiple times or just once per use?

>> No.1620713
File: 420 KB, 1080x1920, Screenshot_20190531-170301.png [View same] [iqdb] [saucenao] [google] [report]

What does this meme mean?

>> No.1620716

that electrical engineers still have a more fulfilling love life than you

>> No.1620725
File: 12 KB, 707x393, kodakflashm.gif [View same] [iqdb] [saucenao] [google] [report]

>boost converter or a car ignition coil?
left side converter, right side ignition coil

>> No.1620739

Huh, that's interesting.

>> No.1620763

I need to measure some voltages on a pcb conformally coated with 1B15H. Will this coating throw off my readings?

>> No.1620772
File: 50 KB, 1000x349, Clipboard01.jpg [View same] [iqdb] [saucenao] [google] [report]

Every example of constant current lm317 i found uses the left circuit. It works fine but it sucks if you need a big current since the virgin resistboy can't handle that and you need to buy some expensive chad resistman instead to dissipate the several watts you need
But now i found the example on right which is also constant current but the load current doesn't pass through the resistor.
Is that bullshit or is it legit? And if it is legit why do all examples show the shit way to do it?

>> No.1620819

$300 isn't bad, but if you are using it for your own personal use (nothing for a company or to make money) you can get an "EDU" version for a fraction of the cost.
You could get a clone, but I've found that clone debuggers are not very good.

>> No.1620828
File: 50 KB, 257x238, 317-pins.png [View same] [iqdb] [saucenao] [google] [report]

a floating regulator is different
datasheet excerpt at >>1619127

>> No.1620832

>load current doesn't pass through the resistor
Of course it does.
Why do you personalise components?

>> No.1620833
File: 56 KB, 684x405, LM317_Pinout.png [View same] [iqdb] [saucenao] [google] [report]

> But now i found the example on right which is also constant current but the load current doesn't pass through the resistor.
Wrong. Centre pin is the output, left-hand pin is the reference. The circuit is just drawn weird. Also: you don't want a filter capacitor on a constant-current supply; that defeats the purpose.

If you need an efficient constant current supply, use a switching regulator.

>> No.1620834

fuck you are right
i am used to the adj pin being in the middle in schematics

>> No.1620853
File: 186 KB, 620x1000, PCB.jpg [View same] [iqdb] [saucenao] [google] [report]

So I have the circuit off a keyboard and I want to wire my own buttons to it (it's for an arcade top).
I tried scraping the carbon off and soldering directly to the board before, but that was messy.
Now I made this pcb to give myself more room to solder, and my hope was to simply screw it to the keyboard circuit and that the contact would make a solid electrical connection, but testing revealed it doesn't work very well.
Any suggestions? I was thinking of depositing little trails of solder on the keyboard end to see if they conduct better into the carbon.

>> No.1620863

conductive adhesive?

>> No.1620865

hmm. thanks for tip, didn't even know that was thing. It would be pretty tough though, the trails are 0.75mm..

>> No.1620887
File: 75 KB, 640x480, soldering skillz to the max.jpg [View same] [iqdb] [saucenao] [google] [report]

> tried scraping the carbon off and soldering directly to the board before

old keyboards, some of them at least, used copper pads instead of carbon.

if you cant find one of those, scrape off the green stuff (solder resist) from the traces and solder 30-AWG wire wrap wire to it.

>> No.1621172

Say I want to make an antenna and receiver circuit with which to listen to schumann resonances and ELF in general. I've heard that 50Hz mains will likely be a significant pain, and that most people filter this out in software. But that means they're limited by the maximum amplitude they can input to their sound card via the signal to noise ratio, so I was wondering if anyone here knows a good high-order band-stop filter design? I've studied a bit of filters before, learning the simple butterworth polynomials and such, but as far as higher-order filters we stuck to stacking op-amp filters onto one another. Is it worth trying to make a band-stop LC filter or other high-Q topology so don't need as high of a filter order?

Ideally I'd be able to plug the receiver into the wall and use the power lines as my antenna, but I'll probably just go for a short-loop antenna of some variety.

>> No.1621205

Ok, turns out searching "band stop filter" gives you next to nothing, while "notch filter" gives you a whole bunch.

>> No.1621223
File: 347 KB, 1536x2048, IMG_20190601_110232.jpg [View same] [iqdb] [saucenao] [google] [report]

I ordered this optical sensor from chinks

But pic related is what arrived. At first i thought it just has different casing because #chinkthings but the problem has that this sensor only has 3 legs and the one i wanted has 4 legs.
So of what do i have a pile of here then?

>> No.1621229

what you have there is typical of a consumer IR receiver. you got chinked, my dude

>> No.1621230


>> No.1621241

also i order a bunch of ir leds and they seem to be all fucking blown.

I connected them with a 100 ohm resistor to 5V in series with normal led, so if the normal led turns on it means the ir led is working, and none of them fucking do work (the normal led is working fine tho i checked that)
fucking chinks i swear to god

>> No.1621245

check their Vf with your trusty diode checker, what do you get?

>> No.1621249

Nothing, because their voltage drop is too high so the multimeter just shows 0 when i set it into the diode checking mode.
I tried it it on a small normal LED where is shows about 1.8V but on my brighter LED that is about 3V it shows 0L and so it does on the IR leds, so i can't really use it to check them?

I checked 10 IR leds so far by trying them in paralel with the normal LED to see if the normal led turns on to indicate the IR leds are conducting and nothing.
I also tried looking at them with my phone camera to see if there is something and nothing.
I can't believe ping pong sent me 20 burned IR leds. Am i just checking them wrong or something?

>> No.1621252

Also what is interesting is that the one guy rating the diodes posted photos where they work fine, by they have transparent casings.
I got some black version, and there is another guy int he ratings who gave zero starts saying none of his diodes worked and he got the black ones as well.
Could the key be in the black paint on them that makes the phone camera not see them? but that doesn't explain why the normal led in series with the ir led doesn't turn on. So did i legit got send 20 busted diodes?

>> No.1621255

Holy shit are you fucking kidding me?
I went to buy another batch of diodes and saw this
Would that mean that the black ones are not capable of emitting IR light and are for recieving IR light only?
I just tried pointing TV remote at the black IR LEDthat is in paralel with the normal LED and the normal LED started to very faintly blink

>> No.1621256

add L at the end of the url

>> No.1621257

IR LED Vf should be in the 1.5V range

yes, the visible-rejection packaging would more likely be on the receiver than the transmitter

>> No.1621258

That is just fucking brilliant.
So instead of 20 IR leds and 10 IR sender/receiver components i ended up with 30 IR reciever ONLY components and exactly ZERO fucking IR emittors.
And now i have to either order more IR leds and wait a month for them to arrive or pay the outrageous electric jew prices in local stores for them
right now i really want to go to my local chinese takeout place and just strangle some random chink

>> No.1621262

Black means daylight filter means receiver
Clear means emitter diode
Longer lead does not always mean anode
If OL, reverse leads when measuring
I found shorter lead = anode and Vf = 1.17V

>> No.1621270

don't be a goy, just order from digikey, mouser, avnet, rs or whatever

>> No.1621273

that better be a joke
for comparison for price of one arduino uno in the jew stores in can get from chinks 7 (SEVEN!!!) if the exactly same ones

>> No.1621279

>arduino uno
you don't buy toys for artists, just get an atmega 328 and upload your bootloader and lib (=arduino software)
or just get a real microcontroller

>> No.1621281

ordering a finished assembled dev board is a bit different from ordering components. that particular seller seems to suck, according to their feedback scores

it seems reasonable to think of "arduino" as "shitty bootloader + soldering this thing onto a board for me" anymore

>> No.1621282

I use many different ICs, like ESP32 or the viagra pill, but only those compatible with arduino IDE because it accelerates everything so incredibly much and it is so fun to develop with in this environment that i will never touch any other ide as long as the projects i make can be achieved with the arduino ide and so far 100% of them can be

>> No.1621284

Even then, i checked local pricing, and for 2.5 IR leds here i can get 20 from chinks. that is a massive difference
Basically at this point i order 99% of my shit from china and buy locally only when i need something right now, like some specific power resistor and such

>> No.1621285

In fact, chinks are so cheap, that when they send me the wrong thing i am only angry because i have to wait a month for the correct thing to arrive when i order again, i don't even bother with disputes.

>> No.1621288

>arduino IDE
true, you achieve actual working things very fast.
but the whole ecosystem is quite limited, don't be afraid to get into general embedded
the learning curve can be steep, but that is part of the fun!

>> No.1621290

I hear that, especially when the alternative entails having had the foresight to video the unboxing

>> No.1621291

this. I ordered a solder iron with temp ctrl for 20 but got a "rework kit" (solder iron + hot air thing) for 30.
can't complain.

>> No.1621309

Where did you manage to find TS100 for only $20?

>> No.1621311
File: 595 KB, 2666x1500, IMG_20190530_180735.jpg [View same] [iqdb] [saucenao] [google] [report]

Hello, I'm trying to fix my LCD computer screen, the CCFL lights broke so I bought a LED upgrade kit on eBay. I've watched several Youtube videos about the trick but the circuit board of my screen in obviously not perfectly similar as the examples. I need to find a power source for the LED between 10 ~ 30 Volts so I can solder that damn last wire but I don't have a multimeter. Is it possible for you to identify a potential suitable power source (10-30 V)?
pics related

>> No.1621312
File: 3.69 MB, 2736x4864, IMG_20190530_180724.jpg [View same] [iqdb] [saucenao] [google] [report]


>> No.1621313

chink version, 60W YIHUA 947-III, manufacturer shop:

>> No.1621316
File: 137 KB, 1048x497, IMG_20190530_180751.jpg [View same] [iqdb] [saucenao] [google] [report]

pic related is the LED control board

>> No.1621336

where can I buy a 1 kwh battery

>> No.1621338

target voltage?

>> No.1621339

enough to power a laptop or phone

>> No.1621344

right here fren, use an appropriate step down converter

>> No.1621358
File: 920 KB, 1071x853, sleepy iron.png [View same] [iqdb] [saucenao] [google] [report]

>> No.1621387
File: 32 KB, 552x293, sleephiber.jpg [View same] [iqdb] [saucenao] [google] [report]

Shaking to wake up is covered by a troll patent, removing from stand or other motion is not.

>> No.1621456
File: 116 KB, 650x488, 2670-00.jpg [View same] [iqdb] [saucenao] [google] [report]

What's the best way to plan out perfboard traces? Software? Just using a whiteboard?
I'm finding it hard to come up with a single layer design without using a bunch of jumpers.

>> No.1621516

waste of time.
breadboard proto -> design pcb -> order pcb

>> No.1621525

no, chink pcbs are so cheap it is 100% certain that child slaves are making them

>> No.1621533

I need to simulate a load, would getting a 1 ohm resistor and then using a pwm mosftet to vary a current through it be a good idea?
So say i have 10V and want to simulate 5A current, then i will flip flop the mosftet with a square wave that is 50% doody cycle

>> No.1621569
File: 53 KB, 593x652, demo tape fuzz.png [View same] [iqdb] [saucenao] [google] [report]

The volume output on this circuit is way too high. What's the best way to lower the output? Higher value pot? Solder a resistor to one of the pot lugs? Any advice appreciated.

Pic related

>> No.1621580

The preferred language of electronics is the circuit diagram.

>> No.1621581

So I made an LC oscillator some anon here suggested, (50Mhz emiter coupled). It works nicely. But now I have a question, how do I stop the oscillations from propagating up to the power supply and other circuits?

>> No.1621593

What's the name of the project? Sounds interesting

>> No.1621606

>What's the best way to lower the output?
Turn the volume control toward minimum.

>> No.1621608

L/C filters
ferrite beads and capacitors

>> No.1621609

and shielding

>> No.1621630
File: 955 KB, 1369x782, 1547497835242.png [View same] [iqdb] [saucenao] [google] [report]


>> No.1621639

>at 5V
Those things have a 1V voltage drop, so you'd be putting 4V across 100Ω which is 40mA. Sure they're rated at that current? It's only 40mW though, and other LEDs of similar packages are fine at 60mW, so you should be ok I guess.


begone shill

It might not have a 12 or 24V rail, since there aren't really any high-power things other than the backlight, which is probably boosted straight from rectified mains. Still, have a look around the mains input to see if there's anything resembling a second flyback buck converter. I'd consider getting a mains to 24V buck module, but since the LED board itself says "10-30V", it likely has its own switching power supply onboard, so bypassing that and directly feeding it the exact voltage it wants might be a more efficient course of action, if you're up to it.
>no multimeter
I'd remedy that as soon as possible, they're vital for troubleshooting.

>> No.1621759


>> No.1621835
File: 614 KB, 700x700, 1543836166781.png [View same] [iqdb] [saucenao] [google] [report]

The output of 74xx 40xx IC should limit current with resistor or it's not necessary?

Name (leave empty)
Comment (leave empty)
Password [?]Password used for file deletion.