/diy/ - Do It Yourself

i had a really good post but then i was banned by an /sp/ mod who did it for free.

Want to get a cheapo multimeter for some very beginner stuff, like $20 tier, what kind of stuff should I be looking for in one or are they all the same at that price point?
Also, I may be retarded, but what's the difference between a "continuity tester" and literally just an ohmmeter?

>>1546196
Can we get rid of some something like Youtube channels from the OP and add a list of recommended equipment? Or at least a link to a pastebin page with that info or something

>>1546213
definitely recommend aneng 8008, if you wanna stretch your budget a bit, should be able to find an oldie benchtop dmm from hp etc.
And there is no difference really, continuity beeps and 'latches' when you measure like <1 ohm. very useful for checking shorts

Welder anon here
The project is going well, I just charged the capacitor bank to about 4.5V and it threw some fat sparks, so at least errors from this part are filtered.
Now I'm stuck because I broke both 3mm drill bits I wanted to use to punch some holes in the MOSFET copper rails, and I need a way to isolate a top layer, which will hold the circuit, the MOSFETs and a connector for the arcade button/pedal I want to use to control the discharge.
Now a question:
When using the cap bank by itself I had problems by trying to solder nickel strips as them ended up stuck on the electrodes instead of the batteries I wanted to weld, so my question is: Is copper a good material for the electrodes or I need to use other materials such as aluminum or something that melts at higher temperatures?

looking for an rf project
Something that just came to mind is to collect some kind of data about a general public with rf. But I don't really know what. Any ideas?

Or any other rf projects for a student to do

>>1546216
is there enough space to do it justice?
could link to EEVblog forums which are fairly based, e.g. their comprehensive multimeter buying guide

>>1546196
>free samples to non-NEETs
GTFO NORMIE REEEEEEEEEE

>>1546220
>isolate a top layer
you could gouge/score some copper-clad to build the circuit surface-mount style, even with THT components. or you could use nylon washers/spacers/nuts. lots of ways to skin this cat
>copper
is best electrode material. you need to control the discharge so that electrodes don't overheat. a simple push button isn't enough, you might want to use a 555 monostable circuit like Pic related (extra stuff at left debounces the switch)

>>1546256
You want something highly digital? In which case you could monitor wifi or bluetooth traffic in a public space. For something more analogue, monitoring general RF noise levels across the spectrum in a few select areas could be an interesting one (plus one I'd personally find quite useful). In both cases I think it would be best to tackle them with an SDR. If you want one that's feasible to design your own circuit for, you could extend the idea of a simple radio demodulator to a number of possible uses: measuring something like an ultrasonic flow sensor but with RF, trying to listen to audio frequency noise produced by people's audio equipment trying to triangulate the positions of radio station transmitters with directional antennae, testing materials for RF absorption at a range of frequencies (something you could back up with very in-depth theoretical data), noise optimisation CW envelopes, data compression and error-correction algorithms, dish antenna optimisation etc.

A project I personally have put a fair bit of thought into is trying to triangulate the position of a radio beacon by measuring the time difference it takes for its transmissions to reach various networked receivers like how satnav works (hint, time-to-digital converters are neat). Another interesting thought is trying to replicate a CAT scan (fourier slice theorem method) but with radio of some sort, probably microwaves.
http://www.dspguide.com/ch25/5.htm
The above looks like a relevant source on this. Note that you'd need a decent method of moving your scanner precisely.

>>1546304
There should be a comma after "audio equipment". Also directional antennae aren't necessary for radio station triangulation, though doing so without it you'd be relying on signal strength differences (perhaps a bit unreliable thanks to terrain/buildings) or the ∆t method as described in the second paragraph.

And here's some info on the CW envelope:
http://www.seed-solutions.com/gregordy/Amateur%20Radio/Experimentation/CWShape.htm

>>1546196
Please explain brainlet what the fuck is on the picture.

>>1546319
crazy high frequency radio magic.

>>1546319
Analog front end of 110 GHz $1.5M scope.

>>1546380
you know it's serious business when there's no more BNC

>>1546380
How can the digital keep up with a 110GHz sample? Frequency mixing?

>>1546380
>110 GHz
Is that a lot?
My computer is only 3GHz

>>1546387
speculating, it doesn't, just has to keep up with the slow side of hella fast FIFO or dual ported RAM

Could you recommend me something like esp32 (which is basically arduino with wifi) that supports mobile data?
I need something i can program with arduino ide and is capable of acceing internet through a sim card. It doesn't need to have wifi.

>>1546395
http://lmgtfy.com/?q=arduino+sim+card+internet

>>1546403
if i wanted to buy the first piece of shit i found on google i would do just that
so thanks for nothing mr edgelord

>>1546405

protip: When you do a google search they allow you to look at all the items.

>>1546403

I always assumed that pile of shit actually returned the google search, but I finally clicked on it, and it's simply shit that a shitloard thinks is cute.

>>1546390
it's a 10 bit adc iirc so the bitrates are massive. it cheats by having multiple channels sampling out of phase with each other but the individual channels are still very fast.

>>1545887
I have been interested in old computers for some time. In theory, with thousands of times more RAM and storage capacity, you could interface them with a Z80 or a 6502. However some problems arise:
1. Most RAM and storage is designed for 3.3V or even 1.8V instead of the 5V logic these microprocessors use, so you would need to use level shifters to interface newer technology.
2. Nowadays almost all memory (be it SRAM or ROM) is serial instead of parallel addressed. The most dense flash memories have to be addressed by blocks of 512 bytes instead of doing so byte by byte. NOR Flash memories are a thing though. Parallel, 5V EEPROMs also exist yet, but they are very small and getting discontinued as we speak.
3. Addressing these chips might pose a challenge. Accessing memory and program locations in yesteryear, with 8-bit computers was done via several dirty tricks you learned by experience, and accessing more than several kilobytes was not feasible.

Thus, interfacing 80s game cartridges such as NES, MSX or Sega Genesis with current technology is a challenge, not because we haven't the technology to do so, but because such systems are getting discontinued and finding proper replacements is increasingly difficult. That's why it's a safer bet to find an FPGA, to synthesize these designs is easier than building with ICs.

>>1546266
Thanks for the 555 circuit, I figured that I might need some kind of analog monostable circuit able to spit 12V when I didn't have the drivers, but now that I received them before I expected, I prefer to go and use one of the many ATTiny85 I have laying around, so I can modify the pulse duration without rewiring the circuit.

>>1546380
That thing must have a shitton of RAM.

Just did a 555 squarewave generator. Actually makes squareish pulses up to 500+KHz, but the they're skewed. Thinking of adding a pot to the VCC-CV-GND, since that tweaked the skew in simulations. It should BE a squarewave. I have a resistor and pot (20k+100k) from output to the threshold/trigger capacitor, and the capacitor is hooked into a pin header socket so I can adjust the frequency beyond the resistance. I ran through designs in the Falstad simulator, but it has it's limitations, might need Multisim for a more reliable design.

I tried to do a pcb layout, but ki-cad is kinda shitty to use. Inkscape was easier. I went through a certain number of routes before I figured I had something good, only to realize I needed to jump a wire on the underside.

Plan to use it for brain stimulation (hence adjustable and more than one decade of frequency), but also maybe screwing about with anything that needs a square pulse. I have another socket, I have dual op-amp chips. 555 doesn't go negative, so I could do a +/- square pulse, and a +/- triangle.

>>1546580
>kicad is kinda shitty to use?
Are you serious? If you read the most basic of introductions on how to use it'd be piss easy, it's pretty intuitive, especially for PCB design. Inkscape might work for circuits like yours with only one chip and a handful of other components, but for anything slightly more complex, you need to use tell PCB design software.

>>1546575
Although memory is "addressed" in blocks, you still read and write on the byte level. The point of the design is a hardware change, that it's cheaper and faster to use larger units of allocation than small ones, and it's also usually more consistent with use behavior.

Level shifting is not a problem, literally every switched-mode power supply is effectively a single-polarity level shifter. It can even be done as simply as two not gates in series, preferably BJT.

FPGA would be more suited to the task of emulating old chips, but there's no reason why general-purpose cpus can't. You can look into TASbot, or the nintendo playstation. The guy who worked on recovering the nintendo playstation was able to dump the bios, and people working on emulators had the extra DSP/processor emulated it before he ran the hardware.

>>1546584
It doesn't stop it from being shitty to use. It links components and footprints in a retarded way, can't use standard keyboard keys to zoom in/out, doesn't do perfboards, etc. It's obviously meant for high-end use but there's just some obvious things missing to make it user-friendly.

>>1546588
The only software that "does perfboards" is software specifically designed for it. You can easily use kicad to design circuits on perfboard, just place your components and traces only on a 0.1" grid. It definitely has its quirks, but it's far better than using Inkscape for anything more complex than what you're doing right now.

>>1546603
I might be able to stomach it by changing hotkeys if I can zoom without panning. It can't do custom layers (jesus christ) so I can't assign a perfboard footprint to a layer and lock it (or even different components/connections).

Have you ever tried VeeCAD?

>>1546631
>It can't do custom layers
But it can? In the form of auxiliary layers accessed via the Setup:Layers Setup menu. There's only one dedicated trace layer per actual piece of copper, because that's how PCBs work, what would you use more than 1 layer for on perfboard anyhow? Plus I can simply zoom by scrolling, don't know what you're on about.

Nothing can beat dedicated PCB software as far as drawing a circuit diagram, assigning part footprints to the components, exporting this as a netlist, and simply importing the netlist into your layout and connecting all the lines it tells you to. I recently did so with possibly over 200 components and it was a breeze.

>>1546585
>you still read and write on the byte level
kek, no
>that it's cheaper and faster to use larger units of allocation than small ones
lern2flash, fren
>literally every switched-mode power supply is effectively a single-polarity level shifter
dude, stop

>>1546580
>only to realize I needed to jump a wire on the underside
do you not look at the convenient "Unrouted" counter in the lower left corner of pcbnew? or did you just skip entering your schematic entirely and went straight to drawing?
>doesn't do perfboards
different application entirely. if you aren't going to send boards out for manufacturing, KiCAD is not for you so stick with Fritzing

>>1546588
>It links components and footprints in a retarded way
how would you prefer?
>can't use standard keyboard keys to zoom in/out
wait until you try copy/paste
>doesn't do perfboards
wow so fucking what get a shitty pcb fabbed in china for less money.

>>1546655
My mouse doesn't have a scroll wheel. I have arthritis.

>why would you use more than one layer on perfboard
Top, bottom, jumpers, separating components into different lockable layers. This is a #1 for any design tool, to not have it forces only one kind of user to find your software accessible.

>>1546660
>kek, no
Are you that retarded, that you don't understand if we couldn't read a byte, that it would all have to be done on the software level? It's not. Literally SSD's these days are done on the group-level for performance and density reasons, and longevity suffers. We know this. It is a fact.

A transistor resistor circuit can easily do 400KHz+ for i2c, a little more can do 2MHz. This will well be beyond most consoles data rate, you may need to get dedicated chips for interfacing if you want 1:1 CPU levels.

SMPS were an example because they easily get 400KHz+ regeneration rates. To say "but how could we directly interface to a slow as shit processor???" is retarded.

>if you aren't going to send it out
>implying your software has to have exactly one use case
Fritzing looks useful. Wonder how difficult it is for custom parts.

>>1546663
Search for part, select form factor, place. Why should I need a schematic in it's software to route a PCB? That's bad abstraction, especially considering schematics don't have top/bottoms.

>doesn't have x feature
>go do something else
Yeah that's what I'll be doing.

>>1546387
There's a neat partial teardown here https://www.youtube.com/watch?v=DXYje2B04xE
It's splitting the 110GHz down to four quad ADCs running at only 16GHz, through fancy ASIC DSPs to fancy HMC RAM.

I have a question about transformers-

How does the primary power supply "know" that the load on connected to the secondary coil is drawing more current?
I mean take an inductor connected to a power supply,why does it suddenly draw more current if you put another inductor next to it? How does this work on a physics level?

I understand how this works from an engineering perspective (IE I know the equations check out).

>>1546671
For jumpers I just add 2 single-pin connectors in the schematic, set each's footprint to a single THT pad, and when I export the netlist into the layout it will give me the two footprints and tells me where to put them. The layers are solely for things on the PCB itself (copper, silkscreen, solder mask, etc.), jumper wires are not those.

>>1546678
>Search for part, select form factor, place.
>Why should I need a schematic in it's software to route a PCB?
>That's bad abstraction, especially considering schematics don't have top/bottoms.
Making a schematic is half the entire point of the software, by not doing so you're missing out on all the features that allow you to optimise your design. The Design Rules Check is a tool that you run to see if any pads are connected incorrectly, or if there's an intrusion into a no-go zone, or if there's some track going some place it shouldn't. The damn thing shows white lines between pads that should be connected, it basically draws itself. The software knows what nets are which and can make a lot of things easier because of this, not taking the 5 minutes to plot your dozen components on the schematic viewer is stupid, at this point there's no reason not to just use GIMP or paint. And I don't know what you mean by "tops/bottoms".

>>1546692
>>1546688
Once you start talking 10+Ghz I wonder how accurate the results are on the damn quantum level. Wonder what the threshold is for room temperature before a 10-bit sampler loses accuracy.

>>1546692
1) the first question is meaningless, the secondary could be higher or lower current

2) Current passing through wire induces a magnetic field around a conductor. An inductor is "more reactive" than air, so it condenses the magnetic field, creating a greater back-reaction (it has a greater "resistance" to an induction of a magnetic field than air). To be more precise, the inductor has a greater potential for a density of a magnetic field, so for the wire to induce this requires more current, and for a given unit of time, can produce a higher reaction (but takes more voltage).

Once the magnetic field is established, there is less reaction, sort of like a mix of newton's laws, that once something is moving it stays moving.

All of inductance is based on flux, that you gain a large negative reaction from the change.

When you couple two inductors, you're coupling that flux. If flux can enter one side but can't out the other, you'll be presented with a very high resistance. If you short circuit the other side, all induced flux can instantly be transferred as current through the wire, which behaves as a short. The ratio of current:current (because that's what transformers actually do) determines how a resistance looks across the transformer. In line/power transmission they transform the impedances, so you can do a single circuit without involving the transformers. More often they use voltage:voltage or turns:turns of course, since transformers can be expressed in terms of V:V, I:I, or N:N

>>1546692
With the hydraulic analogy, you have two seperate circuits full of non-compressible fluid. In each you have a turbine, both are mechanically coupled to a central flywheel. Normally, when the primary circuit is hooked up to an AC source and the second circuit is blocked by something (i.e. open circuit), no current can flow through the secondary turbine because it's open, so it simply acts like a big resistor that constantly tries to slow down the flywheel. Meaning not much current can flow through the primary either. When you close the circuit of the secondary, this resistance ceases and you get the maximum current. Arguably, this makes no fucking sense at all.

>>1546699
Schematics don't normally have tops and bottoms. You rarely see a symbol in a schematic which actually matches the pin out of the components (chips and even transistors). They're good for separating physical design from the abstract circuit, but going from schematic to board is made more difficult.

>muh debugging
When I don't have to care for extra safety features they shouldn't interfere with what I'm trying to do.

>>1546700
>>1546703
Thank you, it makes more sense now.

>>1546700
>Once you start talking 10+Ghz I wonder how accurate the results are on the damn quantum level

Quantum effects aren't even close to mattering at 10 or even 100GHz and there are specialized transistors that still have plenty of gain all the way out to 500GHz and beyond (HBTs, HEMTs, MESFETs, etc.) Various radar and telecommunication systems operate between 10GHz and 70GHz or so and visible light is up there in the 430-770 THz range and we don't have issues signaling with that. Yeah there's limitations with digital circuits and conventional transistors that limit their high speed performance but there are other materials that push frequency boundaries conventional silicon could only dream of.

>>1546585
>you still read and write on the byte level
False. To read a byte of a NAND memory you need to read the entire bulk of 512 bytes and then select the byte you want. The same for writing. To modify a byte you need to rewrite the other 511 bytes and then the byte you want to modify. Look some microcontroller drivers to work with SD cards at the lowest level with SPI.

>It can even be done as simply as two not gates in series, preferably BJT.
Yeah, there are specialized chips for the task but they mess with the transients and put a cap on your maximum clock speed because these switching times. You can do it of course on consoles like SNES, but the serial storage problem arises and you need to either be blessed to find a parallel ROM/EEPROM or juggle with a PLL clock multiplier and a serial memory that lets you access at almost 100MHz.

>extra DSP/processor emulated it before he ran the hardware.
People could even emulate the specialized chips in some SNES cartridges like Super FX in some flash carts. I don't know which black magic do they use to decipher the equivalent transistor circuit from die photos.

>>1546380
I would bet money that in 99.99999% of the possible usage cases actual measurement errors are tenfold higher than the devices precision.
What in the name of all things holy do you need this for? They're really pushing the boundaries of the uncertainty principle here.

>>1546704
I'm trying to tell you they do that for you, at least for the (substantial) list of ICs programmed into the thing. I pick a NE555 in the schematic view and place it. I select a DIP8 footprint for it and export its netlist. I bring that into the layout editor and since all the pins on the schematic's 555 were numbered 1-8 from the get-go, the DIP8 footprint already has the correct connections and is telling me where to put them. For example:

>>1546575
Considering that you can run dozens of Z80 or 6502 emulator programs simultaneously on a modern multi-core computer, it makes no sense to try to interface all that old hardware into one 'do all' computer.

That being said the way to approach the problem is by using another old technology: passive backplane. You design a CPU card for each different processor, and card(s) with peripherals. You enable only one CPU card at a time, and it masters the bus. The trick is designing the common bus architecture so that every CPU card in the system can work with it, and/or designing the CPU cards themselves to be compatible with the bus.

>>1546799
And when I reach a point where I need a jumper (as is inevitable in coarse-pitch 555 design) the DRC tells me that the ground-fill has failed to make a connection and that I need to add one some other way, which you can see here as a white line from the ground input to the 555's ground pin / pin 1.

>>1546803
So I go into the schematic viewer and add two jumpers at the bottom of the schematic (pic related), assign them names and footprints, and export this updated netlist.

>>1546803
>outline zone mode
disgusting

>>1546804
>export netlist
or, in KiCAD 5, you hit F8 and all your changes propagate from the schematic to the board, no netlist file required

>>1546804
And importing this netlist gives me the two extra footprints and tells me where to put them. Then I just have to update the ground fill (and add some extra fill in the middle) and it's done.

Makes me wonder how compact you can get a single-sided coarse-pitch 555 multivibrator layout.

>>1546805
It covers the component names otherwise, I hate it too.

>>1546807
Oh neat.

>>1546809
if you're using the new OpenGL canvas (pic related), you can change the transparency of the layers so that you can still see the silkscreen/other layers. and/or, select the silk layer of interest in the sidebar to draw it above the other layers temporarily

>>1546819
>you can change the transparency of the layers
Ah, that looks like something I'd want. AFAIK I'm on the latest version. Selecting just the silkscreen isn't enough since I'd want to see the other labelling layers too.

>>1546823
hit F11 to turn on the OpenGL renderer (hope you have it)
then, to change transparency, double-click on the color swatches in the layer selector

So I was repairing a clock radio the other day when I noticed a resistor attached to a leg of a 12v bulb that had AC straight from the transformer across it. So I can that the resistor is supposed to limit the current flowing in case the bulb fails short, but what would happen if it wasn't there and it shorted anyways? Would the transformer(coupled directly to mains) blow up?

>>1546732
>we don't have issues signaling with light
I mean, we only do AM with light, right? That's different. We don't have electronics that'll work with a carrier signal at visible light's frequency as far as I know.

>>1546803
>DRC tells me ground-fill has failed to make a connection
Nice. I don't think Altium does that

>>1546911
>FM with visible light
Man that would look cool as hell.
Wouldn't that theoretically allow us to produce pure white light?

>>1546911
>>1546913
Mate, you can do fucking anything with lasers these days. I guarantee that it's possible to do FM in the visible range with common components like non-linear optical materials, mode-locked lasers, etc.

It is possible to use this to make lighting with higher CRI than white LEDs (plus the reds everyone wants), but it's fucking expensive. Arguably you could use the incandescent glow of a farnsworth fusor, since they technically can get up to millions of degrees.

>>1546914
pic

>>1546914
>1546914
Ok, I'll give you that, but you can't decode it electronically. You could use a spectroscope, but again, that's not kind of a loophole. We're talking about the speed of electronics, not using other phenomena.

>>1546916
What do you think happens in fibre modems? The bitrate is comparable to the frequency of the IR light they use, and I don't think it's anything close to standard CW, but more along the lines of a complex AM signal. And by modulating the amplitude with enough finesse you can possibly get an FM signal out the end, though I'm unsure how possible it is.

>>1546918
IR starts about 3 orders of magnitude below visible light. I maintain that we don't have semiconductor or RF technology that can handle signals in visible frequencies.

>>1546919
The IR in optical fibres is near IR, not far IR. It's 180-330THz, visible light is 430-770THz. The maximum bitrate for fibre optics is 43Tb/s. To modulate the IR they do so by switching on and off an electric field across a nonlinear optical material that switches the light in or out of phase with itself, causing interference. Not too sure how they demodulate it, but both processes need to have electronics that can operate somewhere near 43THz.

>>1546915
That's very cool.

>>1546809
Why would you ever go through all this trouble if you could have just added a couple of vias and a trace on the bottom layer that you don't use anyways, and put your wire in place of this trace when assembling?

>>1546931
I was assuming single-sided, because the guy was asking about perfboard and jumper wires in the first place. If you were making a proper PCB, you'd just use narrower pads with smaller clearances and just run extra traces between the pads of the 555, I think that solves the issues without the need for another copper layer.

>>1546932
You don't get my point. I assume a single sided board as well with whatever crappy tolerances you can get, be it made on a perfboard, with a toner transfer method or even with a permanent marker. You have your unused bottom layer which you dedicate to your jumper wires, and your vias which give you nice pads to solder to. As added value, you don't clutter your schematics with meaningless and misleading header connectors, see exactly where jumper wires will go to avoid placing them over components, and need to do ten times less work, especially when you need to reroute a jumper wire to some other net.

This is kind of a silly question, but is there a good way to stack large amounts of PCBs?

I'm working on a modular design which should be able to accept an arbitrary number of modules, but I can't seem to find a good connector for it.
The obvious answer is the ubiquitous stackable headers, but I can't find a datasheet on what their rated resistance and current handling is.
I'm probably looking at ~20 PCBs stacked, with each PCB drawing around 50ma on +/-9v rails.

Pic related is similar to how my design is set up.

>>1546951
Did you consider PCI or PCI-e headers? Should be common and cheap enough.
> each PCB drawing around 50ma on +/-9v rails.
That's nothing, you can use literally anything you can find.

>>1546954
Oh well, disregard this post. I thought you plug them into a mother board.

>>1546954
>>1546955
Yeah. I'd go with a backplane, but I'd like to be able to accommodate as many modules as desired.

I didn't know if the stackable headers would be a problem for power distribution though, since the PSU sits at the bottom of the stack and everything else daisychains off from there.
I'm a bit concerned that by the time the power gets to PCB #20 there will be a non-negligible resistance from all the headers.

>>1546960
Well, you counter this by designating more than one pin to your power supply and ground. E.g. SATA power connector: there are six ground pins in total, and three pins for each of three different voltages.

>>1546966
I'll probably wind up doing that.
Thanks for the input.
I'll be sure to come back here when I've got more to show.

>>1546960
Compare the header size to the closest solid core wire size and you can probably make a rough estimate of the header's ampacity. I'm with the other anon, I think most standard headers should be fine.

>>1546918
Pretty sure fiber modems are all digital. It would be shit to convert AM to digital (can you call one-bit AM digital?).

>>1546922
Neat. Sounds like QAM. Wonder why they can't use lower frequencies, something to do with the difficult of THz generation, maybe.

How compatible are those common flash chips?

Could I just solder one from a fucked pen drive into a different one one if it meets certain conditions (e.g. same size, same speed, same flash manufacturer, ..). Or does it need to be the same exact chip for a pen drive to work with it?

>>1547044
Since they couldn't do testing on millions of devices, as long as it's the same spec I'd assume it'd work. Different batches could be a little different though, and that's ignoring hardware revisions

Anyone here tried to add an external wifi antenna on their phone?

>>1547071
My old s3 had a port on the back for testing that would supposedly work for that

why do the citizens of ohm hate ssrs so much? they seem pretty cool, completely silent, switch super fast and provide complete isolation

>>1547075

They are awesome. Compared with mechanical relays, they can cost more and they can require heat sinks and/or fans. Advantages are what you listed plus they can be directly driven by a microcontroller.

>>1547074
If i solder some foil on the onboard antennae, will that work as intended?

>>1547086
nope. antennas work on a specific resonance and any change in antenna shape will fuck it up and result in most of the transmitted energy being bounced back into the device instead of flying into the air

>>1547086
That's basically a Faraday cage. The more conductive a material is, the harder is for radio waves to pass through it.

>>1546919
but you're not working with electrons anymore when you're working with photons

>>1547044
>how compatible
8/10
>same size, same speed, same flash manufacturer
9.9/10, making some allowance for the possibility of low level formatting (bad block maps etc.) at manufacture

>>1547005
CW is noisy

>>1547005
>one-bit AM
usually known as on-off keying or amplitude shift keying

>>1547086
l/4=30.6mm

Hello. I may be a bit over my head (i have a soldering iron and have done simple repairs, fix broken cables and replace caps, etc.) so I'd like to check the feasibility of what I want to do with you fine gentlemen. I'm a turd world uni student (geophysics), and every year we hold a large public outreach event where we spend a week explaining all sorts of science shit to kids ranging from elementary to high school, and also the general public since anyone is free to stop by. I'm second in command of the Seismology stand, and was thinking of a project we could do that would be cheap, effective, interactive, and "impressive".

Here's my idea. I would fill a large plastic container with literal dirt, and then build 4 """geophones""" which would lie on the dirt. The geophones would simply be some cylindrical neodymium magnets hanging off a spring inside a solenoid wound with magnet wire. Any vibration would a (small) voltage. This would go to an amplifier circuit (need suggestions for this), which would feed into a microcontroller. By measuring delta_t of voltage spike, and knowing the geometry of their layout (also assuming constant wave velocity and linear trajectories), it should be easy to triangulate the location of where the waves originated (i include the 4th geophone to leave me with only 1 point instead of 2, even if the other is known to be false/in the air, just to explain it better to kids). I tell a kid to tap anywhere the container or the dirt, and a projector (microcontroller feeds data to laptop) shows exactly where he taped on the container. That's how we locate epicenters, boys and girls!

What I'm worried about is how fast the electronics can react. How finely can a microcontroller measure the time differences?How quickly and, more importantly, consistently will the geophones behave in terms of that first spike in voltage when the wave incides? If anyone has any input before I sign off school money on shit that might not even work, I'd be grateful.

I'm trying to design a flyback transformer for a Geiger counter, I've been playing with this 1:10 (26:260) flyback I've made, trying to clear out basic concepts, it kinda works fine stepping from 5V to 50V, but I still have many doubts.
The mode of operation will probably be DCM, since the output current is tiny. I know the output is roughly Vin*(N2/N1)*DutyCycle, but I've still got to figure out the number of turns in the primary, the primary inductance, the air gap required, and apply concepts like permeability. Any rundown or recommendation on how to design this stuff? I'd probably pick up a feedback controlled variable duty cycle driver once I've got the basics.

Also, what's a good transistor that can switch a decent load at 4 volts Vbase/Vgate?

>>1547258
I think you'd have more luck with electret microphones or piezos, as the magnet probably is too heavy to respond fast enough.
Imagine the sound travels at 1000m/s, probably, more, the differences between peaks will probably be in the order of fractions of millisecond. It might be feasible, but I'd check the time difference on an oscilloscope to be sure.
You can look at amplifier circuits online with OP amps, you just need a really high amplification to produce the spike to be read by the microcontroller. I guess if you use a fast enough micro with interrupts you might get a decent location.

I think what you described is already in use in some gun ranges, the bullet impacts a metal plate target, a computer measures the time it took for the wave to propagate, trilaterates the signal and displays a dot where you hit.

>>1547075
They have their place. I think for anything that turns on and stays on, mechanical relays are still better because they dissipate less power.
For anything that cycles regularly, SSRs usually make more sense.
I've also often seen SSRs driving the coils of mechanical relays. With that you get the advantage of the isolated input you can drive with a microcontroller, without the SSR having to handle the power.

>>1546922
>need to have electronics that can operate somewhere near 43THz.
I don't think so. These modulators and demodulators use laser diodes and photodetectors, they don't work directly with visible-light frequencies.

I bought a chink made triac power controller.
It has terminal block on the side for N-L N-L
I looked at the board and N from one pair is a direct connection to L on the other pair.
Thinking it couldn't be LINE and NEUTRAL I started looking for similar circuits.
I found pic related in nuts&volts magazine online.
As you can see, both terminals are marked N - wut?

>>1547286
The device I bought looks almost the same as pic related (no fuse on mine)

>>1547287
The schematic of my device.
I used >>1547286
as the starting point hence the similarity.

What does N and L mean on these?

>>1547288
>What does N and L mean on these?

mains power. L = line = the hot wire, usually black in america. N = neutral = the return wire, white in america.

>>1547288
>>1547298

I just noticed that L is shorted to N. So I have no fucking idea what is going on there.

>>1547298
>mains power
'splain >>1547286
>File: NV_0202_Marston_FIG6.jpg
two Ns - no L

>>1547310

nah, I gave up when I saw N shorted to L here >>1547288

>>1547258
i recommend you get in touch with some turd electrical engineer uni colleagues and ask them

>>1547269
I was just looking at this app note the other day, which might be just what you need
https://www.ti.com/lit/ml/slup127/slup127.pdf
>switch a decent load
IRL540N?

>>1547302
>no fucking idea
neither do they. apparently they believe that N and L are interchangeable, which is only sorta true, sometimes

Does anyone here have any patents? If so, what was the process like? Is it expensive?

>>1546859
the lamp would blow up before the transformer. transformers tend to have large thermal time constants. lamp might even work just fine with a shortened lifespan unless it's a led in which case it'll blow up immediately.

>>1547332
i think eevblog has a video about why patents are a bad idea for individuals. since you need resources to find and litigate patent violations, and publishing your idea makes it more likely that someone will try to use it in the first place. also they cost thousands.

Hey guise, it’s me, anon.

Two retarded questions as I attempt to grasp how electrons move:

1) I made this stupid thing awhile back, small lead acid battery inside and a 12V/3A power supply to run it off 120V if I were near an outlet. Thought I could wire the battery and power supply in parallel (I think?) so the battery would power it with no power to power supply, and then power supply would at least charge the battery to 12.0V when plugged in and power the lights. But when Indid that, something in the output of the power supply started glowing and smoking.

If I put a diode on the output of the power supply, before it meets up with wires to lights and battery, would that solve the issue? Is that how diodes work?

2) I forgot the second question. Oh, wait... capacitors!

Anon was making the spot welder, how do you wire up capacitors for that purpose? Does one post go to the battery ground, and the other post goes to battery + as well as + to the welding nail thing? Or is it inline with + in and + out like a switch? Bear with me, I’m retarded and when I try to Wiki or Google something, I end up with 30 tabs open.

>>1547378
get bent bepis

I need to replace a 24v brushless dc fan/ add two 12v dc fans to a 24v system. Can I just wire the fans in series?

I figure no because they're not simple resistive loads but I thought I'd ask.

Should continuity exist between the mosfet source and ground when the source is not actually grounded but voltage is going to the gate? I'm not sure if this is normal behavior for a mosfet, I also get continuity between S and G but only when black lead is touching G

>>1547411
maaaybe, if they're identical, and you put identical large caps across each one. I wouldn't vouch for reliability tho

Anyone ever worked with LCD backlight LEDs?
I'm trying to find something that will drive pic related properly but I'm not sure if the cheap chink drivers will do.
How far out of spec can you go with these LEDs generally?

>>1547399
haha bepis xDDD

>>1547421
The ground of what?

>>1547431
The ground in my circuit board that goes to the ground pin of a usb daughterboard. My mosfet source pin is not connected to anything yet theres continuity from it to the usb ground pin which I dont think is normal. I do have an inductor that is connected to the drain which heads to ground but theres a diode and cap inbetween so there shouldnt be any kind of reverse current leak. Im new to circuitry so some of this stuff confuses me

>>1546951
Have a look at long pin press fit headers typically found on PC/104 modules

>>1547429
Sounds like you want a constant current power supply capable of 55 or so volts?

>>1546196
>~10y old display making humming noise
>check cables, it even makes a pop when I turned it on after switching cables
>disassemble it to see if I can find the source of the pop
>can't see anything, assemble back and turn on
>makes a worse pop
>displays only green-ish
>disassemble again
>still nothing obvious to me
Well, the only things that I could recognize are bulged capacitors and black burned shit. Tried searching for service manual and schematics but most of the sites that come up want my money and only one actually let me read the service manual (but only one page at a time, good goy has to refresh ads) but it's not really helpful when I can't find R8 among like 200+ things on the boards. Considering just scrapping it.

Also, fuck this shit, I spent a couple hours refitting connectors in power cables (I was suspecting that just screwing stranded wires to contacts without anything to hold the strands together wasn't enough) but it was just the display shitting itself (again).

What is a proper way to recharge a nimh AA 1.5V battery?
What kind of voltage should i connect it to and at what kind of voltage in the battery should i disconnect it?

>>1547430
This is why generals are cancer

>>1547486
>What is a proper way to recharge a nimh AA 1.5V battery?
In a NiMh charger
and they're 1.2v - not 1.5

>What kind of voltage should i connect it to and at what kind of voltage in the battery should i disconnect it?
It has a charging curve, not direct points to shoot for.
This is why you should use the proper charger.

You can buy a charger on eBay for <$2 delivered to your door.

>>1547437
>My mosfet source pin is not connected to anything
mosfets are controlled by a voltage between the gate and the source. if the source is floating, it'll be pulled to a random voltage due to leakage. this can cause it to turn on, and depending on what the circuit looks like that can connect it to ground via the drain. also inductors are shorts at DC. your multimeter won't see them.

>>1547490
stop being a namefag and the cancer goes away.

>>1547288
M8, you have to switch one N-L pair in your schematics.
The best would be to keep your N(eutral) connected, an chop your L(ine) using this Triac switch thing.
The aim is to chop one of your phase (here the line) as this device can't chop both phases at once.
Don't short Line and Neutral nigga. I mean, it will probably work if you put a lamp on it, but it will be connected "backwards" on your power system.
This wasn't such a big deal back in the day and with resistive loads, but now that you have differential security systems, it will fuck everything up. In the sense that you will not be protected, not that it will not work.
Furthermore, if you are unlucky, the resistance between hearth and neutral could be small enough to simply blow your fuses is connected to in reverse. (hearth and neutral are connected together in your power supplier's transformer, the closer your house is to the area's transformer, the smaller the resistance between hearth and neutral is).

>>1547486
This >>1547509
Or if you want to diy, use la linear regulator configured as a current regulator. LM317 are an old as fuck design, but will work with like a resistor and a capacitor maybe.

>>1547558
You kids need to get over this '''namefag''' crap. It's legal according to site rules regardless of your not liking it, and protests about '''muh chan culture''' is just stupid garbage. Get over yourself.

>>1547486
The proper way is to charge it with a current-regulated source. C/10 is the generally-accepted overnight charge rate, where 'C' is the cell capacity in amp-hours; so if it's a 2200mAH cell, then you charge it for 10 hours at 220mA. You don't use a voltage-regulated source to charge secondary batteries because if you did then the current would be excessive, the cell would likely overheat and become damaged, or at the very least it would have it lifespan severely shortened after a relatively low number of charge/discharge cycles.

Some cells are designed to charge at a higher rate, like C/5, C/2, or even some can charge at C/1. Check the datasheet from the manufacturer for the specific cells you're using. All cells can be charged at C/10 however and that's generally recommended for best cell lifespan and safety.

>>1547486
>>1547509
>>1547583
>>1547648

https://batteryuniversity.com/learn/article/charging_nickel_metal_hydride

has anyone successfully used mcu integral pullups for switch inputs? i've read claims that it's a bad idea for some reason, but my particular mcu's pullup sources ~300uA which is about what i'd get out of a 10k resistor.

>>1547665
Sure, that's nice and all, and it's more advanced than what I was descibing here: >>1547648, but I don't think the guy originally asking is willing to go out and spend $100 on a NiMH charger, nor is he willing to invest in an Arduino or some other microcontroller just to charge a shitty NiMH battery. If he was serious about some application designed to use a secondary battery he'd use Li+ anyway, NiMH cells are '''mature technology''' and nobody really goes to all that much trouble over them anymore. Charging over time at C/10 (or less) is fine. So long as you don't let them sit there for days like that nothing bad will happen, it's not like Li+ cells that lose their shit completely if you overcharge them and turn into bombs and burn your house down.

>>1547669
*shrug* spin your PCB for external pullups and if you don't need them then don't stuff them and you lost nothing really, if the internals aren't enough then you have a footprint for them externally.

So I took apart some busted laptop PSUs to learn how they work, see the kind of rating the components have and hopefully scavange a few parts.
tl;dr
I have a bunch of those high frequency transformers, they are all caller ER-something. I can`t for the love of god find any info other than operating frequency and dimensions. I was planning on using one of them in a dc/dc converter but I have no idea of the coil ratio, inductances etc.. Anyone here knows where to find that kind of info?

>>1547675
>Anyone here knows where to find that kind of info?
Nowhere. They're likely custom parts and specific to the manufacturer of the power supply, or at the very least they're 'house marked' with custom part numbers for that power supply manufacturer, and they're highly unlikely to give you any information about them other than "you should not try to service our power supplies, please purchase a new supply if yours has failed, have a nice day *click*".

If you want to learn how switching power supplies work you should go look at switching power supply control ICs from companies like Linear Tech. The datasheets will have application circuit schematics in them, and usually companies like Linear Tech also sell 'demo' PCBs for their products to use as starting points for custom applications; you can also use those to learn from and experiment with.

>>1547675
you can characterize it with $20 worth of ebay equipment. you just need an LCR meter and a sine generator. characterize each winding by shorting and opening the other windings as necessary.

>>1547671
>I don't think the guy originally asking is willing to go out and spend $100 on a NiMH charger,
>>1547509
>You can buy a charger on eBay for <$2 delivered to your door.
It's specifically a NiMh charger for AA batteries.
They're available for 1-4 cells and charge/monitor each cell individually.
Why risk cobbling something up when a charger is less than the cost of a battery.

>>1547679
Yeah I know that. All I have is a scope, things aren`t cheap in the third world.
>>1547678
So I endedup with several meters of magnet wire?

>>1547680
Well if you go by that article that was posted then it makes it sound like you need a Masters in Engineering just to charge a shitty old-fashioned NiMH battery when that's just too extreme for how cheap and low-capacity they are. If you have an actual commercial charger or can get one cheap then sure do that but if you just need to charge one cell and have a current-regulated power supply or can cobble one together from a few parts in your junk drawer then why not do that? As so many here say this is /diy/ not /buyfag/.

>>1547681
The takeaway is it's application-specific. You could I guess dismantle them and re-wind them for what you want to do.

>>1547680
>>1547671
don't even use nimhs they are ouddated trash.
use liions instead or lipo bags. chinks sell them for like $.5 a pop, in normal store those fucking jews sell them for like fucking $20 a piece

>>1547692
The guy who was asking about that obviously doesn't have or want to use Li+ so it's kind of being a dick to tell him "throw that away it's trash get something else".

>chinks
Also there's no call for your racist bullshit on /diy/, if you want to talk like that then go to >>>/b/ or >>>/pol/ or >>>/trash/

>>1547699
slit your wrists, faggot. not even him but this thread is largely about shopping for $2 chinese garbage so it's perfectly relevant.

>>1547699
wow son, how did you get here all the way from reddit?

>>1547703
>>1547704
Faggot.

>>1547709
no u

>>1547709
>Faggot.

Also there's no call for your homophobic bullshit on /diy/, if you want to talk like that then go to >>>/b/ or >>>/pol/ or >>>/trash/.

>>1547258
You're looking at measuring individual µs. Hardly impossible, but perhaps you'd have more fun by putting microphones/speakers against/in the ground and ask the kids to stomp their feet or jump. There you'd be working with 10s of µs, and would have a surface to potentially measure shear, pressure, and surface waves. Plus you'd be dealing with a 2D plane and would only need 3 sensors, making the math a lot easier.

If you used time-to-digital converters you'd easily be able to measure those time increments (they're designed for use in laser rangefinders and such), otherwise with a 4MHz crystal on an MCU you'd probably not be able to measure anything smaller than 500ns or so. Depends on how many significant figures you want.

>>1547275
Laser diodes can't respond nearly fast enough. They instead switch the electric field across a nonlinear optical material so fast that you can get up to 40A of current going in or out of the capacitor.

>>1547669
I use 'em all the time

>>1547681
LCR meters that are sort-of-accurate in the ranges of interest are $7 on alibay
if you have a scope and a frequency generator, you have a way to determine the amplitude of a waveform. put a cap across the winding, feed a frequency into the parallel LC through a resistor, adjust the frequency for maximum amplitude across the parallel LC, and calculate L from c and f

>>1547685
>they're cheap, why not abuse them
>why doesn't physics bow to me reeeeeeee
because you want them to last more than three charges, maybe
NiCds are nice and abuseable. NiMHs not quite as much, but they won't explode if given the side-eye like Li+

>>1547722
>4MHz crystal
bluepills are cheap, and their timers can run at 36MHz or better. plus, a few of the timers on each chip have capture counters built in

>>1547718
Came here to say this
>>1547709
and this
>>1547692
Also there's no call for your dickphobic bullshit on /diy/, if you want to talk like that then go to twitter or reddit

>>1547699
>>1547718
>>1547734

this. it's high time we cleaned up this board.

>>1547741
this, no more antisemitism. all of us with degrees here understand that usury is a human right.

>>1547727
>their timers can run at 36MHz or better
Shit, really? That's not too bad.

>>1547741
>>1547746
Take a look at what happened with /g/'s csg threads, the anti-racist guy(s) were more trouble than they were worth, it just incited derision and more shitposting.

>>1547455
I'll need a way to dim them too.
Won't be making a board for this, just want to know if I can cheap out with a generic board really.

babby's first high-power layout. I'd like to stretch it to 48V but I don't want to set the building on fire. components can, but can the board? constructive roasting appreciated
>inb4
fuck J104 refdes

>>1547764
and they're nice timers, too. four channels of capture each on the "advanced control" and "general purpose" timers, no need to sweat about skew. if you do the signal processing in the analog domain, timers and pin change interrupts can be configured to do most of the work for you

>>1547797
as long as its specs meet your specs, sure
I think dimming might be a hard feature to find

>>1547375
>eevblog
I think I found what you are talking about, anon.
https://www.youtube.com/watch?v=Z7BL1O0xCcY

>>1547378
>and then power supply would at least charge the battery to 12.0V
12.0 volts is actually kinda low for a 12v battery, they usually sit between 13-14 volts when charged. if it was higher than 12 volts when you hooked it up, it started discharging current into the power supply. power supplies don't like that.
Even if that did work, sometimes a power supply can pull a little current from the battery and run it down.
A diode would solve both of those things, but has a voltage dropped across it (something around 1 volt) so your battery would never get charged very much at all.
So, really you should ditch that power supply idea, and put a little cheap battery charger into your box instead. It'll charge the battery right.

And you shouldn't charge your capacitors straight from a battery like that, they'll draw too much current from the battery. discharged caps are like a dead short circuit.
if you can't charge them with constant current, at least put a resistor.

>>1547803
Dimming comes standard on monitors though, it's how they control brightness.
The main driver board will have a pin for on/off and another for dimming control, which will interface with the led driver.
And finding something that meets the specs exactly is hard, boards I've seen are like 40-60V and 350mA current.

>>1547411
I just tried this with two (identical) fans and it worked perfectly. Even when one of them was loaded or stalled, they matched within 2 volts.
I didn't figure it would work either, but shit, it did.

>>1547722
>40A of current going in or out of the capacitor.
that's gotta be, what, picofarads?
jesus that dV/dt will tear you to shreds.

Stupid question. Do any serial to 4G or LTE modules exist? Are they good? I want to leave some device from a rural area transmitting data to a server.
If so, can you recommend me one to start looking up or should I settle with a Raspberry Pi and a connectivity dongle?

>>1547836
modems do still emulate a serial port and accept (somewhat) standard hayes-style commands.
what do you need the serial interface for? do you just want advanced control of the modem or do you need to hook it up to a computer with a serial port?

>>1547813
you might be able to change the output current by changing a sense resistor (e.g. R505 in my previous pic). also note that PWM might not change the peak current through the LED string, only the average. depends on the driver

>>1547836
they're out there but rare and expensive, e.g. SIM7100C
the connectivity dongle may be the way to go

>>1547841
it might be a question of not having unnecessary layers of USB interface in the middle, such that a lolduino might have been enough

>>1547841
Just transmitting some low res images to a server with a given IP

>>1547863
what thrilling news

>>1547863
what's wrong with the Pi+dongle that would be "settling"?

>>1547863
Yeah, a pi is definitely the right way to do that. Either a raspberry or any one of the numerous other fruit pis out there right now.

>>1547764
We're having technical discussions in this thread, there is NO LOGICAL REASON for posting racial epithets of any kind. It's just more /pol/ bullshit, invading other boards trying to spread their bullshit. You either stand up to it and say NO to it or it just infects everything, is that what you want?

>>1547985
Calm down it's just slang at this point, no one uses it to be racist.
/csg/ is doing just fine. Except whenever someone tries to change it of course, because nobody likes it when one fag thinks they can bend an entire general to their will.

>>1547985
>being this new

>>1547985
stop derailing the thread. unless you're doing this on purpose in which case bravo.

>>1546319
Technology is indistinguishable from magic to an unlearned mind. Put some time into learning about electronics and how they actually work on the physical level, after that you will either find yourself fascinated and eager to learn more or bored with the concepts.

What is a good entry level Soldering Iron? I find myself fixing more and more electronics and would like to be able to replace components if I can't find continuity. This may lead to a hobby, but mostly just trying to overcome my absolute fear of electricity.

>>1548052
$25 Yihua 908D, it'll let you git kinda gud

>>1548059
I can only find it with a EU plug, is it worth getting an adapter? Or should I just go for a us spec soldering iron?

Or is this some kind of sick joke where I'm supposed to solder a 120v us style plug onto it without a soldering iron.

>>1548064
just look a little harder. here's a fairly well-stocked low-end kit with US plug for $40 shipped, including a bunch of other tools that will make repair a lot easier for you
https://www.aliexpress.com/item/60W-220V-110V-EU-US-Temperature-Adjustable-Electric-Welding-Solder-Soldering-Benis-5pcs-benis-Tsui-Rosin-To-Rub-On-Your-Benis/32618139241.html

>>1548052
Personally I'd recommend a T-12 soldering station, the tips have the thermocouple built-in so you get less thermal delay between the tip changing temperature and the iron's controller reacting. The thing also has a SMPS so runs off any voltage. Just my 2 cents though.

>>1548059
Do those things run off mains directly? If so I'd quite like to see the circuit. Making a similar circuit to power a T-12 tip directly off mains is something I'd quite like to attempt as it would pack quite the punch and be all kinds of simple.

Is this safe?

>>1548090
that would be asking for shock hazard left and right. there is easily room for a 60W SMPS inside the temperature control box

>>1548107
yes, it's safe, it's very safe, it's so safe you wouldn't believe it

>>1548112
Seriously though this is my first DIY project and I'm worried that this will cause a housefire or something

>>1547807
Yeah the battery is old and likes to sit around 11.5V, but it was mostly built to see if I could do it. I may use those lights for another project in the future because they’re sweet. I figured the power supply would at least get the battery up to 12.0V, which is better than the low-mid 11s it sits at.

How do you know which size diode to use? The lights are like 12V-1A.

And the capacitor part,’pls remember I’m retarded and trying to learn and don’t know how to draw these things properly. But if I were trying to make the spot welder work, is this how you would wire up the cap(s)? Switch/relay not included in the pic, just trying to figure out what the two posts connect to if you’re trying to use a cap to provide a short extra burst of power.

>>1548107
Meh, I’m a fool but it should be fine as long as it physically holds and the wires are connected well enough to handle the current and don’t get hot. Just test the thing. You got quite a bit of bare wire on the ground connection though and you wouldn’t want that to tap the + part. If you want it to look more neat you could at least clean up the flux with some alcohol.

Is there anything wrong with this circuit? At one point it works perfectly, I get almost triple my input voltage the next day it either only slowly ramps up by a few mv or even drops by 1v. What the hell is happening, this already occured twice I'm not sure if my mosfet is just getting destroyed

>>1548114
seriously though, it should be pretty okay, especially if the two strips are firmly attached and enclosed. you should probably not leave so much of the wire ends exposed tho. use tweezers, small pliers, or something like that to hold the wire as you solder it onto the strip to avoid melting the insulation and your fingers. at least quick fix them with liquid electrical tape or hot snot or something like that

>>1548126
that would suggest that the MOSFET and/or the diode is getting fucked. first, do all your values in the real circuit match the ones in the schematic (especially L1)?

>>1548134
>do all your values in the real circuit match the ones in the schematic (especially L1)?
Yes

So i can use an ultrasound sensor to measure water level but, what kind of cheap chink sensor can i use to find out if the water is boiling? it's for cooking so it will be attached above the pots so i can't use any of them submergable temp sensors like potentiometrs

>>1548163
>potentiometrs
thermistots

>>1548163
You can use type k thermocouples to determine if the water is boiling and they CAN be submerged in water. I think those probes are typically rated to measure up to a maximum of anywhere from 250 to 500C, the exact range being dependent on the exact probe you buy. Either way it'll survive boiling water.

>>1548168
No, what i mean is don't want to put anything in the water because it's going to be eaten later it has to be be placed above a measure it form distance of up to about 1m

>>1547817
Sweet thanks.

>>1548169
I suppose you could put the thermocouple on the outside of the pot instead. You would need to measure the temperature on the outside of the pot when the water inside starts to boil and add an appropriate offset in software. It might be a bit buggy but would probably work.

>>1548163
Google infrared thermometer module.

>>1548176
>infrared thermometer module
Crikey mate, thats exactly what i needed! it only costs 5 bucks too, perfect for my arduino project, i will hook that shit up to my google home and then i will just be like
"hey google, tell me when the waters boiling"
fucking perfect!

>>1546380
holy fucking shit

>>1548177
fuck turns out those fuckers only work from like 5cm
back to the drawing board

>>1548213
if you're talking about the MLX series, they work from any distance. they just have a field of vision so they need to be positioned so they're not picking up background temps you don't care about. but that could be a good meter away.

>>1548218
it says the focus is 8cm at 10 cm so at about 70cm i need to place it it would be 7*8cm and that is massive, the average pot radius is like 30cm tops

>>1548224
there are MLX models with a 5 degree FOV. they do get a bit more expensive as that number gets tighter though.

>>1548116
the diode's rating should probably just match the output of the power supply. So, something like a 1n5400 diode. But with the forward voltage across the diode being 1 volt, your battery will only charge to 11 volts which is not much.

Yes, that's how you'd do it with the capacitors. Keep the wires fat and short, because they'll add resistance.

>>1548126
There is literally quite serious mismatch between component parameters.
So question would be input and output parameters (voltage and current), where it is used etc.

I would suspect dying M1 (where did you get this designator?) due to overvoltage since L1 a bit too high for inductance (I would expect something like 3mH to 300uH). Another question is about currents since transistor is rated for more than 10A continous while diode is 1us 2A peak pulses.
Also measure voltage between ground and after inductor, maybe set to max hold and see what you get.

Other notes - if it's high power application - output cap is shit, so high output ripple is possible.
In other words, design is questionable and I would test switching characteristics with scope under minimum and maximum loads intended.

I'm putting together a nice little remote garage opener.
The garage motor at my work only has one remote that came with it and dipshit coworkers keep misplacing it or taking it home in their pocket, so i am making this thing, which will allow anyone with a password to open the door via wifi from their phone.
It will plug directly into a 230v outlet which the little chinky power source turns into 5V DC which in turn supplies an esp 32 and a relay. The relay will be connected to a terminal normally intended for a push button on the door motor.

It fits perfectly into a plastic case, which i got from an aliex seller who sells scam rodent propellers (they only have a led and a button inside kek), but as it turns out those cases are perfect for dyi projects like smart plugs and shit

>>1548254
Also, if you look at the AC source pins you will notice that they are wire wrapped. That thing is cool as tits. I will replace the other dupont wires with wrapped wires as well once i'm done testing, they are just so compact and don't leave ugly black rectangles sticking out taking up space everywhere.

>>1547985
I'm gong to post this one time and not respond.
Rants like yours are equivalent to attacking a namefag or being a namefag. - both are just a distraction.
I filter namefags and tripfags so i never see their wit and wisdom. - I don't think I'm missing anything.
Posts I don't like that aren't named or tripped I just ignore.
If EVERYONE would ignore them, most would go away.
DON'T FEED THE TROLLS.
Stop falling for bait.

>>1548126
Your circult is open-loop: there's no feedback mechanism to regulate it, so chances are you're pulling too much current through the mosfet switch, and it only puts up with it for a limited amount of time before failing. Remember that when the switch is on it's basically a dead short across the 5V supply. Take a look at boost converter schematics and you'll see what I mean.

>>1548254
> via wifi
Please purchase a certificate and do https, otherwise enjoy your robbery.

>>1548336
>purchase a certificate

well, even if wifi wasn't already encrypted, and anon was for sure using http rather than any other protocol, you know he could simply generate his own certificate, absolutely free of charge.

Hello, ozone generator guy here. Had some health and personal issues and I`m slowly getting back to things. One question, how stupid would be to rectfy mains and feed it to an H-bridge that supplies the CW generator? (To increase frequency and as consequence, current)

>>1548247
A DC motor is connected where the volt meter is, I remember it pulled 14V max when it worked but the current was small like 20ma without a load on the motor. I measured 400ma of current from source to ground which is whats heating up the mosfet. Theres about 4.8V from one end of the inductor to ground however when I measure drain and source then check the voltage of one end of the cap to ground it reads 7v then descends under 4V. Im guessing measuring D and S induces a collapse in the inductor? Output cap is fine its high power. My schematic is how all basic boost converters look like so I have no idea why it ends up failing.

>>1548336
The only known way to breach WPA2 is by bruteforcing the password, so if you use long enough and strong enough phrase it will be completely secure.
Also the device won't be connected to the internet at all, it will act as isolated AP

>>1548340
> he could simply generate his own certificate, absolutely free of charge.
He mentioned his colleagues. That implies that he would have to make them install his CA, which obviously won't happen.

>>1548369
You don't know how ssl certs work do you?
THe browser just asks if you want to accept the unsigned cert or not and then uses it.

>>1548254
>rodent propellors

>>1548394
repellers
i thought spell checker is supposed to make you sound less retarded

>>1548394
>>rodent propellors
spud gun works great

>>1548169
then get one of the many certified food-safe thermocouple probes that are available, geez

>>1548343
about 3/10 stupid, just keep safety first

>>1548349
>induces a collapse in the inductor
no, inductors don't store energy that way. the fields collapse when current changes, and the collapse turns into current that serves to resist the outside current change

>>1548349
still trying to pull power out of nothing, I see

>>1548126
>P=[V/(f*L)]^2*L*f
>5V/(19.4kHz*330mH))^2*330mH*19.4kHz
> = 4mW

what this tells me is that this shitty boost shouldn't be spinning your motor at all because the inductor is barely transferring any power. so if your motor has a high leakage inductance presenting an effective open circuit for a few cycles then it might be killing your cap. if it's successfully spinning your motor at any point then either your PWM or your inductor isn't what you think, because only those two factors really determine your power output. if you've ever turned this on without the motor attached then you killed your fet, since boost converters transmit a fixed amount of energy each cycle and will keep doing so until something (the fet or cap) breaks.

>>1548435
>so if your motor has a high leakage inductance presenting an effective open circuit for a few cycles then it might be killing your cap.
disregard that bit, it was going to be another thought but i decided it was retarded and forgot to delete it

>>1548431
>food-safe thermocouple
I am not sure where i would get such a thing. I don't really trust chinks enough to order it form there and stores here just sell regular thermocouples

>>1548439
>don't really trust chinks
dude, it's a type K tc inside a stainless steel probe. really hard to fuck that up
anyway, restaurants commonly use type K thermocouple probes for food safety checking. if you're paranoid, try a restaurant supply shop

>>1548442
Because one day then can just go like
"oh shittu, we run out of stainress steer, let's just use this water dissolvable cyanide material instead" and give zero shits

>>1548435
>if you've ever turned this on without the motor attached then you killed your fet, since boost converters transmit a fixed amount of energy each cycle and will keep doing so until something (the fet or cap) breaks

I have and that's probably what killed it. However aren't you using the wrong equation? I think instantaneous current should be included and since like I said before I got a max of 14V, this motor was spinning fast.

How would I do this? If it's lossless, shouldn't the impedance be equal to JB since all of it is imaginary?
I've been reading the lecture notes for hours and they're not helpful in the least.

>>1548478
Could it be done with Z0 = sqrt(L/C) and B = W*sqrt(L*C)? I'm not completely sure of the difference between the low loss model and lossless.

>>1548163
>figure out if it's boiling
Gotta measure the amount of steam being produced, use a flow meter on a hole on the lid. Patent pending. If you know what you're doing you could always calculate the rate at which the water level is changing, or check for the froth going too high.

>>1548343
At that point you'd be better off making a flyback converter. You can buy silicon stack diodes that go up to 20kV for a couple dollars (source: I bought 10) though they do have a forward voltage of 50V each. I plan on putting them all in series to get a few hundred kV to charge up a diy capacitor or two.

>>1548343
There's also a parallel-feed version for higher currents.

>>1548501
I`m using one negative and one positive CW generator, that way instead of a huge shlong with little current I got two small shlings with extra current to achieve the same voltage.
>>1548499
>flyback converter
lol no, those things are fuck huge
and
>death
And If I`m going to buy anything to this,it`d be a neon lamp transformer.
>couple dollars
not in banana country. Here everything has 50% tax and 1 trump = 4 banana monies.
>>1548501
>pic
Really activated my almonds, although if I`m going down that path I ought to find a small isolation transformer.

>>1548505
>those things are fuck huge
Not really, the higher the frequency the smaller they are. At 200kHz you'd only be running a dozen primary turns, not to mention you only need 1 transistor, not 4. Grab a transformer out of a CFL bulb or wind one yourself, probably best to wrap it around something insulating (not ferrite) because at the end you'd have a pretty high voltage. I can understand that you'd have an easier time getting an H-bridge than a diode stack though. If you do go for this bridge method you could use a half-bridge since the duty-cycle will be 50% and pretty speedy, but I've never seen an ozone generator have a trouble with current. The air they're discharging into has a pretty high resistance, there's a reason the things typically use a simple CW multiplier.

>>1548514
Ozone generators use transformers, ionizers use CW. At least in my experience. Thanks man. I don`t think I can get ac with a half bridge.

>>1548517
Half bridges just have two capacitors to supply the current, if you switch the thing fast enough there won't be any difference.

>>1548505
Oh doesn't it not make sense to have a positive and negative CW multiplier because only the positive end actually does work? Not like you can just ground the opposite end since it's already referenced to 0V.

What voltage are you aiming for anyhow?

>>1548468
it's intuitive if you get the cyclic nature of a boost converter. in the on state, all that's happening is the mosfet is permitting current to pass through the inductor. that accumulates a fixed amount of energy no matter the output voltage, since the inductor is just being shorted between the input and common. so for one on-cycle:
>I = V/(Lf), standard inductor current formula slightly rewritten
>E = I^2 * L/2, standard inductor energy formula

when the switch turns off, all of that energy is going to go somewhere. as long as the parts are all working, that somewhere is going to be the output. since all the energy goes to the output, and that happens a fixed number of times per second, power is:
> P = E*f

discounting complications like continuous conduction (prob not true here and not really relevant), that's all there is to it. if you're seeing behavior that contradicts the above then it's only because some other circuit behavior is not what you think it is, or just not precisely like the pic you posted.

>>1548499
>50V/diode
Heavens, that'll ruin a lot of your gain. Starting with 230 won't be that egregious but still, that's a lot of loss.

>>1548517
There's no hard line between an ionizer and an ozone generator. None. All ionizers generate at least a small amount of ozone, and the more current/voltage the more ozone and fewer ions you have. If you're having lung problems now, that is the first symptom of ozone toxiicity. If you have asthma or bronchitis, don't even be near it when you're generating ozone. I assumed you had some awareness of this before, but perhaps you think you know better.

>>1548535
It's a flyback, you only need a HV diode on your secondary side. i.e. you'd already have a few hundred times that voltage on the secondary so it wouldn't matter.

>>1548530
That's a nice boost converter explanation, thanks

>>1548536
Oh, I misread and though someone was gonna use 50V diodes in a ladder.

>>1548539
I don't think that would be worth doing anyhow since all the diodes are in series and would only ever have V_peak over them.

>>1548527
>Oh doesn't it not make sense to have a positive and negative CW multiplier because only the positive end actually does work? Not like you can just ground the opposite end since it's already referenced to 0V.

It does. Imagine a n-stage CW with supply in the middle stage. On one side there will be a positive V and on the other a negative one. Betwen those 2V. (I know because it`s already built, it makes the zaps and a bit os ozone.) I was planning on putting it in the H-bridge. The benefit of using 2 smaller ones is that a small CW has better regulation.
>>1548535
>There's no hard line between an ionizer and an ozone generator
I don`t understand, I thought there were big units with fans a big coil rolled up on a glas tube that were used to clean hotels and hospitals and then there where the ones with needles to make that ionic wind. I`m planning to leave it on when I travel for a couple days.

How do you hamfags amplify the signals from your oscilators? I have a LC oscilator that gives 2Vpp at 500khz and 9v supply.
I want to get it to 8-9v then boost power
is using an voltage follower and then an amplifier enough (using single BJTs, AC coupled)

>>1548435
Looking at the inductor again its 330microhenry I confused the shorthand terms

>>1548604
That's not using electricity, that's using UV. UV generators explicitly generate ozone. Negative dischargers make both. Don't try to no-true-scotsman me you wanker.

>>1548606
You have to check your stuff. Some transistors/mosfets are good into the MHz range, but usually don't provide that much voltage gain or power. Antenna work is usually a fixed-load antenna (50 ohms) and they use amplifiers designed with that in mind (power amplifiers, E and F for example).

Ham radios often used tubes for that reason, tubes offer higher bandwidth/power than transistors for a very long time. These days if you're using mosfet drivers you can get away with higher power, but they still usually use 50 ohm antennae.

>>1548604
The ion generators or whatever it is strip electrons from atoms by using a positive electrode, that electrode holds a net positive charge relative to their surroundings. It doesn't matter what kind of voltage you have on the other electrode, it's not going to strip electrons off anything. If you could isolate the CW multiplier and ground the negative electrode then you'd be able to boost the positive one up to 2V, but as is they're both ±V from 0V, and therefore ±V from the air.

>>1548604
Positive generators make nasty byproducts, worse than ozone iirc.

>>1548638
>>1548639
ok other way around, same concept though

>>1548606
JFET cascode for voltage gain and a class C amplifier for power gain. Thank me later.

>>1548635
He doesn't need fucking tubes to amplify a 500kHz oscillator. Jesus, talk about overkill. Any VHF JFET or BJT can do it just fine. For that matter general purpose JFETs and BJTs should work too. Hell, he could probably even get away with a sufficiently fast op-amp for voltage gain at that frequency, not that I'd recommend it mind you.

>>1548642
I didn't say he should use tubes.

Can I use a class-D amplifier as a power supply?Just hook up a pot to the input and go from there? The only thing I can think of that would prevent this from working is if the circuit had a DC-blocking capacitor somewhere, but I don't think they do. Could also use it as a buffer for a function generator of course.

>>1548665
sure, if you like this

>>1548668
I could always add my own LC filter to the end to drop the corner frequency to ~10Hz or so. If they make music sound good enough to be played on speakers I can't imagine I'll be getting more than 1% ripple. Or is it just hip to go linear these days?

On a related note, how easy would it be to use a LDO at the end of a variable switchmode power supply (or class-D) to drop the last volt or so and eliminate that ripple? I can't figure out how you'd drop the voltage of the feedback into the switcher itself. From what I've seen this is what high-power commercial units do.

>>1548673
You should use a capacitence multiplier as well which should kill a lot of your ripple if designed well. Using a class-D amplifier is still not really a great idea for a power supply because no feedback. If you're just doing this to prove you can then whatever but I'd recommend a proper switched mode controller if you actually want to build a serious usable power supply.

>>1548665
the real answer is no, amplifiers are designed to drive power into a narrow range of loads (impedance, type) while power supplies are designed to cope with any or no load. it's a matter of stability and compensation

>>1548673
>a LDO at the end of a variable switchmode power supply
a commercial three-terminal linear regulator? meh, I wouldn't. if the linear stage's current consumption is low enough you might be able to use a double-ganged pot, but it had better not scratch. or you could use an adjustable current source into the feedback of a four-terminal regulator to push the output voltage down like Pic related

Probably a stupid question but from what I understand, in a circuit with a LED and a battery, electrons flow from the negative side into the bulb, vibrate the filament which creates photons that are emitted, and the electrons go back into the battery. I got 2 questions though

1: why is it when I cut a wire that goes into the positive side of the battery, the light won't shine anymore? Aren't electrons still going into the light and making the filament emit photons?

2: what is the purpose of electrons going back into the positive side of the battery? I originally thought that electrons just went to a light and it's energy was converted into photons, so it doesn't break any laws of thermodynamics, so I'm wondering how are electrons still even there when they come out of the light?

>>1548680
>LED
>filament

>>1548682
Sorry, I have no idea how LEDs work and assumed they were mini lightbulbs. Just replace the word LED with lightbulb in my question.

>>1548676
Aren't capacitance multipliers purely linear?

Class-D amps do have feedback, they don't just rely on modulating a duty cycle proportional to the input voltage value. Hence I assume they'd work with loads of different impedances and reactances to some extent. Of course they haven't got a lick of current regulation.

>>1548680
Ok, first of all there isn't any vibrating filament in an LED, but that's a whole topic in itself and it's kinda quantum physics. The filament in a normal bulb doesn't vibrate either, it just gets hot.

Electrons hate each other. Similar to poles of a magnet, electric charges repel one another if they are of the same polarity and attract one another if they are opposite polarities. So by having only one side of the battery hooked up, the electrons travel through the wire, through the LED, and run into a cut end of the wire and would start to bunch up. But the electrons would build up there and repel each other so strongly that the current would effectively stop immediately. In practice, the electrons actually move incredibly slowly, and it's the propagating electric field that moves quickly down the wire. As soon as you cut the wire that electric field ceases. It is important to note that batteries are not just electron sources, but electron sinks as well, and that the mechanism behind their energy storage requires that a current flow into one terminal and out of the other.

what sort of adc resolution do i need to read a thermocouple? what sort of voltage should i epect from it with 1C change?

>>1548680
>what I understand
You don't understand.

As retarded as it is, the water flowing through pipes model of electricity would be appropriate for you because your current level of understanding is fucked beyond belief. Electrons are like flowing water, the battery is a pump and the LED is a turbine that is spun by water. If you cut the connection to the inlet of the pump it won't have any water to push out and the water will stop flowing through the turbine and the turbine will stop doing work.

Electrons do not become light; you can make them gain energy, vibrate them the right way and make them turn the energy into light, but the electrons themselves do not change. You're missing like an entire high school education of physics here so search wikipedia for what light is first.

i love the slurp127.pdf

Why the fuck are current feedback op-amps so expensive? They aren't really anymore complex internally than standard op-amps despite the different topology.

Do they only command premium prices due to their insanely good performance?

>>1548878
CF op-amps require more transistors according to http://www.ti.com/lit/an/slva051/slva051.pdf and have a tendency to self-oscillate if they don't have proper feedback compensation. To make them ideal voltage amplifiers (the most common use for an amplifier, surprisingly) requires high-value resistors, since the bandwidth and stuff is dependent on that. Ironically integrating high-value resistance is difficult, which may limit their usefulness. The non-ideal components are also coupled to the inputs.

Seems although they have some performance benefit at really high frequencies (100MHz-GHz), at lower frequencies their gain isn't so much different than op-amps. I also don't see that they have decent voltage support, so it's not like you can amplify something to +/- 9-15V, as is doable with even some of the cheapest op-amps.

Best guess: their ideality is most obvious at low voltages, really weak signals, and really fast signals, which are kinda pointless outside of integrated circuits meant for specific uses (fiber optic amplifiers are one example I saw). They've had cheap transistors that have 100MHz GBP for decades, so if you needed only an amplifier, that'd be easier than a CF op-amp. Designing for transconductance amps is more difficult, and we just don't teach it, so there's that too.

>>1548685
>Sorry, I have no idea how LEDs work
Light
Emitting
Diode

It's a diode that gives off light.

>>1548844
bout tree fiddy
http://www.ti.com/lit/an/sbaa354/sbaa354.pdf

>>1548945
So i suppose 12bit ADC won't be enough to get a temperature with 1C precision then?

Is there anywhere to get cheap matched transistors, preferably under $0.50 per unit in one-off quantity.

Right now I'm looking for matched NPN pairs and matched complimentary NPN+PNP pairs as well though I wouldn't mind some matched N-JFETs as well if available.

I want to clarify I'm talking about two transistors on the same die in the same package, thermally coupled. Not two discrete transistors with closely matched properties.

>>1548957
a pre-amp would help (50x gain?) but you'd have to deal with input offsets, cold-junction compensation, zero point calibration, etc. yourself

>>1548960
>on the same die
if adjacent devices from the same wafer is good enough,
https://www.digikey.com/product-detail/en/diodes-incorporated/DMMT3904W-7-F/DMMT3904W-FDICT-ND/717879
covers the dual-NPN case
>matched complementary pair
the physics of that aren't favorable, good luck and let us know what you find

>>1548964
I probably should have mentioned I was hoping for THT to make breadboarding and prototyping easier though I may be forced to go SMD anyway because the THT component prices are always worse. Though I have to question if SMD really is economical factoring in the cost of all the breakout boards I now have to buy?

>>1548967
consider flipping the ics over (and glueing them to a surface) and soldering wires directly to the pins. it's really not too bad if you're only dealing with sot-23 or anything under 8 pins. google deadbug soldering for info

>>1548642
>tfw no discrete jfet or signal fets
>can`t design a bjt cascode for shit

>measure oscilator
>1.2Mhz
>measure same oscilator with a buffer
>1.5Mhz
why

>>1548967
chinkshit SOT-23 boards are 3-4¢ each
gold plated machine pin headers are on the order of a penny per pin, and if you don't mind wearing outyour breadboard a little you can use "Dupont" pin headers
also dead-bug is not a terrible idea, given a solid frame

>>1549014
>he doesn't know about loading
try sketching out the equivalent circuit and see if you might have unwittingly placed some capacitors in there with the direct connection to the frequency counter

>>1548847
Hydraulic analogy wins again!

>>1548844
https://www.omega.com/techref/pdf/z204-206.pdf (for type-K)
Looks like around 40µV/C, and a 12bit ADC with a 1V reference voltage will measure down to steps of 244µV (from 0V to 1V), so it's not feasible without an amplifier. Use a gain of ~20 and you'll be golden, as that covers from 0V to the maximum 50mV
The voltage it provides is proportional to the difference between the two sides of the thermocouple, so you also need a conventional temperature sensor (thermistor, diode, etc.) on the near side also. A MAX6675 (module) will do all that for you and output I2C data for your MCU.

opinions on this starter kit? programmer who knows very little about electronics, want to get into embedded stuff maybe
https://www.amazon.com/dp/B01EWNUUUA/

>>1549050
overpriced jewery and you won't even use half of that crap
not to mention that arduinos are archaic, buy esp32 instead

>>1549050
It comes with a bunch of shit you'll probably never use and is probably overpriced, but in general it does contain a handful of neat curiosities. The stepper and servo and screen are things that are handy to learn how to drive, the rotary encoder is handy, plus you get a bunch of passives that you'll certainly use. You can probably find a similar combination of parts on aliexpress for perhaps $20. At least I wouldn't pay any more than that for it. There are also kits with just a bunch of sensors/modules for $10-$15 with no breadboard or wires or passives, which you might want to consider (instead).

>>1549053
>buy esp32 instead

ESP has wifi but in every other respect it is vastly inferior to STM32 devices.

and there is a better thread to discuss toy devices like arduinos, ESPs, and STMs: >>/1545302

>>1549023
>>he doesn't know about loading
I don`t. I suppose that`s one of the practical aspects they don`t approach in theory classes.
It means that the ``true`` frequency is the higher one and the emiter follower allows my measurements to not skew results? I`m using a scope.
>>1549050
See a project you want, buy 20x of the cheap parts, 10x of the medium ones and singles of the expensives, start with a cheap meter and go up from there. Soon you`ll have loads of parts you`ll use here and there when needed. And always be on the lookout for garbage. Most things thrown away have some quite usefull components
>inductors
>ferrite cores
>li+ cells
>leds
>big capacitors
>tiny and big motors

>>1549064

retard: >>1545302

>>1549053
>arduinos are archaic
>80% of all hobby projects we see arround could be made with a 80`s PIC in assembly.
>some retarded shit with magic wireless

>>1549072
this

>>1549058
That makes sense, thanks, I'll check ali

>>1549023
I need a SOT-363 board for the particular matched transistor package linked above.

>>1549066
they do approach it in theory classes, at least good ones, where they talk about how real components can/should be analyzed as combinations of ideal components. e.g. in addition to the obvious resistance, a cable might have several pF of capacitance per meter and some inductance too. the emitter follower is isolating your oscillator from ~99% of the cable capacitance. if you add another emitter follower stage, you might notice a tiny bit more shift of frequency (several hundred Hz?) but nowhere near 20%

>>1549105
then flip it over

so let's say I find an old LED, but I have no idea how bright it gets or how much voltage/current it can take. Is there a way I can see what its voltage rating is without running the risk of burning it out?

>>1549107
Maybe I'll pick some up then. It'd be nice to have on hand. Wish I didn't have to go to ali though. I always have trouble ordering from ali because my bank automatically places a freeze on my card every single time I try to buy from there, suspicious transaction, they consider it fraudulent or something. Spent a half hour on the phone with the bank last time trying getting a lecture from a guy on security and why I shouldn't make the purchase all because I referred to ali as "Chinese ebay" and that made the guy suspicious. I really wish they had paypal support and I didn't have to either hand over my card details directly or go through sketchy Chinese intermediaries. It really puts me off which is a shame given the cheap prices.

>>1549108
The maximum power is typically a function of what size package it is. Somewhere around 50mW is fairly standard for a 5mm LED, SMT LEDs can get higher since they have some heat sinking into the solder pads, etc. Find an LED in the same package on digikey or wherever and go into its datasheet to find thermal information to calculate its maximum power at room temperature and on whatever substrate you plan on mounting it. Also note its polarity if it isn't obvious. Then you just need to feed it a low current in the right direction (10mA is fairly safe for small packages) to see what color and forward voltage it's after. Color is a pretty good indication, basically 3V-4V for UV or blue or white, 2V-3V for yellow or orange, 2V for red, and 1.5V-2V for IR. But this won't tell you its reverse voltage, which you may be able to find out by putting a very low current through it backwards but don't quote me on that.

>>1548372
No, you are.
"Unsigned" certificate window = MITM. Not to mention browsers nowadays are picky about it, not even sure if mobile chrome actually allows to accept it. AND they would do that alert every time. Unless there would be some sort of an option to install this thingy, forgot how it's called. Something something Root CA?

OR

he could just buy one for as cheap as 10 bucks

>>1549108
I hook it up to a power supply with the current at 0, then turn it up slowly until it looks bright enough.
If you don't have a constant current supply, just use a (few Kohms) pot in series with it, across a 5 volt supply (or whatever you can find). Then turn the pot down slowly until it looks bright enough. you can measure the voltage across it, and you can measure the pot's resistance too and work out how much current was passing through it.

>>1549112
a bit more expensive than ali, but here's an ebay
https://www.ebay.com/itm/50Pcs-0-95mm-0-5mm-SOT23-SOP10-to-2-54mm-DIP23-DIP10-SMD-IC-PCB-Adapter-Benis/311008365133

>>1549128
Benis :DDDD

>>1549050
I have that exact kit, but I’m slowly learning about all the electronics so it’s way above my head. Then again, I haven’t really had the time to mess with it. I need to send the girl away to a spa for a weekend so I have time to fuck with it. Hopefully it’s not like “Plug this into this and that into that” without ever learning what is actually happening.

If you know absolutely zero about electronics, maybe get a multimeter and work on some of the other beginner projects. I just got some potentiometers the other day so I was playing with those and the power supply and I sort of understand what resistance it. But I’m a dumbass and do much better with mechanical and tangible stuff I can see so maybe you will do better than me.

>>1549141
https://www.ebay.com/itm/I-THINK-YOU-CAN-PUT-HERE-ANYTHING-YOU-WANT-ITLL-STILL-WORK/311008365133

>>1549148
https://www.ebay.com/itm/311008365133
nothing is anything

>>1549146
>I need to send the girl away to a spa for a weekend

stop posting in /ohm/ you fucking cretin. there's a perfectly nice home for you here >>1545302 and I will answer all of your stupid assed questions there.

just curious, and answer here >>1545302 so as to not nigger up this thread any further, but does /out/ laugh at you or do the innawoods knuckledraggers embrace you as one of their own?

>>1549128
You can get these from OSHpark or pcbs.io and support the local economy, if you live in burgerland. The quality of pcb is better too.

Search the shared boards for sot23, or breakout.

>>1548960
I have a random chip with 4 matched NPN's: CA3406. So old it's not even sold anymore. Sounds like you're trying to make a cascoded B amplifier, since you can use exactly two matched NPN's and a matched PNP/NPN output buffer.

>>1549112
Some things on ali have paypal for payments

Am I crazy to want to build an smd pick and place cnc? I want to get into pcb printing and this is one of the things I want to go along with it.

>>1549163
if nothing is anything, then everything is

>>1549141
XDD

>>1549200
>3406
not 3046?

>>1549278
the line between crazy and ambitious is wide and fuzzy. that said, you wouldn't be the first to roll an own pick-and-place bot

I want to build an SBC cluster, and I'm thinking of wiring up a PCB to route power + ethernet to avoid cable hell.
Power I can deal with, but is there anything I'm likely to fuck up when routing ethernet?
Stack overflow says try to get the traces length matched and match impedance to 100 ohms.

>>1549281
Ethernet will want to be twisted pair if you're going near Gbit speeds, I'd just go with a bunch of 10cm cables to a Gbit switch in a good position, perhaps installed protruding from a project box along with some barrel jacks or USBs so you can just plug the computer boards directly into to the chassis. This way cable management doesn't matter as it will be constant and contained all the time.

>>1549281
curious what you're doing about the physical layer, since RJ-45 males don't usually come in pc mount

I have the need to capture a voltage trace of a solenoid's pull in and drop out because reasons.

The solenoid is powered from a floating power supply.

When I measure from positive I get vastly different readings than when I read from the negative end. Both seem rather noisy. I tried using two probes with nothing tied to ground and doing math to add or subtract the two traces. I tried tying positive or negative to ground. Do I need a differential probe? Should I underground the oscilloscope? Should I just scream and shout and let it all out.

I can do a current trace with a shunt resistor and the thing looks cleaner but for reasons which I cannot specify that is not acceptable.

A non shit differential probe costs as much as my scope, would it actually fix anything or will that show the same noise?

>>1549306
Nah bro, it's simple. Just have the PCB extend into regular cheepo jacks and have carefully positioned holes for the crimpy bits to smush into.

Buy a couple of these to disassemble and reverse engineer.

https://www.amazon.com/dp/B0058RW5CQ/

>>1549304
I've got this now. And the boards themselves don't go over 100M (Orange Pi) which is still fast but I figure if it can survive a few meters of cat5, it can survive a few cm on my board.
"Just stuff it in a project box" is what I usually do but I'd rather have something more planned out

>>1549306
I'm just now figuring this out, I was expecting to find dozens on digikey but was sadly mistaken. I bet I could jury rig an SMT cable like >>1549308 suggests. I've also got a part number (AJP92A) for such a connector, but I'll probably still go with the self-built solution.

>>1549308
I think your axes are misaligned. but I could imagine a rectangular cutout to fit the jack through after lid removal, with some pads on an inside edge of the cutout, on both sides of the (thickness carefully regulated) board. sick, but intriguing
I probably would have considered taking the standard crimp-on RJ-45 end and welding flex circuitry to the top on the one end and an FFC connector on the other. or an electrically short length of pitch-matching ribbon cable from that jack to the backplane, or something like that. just enough to allow for usual qaulity physical tolerances

>>1549312
I was thinking something like this:
As it would be more rigid than a PCB with board-mount jacks on it, if that's what you were thinking. Also try twisted pair on a 2-layer board with vias and shit, that could be neat.

>>1549316
Like this?

>>1549317
Yeah, exactly. I bet the high-frequency action of all those jagged turns is worse than two straight lines though. Just do a normal straight line lot and shield the box, should work out fine.

>>1549146
>“Plug this into this and that into that”
But the entire electronics is like that son.
You plug component A into hole B and ejaculate solder all over it to seal the dead.

>>1549318
impedance control would be fucked too
you have much to learn about high-speed layout, grasshopper

>>1549312
yeah, just impedance-control your traces and you'll be fine

>>1549319
>not getting fluxy soldery wetness all over your fingers and groping the board until all melts
jfc don't they have safer sex ed in electronics class anymore?

Anyone have a timer chip suggestion? Need to convert a 5v negative pulse (i.e. idle 5v pulse 0v) ~4-8us wide to one always 6us wide also negative.

>>1549327
A 555 monostable should do that, plus or minus a transistor inverter or two.

>>1549332
The 555s I've looked at say minimum pulse 10us

>>1549333
CMOS 555?

>>1549334
aah, gonna do some spice sims to see what that can do

>>1549344
RTFDS
fmax Maximum frequency in astable mode RA= 470 Ω, CT= 200 pF, RB= 200 Ω (2)
(2) RA, RB, and CT are as defined in Figure 1.
But there is no Figure 1 in this document and the link sends you offshore. It should link to Figure 6.
This means the TLC555 can handle signals down to 500ns or so.

>>1549327
74HC123?

>>1546196
Falstad circuit sim does not have coupled inductors or multi winding xformers, does it?

>>1549307
if i understand correctly you left the ground leads off when using the math function which I'm pretty sure doesn't work very well. try grounding at some random place and retry

>>1549112
>chinese ebay
kek

Have any of you guys come across a wideband DC-DC converter? Vin would need to be 6-15V and output ~12V and rated for 800mA.
Its to keep an ECU running during engine start up. The lowest I can find is 8.4 volts.

Making a "precious plastic" injection molder but I have to make a PID but I have no experience at ALL beyond soldering wires twice.

im trying to read this, there are 2 positive and 2 negative coming from the power of the same color. Unless one of those is ground which I don't believe so how would I connect 2 points from 1 original wire? Would I split the wire in half or solder 2 wires in a Y shape? Something else?

>>1549569
here's a pile of trash that will do what you want.

https://www.ebay.com/itm/143074980640

>>1549572
Firstly, in AC there is no + or -, but there is live and neutral. I'd assume that all the blue wires coming from the "Power" box are commoned to live, and all the red wires coming from it are commoned to neutral. But also note that since the SSR requires LV DC to switch, the switching wires coming from the controllers are at a different voltage even though the wires are the same colors. Pretty shit diagram if you ask me. Plus that voltage controlled switch is a mystery and a half.

>SSR RELAY
>Solid State Relay Relay

>>1549572
diagram sucks/10
only guessing at the pinouts of the controllers, and assuming the purpose of the switch is manual heat-on, but I think you want to do what's in yellow (which you can treat as a switched red): connect both of the controller #2 together on one side of the switch and mains red to the other side of the switch
>hand lotion for mold release
>tearing up mold by using nippers to remove the part, instead of a pin through the very convenient hole in the back
>clamping mold with bolts
>no insulation or heat management anywhere
>A HEAT GUN to heat the otherwise unheated
>BRASS NOZZLE
cringe

Is it safe to assume that mosfets only heat up when they're in the linear region? Also, is there an online calculator for mosfet power dissipation given vgs, Id and Vds?

>>1549632
no, MOSFETs will also dissipate heat in saturation according to Rds * Ids
no, there is no online calculator for that because Rds is also a function of geometry. read the datasheet

>>1549630
Ah yeah that makes sense. Here's a recolor that clarifies things some more.

>>1549569
>webum
I swear to god that people that are suposed to make regulations have no fucking clue of what they are doing
>country has both 220 and 110 V
>most homes and buildings lack grounding
>90% of the plugs in place could fit either the two prong round or rectangular plug

they go full br
>lol hurr dur think of the children, they might stick neutral on their tongues and lick the socket!!11!
>make it a non simetrical recessed piece of shit
>safety first lmao, u don't want your house burning down rite?
>make a 10A and a 20A plug, the 20A plug has thicker prongs (because breakers don't work rite)
>instead of making a 110v and a 220v plug because they will NEVER manage to make the whole country 220v
>only the swiss have a plug that is ressemble this monster
The brazilian government is the only credible evidence in favor of ancaps. Instead of making something simple and common, they make something retarded and without backwards compatibility with any of the 12390 plugs we had. It doesn't surprise me tho. It's literally the world's largest producers of "estabilizadores de tensao", which are isolation xformers sold as devices that protect appliances and "correct" voltage.

>>1549572
That is a very bad diagram. You should NEVER leave ambiguity or confusing things such as wires with the same color without names or numbers.
>but it's obvious
Maybe, but not for you a week later or some retard that finds it later and ends up burning something.

>>1549646
>>1549630
>>1549622

thank you I was concerned I was a brainlet but this shit is stupid.

>>1549638
is what we can assume the blue live wire really going directly to the heating element?
And that new brown wire joining both #2 to the switch. Is there a proper way to connecting both to said switch? Like the 2 wires being joined and creating a Y to the switch or just independently connected directly.
I know how that "Y" shit sounds but idk how this is really suppose to go.

I really wish there was a better tutorial to wiring a PID temp or even a better diagram to begin with. I have to change this to a 110Volt system anyway.

>>1549658
>is what we can assume the blue live wire really going directly to the heating element?
Yes. You only need to switch one of the two as the heater needs a closed circuit to operate. If it wasn't obvious, the SSR switches the pink wire on and off.

>Is there a proper way to connecting both to said switch?
The temperature controllers require power to be fed to them. Since the only other wires they're connected to are the thermocouple and SSR, pins 1 and 2 must be the power wires. Since "1" is already connected to the blue (again, you only need to interrupt 1 of the wires to stop the circuit from drawing power), we can only assume that "2" is meant to be connected to the pink wire. So we just connect the dots and change the orientation of the switch. I've marked both the switched mains sets of wires in dark red, as they both turn to whatever voltage the pink is when the switch is closed. If I were wanting to turn the controllers on and off, I'd put a switch in the same position.

>>1549658
whatever's convenient for you. you might find terminal blocks like Pic related to be helpful. each pair of screws is connected, so you can put a wire or two in one side under the bracket and a wire or two in the other and just screw down firmly for a decently stable mechanical and electrical connection that's quite unlikely to come undone unless disturbed
you might want to find some hands-on help locally to look over your work, mains wiring isn't to be toyed with

>>1549636
So assuming the mosfet has a 5 miliohm rDS at 5v gate-to-source and it's conducting 10 amps, then it's 10 amps^2 x 5 miliohm = 0.5 watt or something like that?

What a good free PCB design MOOC I should take?

>>1549688
spot on, sir

>>1549688
Vgs won't matter. If you have a black box that has a voltage drop and a current through, it's power will be i*v always.
>i= charge/time
>volt = energy/charge
>i*v = energy/time = power

>>1549307
If the power supply is floating, why can't you bring one end to ground?
All you should have to do is hook your scope probe across the solenoid's coil, probe on one contact and ground lead on the other.

you only really need to worry about your scope's grounding when you need to make differential measurements directly on mains-powered equipment, where you can't pull any point you want to ground because it has a voltage with respect to ground.

Using two probes should have worked though. make sure both were DC coupled.

>>1549537
it does work actually. If each probe has a 1Mohm to ground, then making a differential measurement without connecting ground is just measuring the voltage across a 2M impedance. Subtract the channels (or invert and add) and you get the right reading.
only works with two identical probes, don't expect to get readings that make sense if you're using a 1x and a 10x probe even if you compensate.

>>1549700
Rds is a function of Vgs
lern2mosfet

>>1549713
He gave resistance and current. It could be your mom's ass and it'd be the same thing
Learn to read

>>1549666
>>1549673
thanks

i made this as an ideal version

I changed the power voltage and heating element voltage, not sure if the SSR need to be changed. This diagram makes sense to me, thoughts?

>>1549715
>>1549673
>>1549666

forgot pic

>>1549715
24 < 110 < 380

>>1549715
>>1549716
SSR should be fine since it's 24-380V, though a bit overkill voltage-wise and you'll want to check that it can handle the current of the new element. On that note, you'll need new heating elements with half the resistance in order to get the same heating power, lest you want to put twice as many of the existing ones in parallel. They'll pull twice as much current, which is why checking the current rating of the SSR(s) is a good idea. You'll want to check if the existing PID can run off 110V (it probably can). The PID may have a ground pin also.

Can't help but notice that you've only got one PID and SSR and thermocouple and are running the power to 4 heating elements from that, there a reason for this?

>>1549716
Oh and your relay might want to switch the live/red wire, not the neutral/black wire. At least the former would be safer. When dealing with water it's going to be a good idea to ground things and have them running on a GFCI/RCD to ensure no current is leaking through ground.

>>1549716
good enough
just consider how you want to join all those heater bands and indicate that on your drawing if you need to, e.g. Pic related
also you might want to think twice about a 110V version of this project unless you have outlets rated for the current you'll be pulling. most outlets are rated for only 15A (1800W) which may or may not be enough. you might also consider shortening and/or narrowing the barrel a bit

>>1549719
manual plunger-style injection molders tend not to need a bunch of temperature zones because shot-to-shot consistency is going to be shit anyway. even larger barrels will often be equipped with several heater bands in parallel per zone

>>1549726
I think there are some terminal blocks or wago strips or whatever that common all the wires, avoiding needing to run a bunch of straggly copper from screw to screw.

>>1549728
very true and fine idea

PID bathroom faucets when

>>1549730
not soon enough

>>1549719
ok so make sure the heating element has half the resistance of the original diagram

Well the original had 4 down a little more than a foot about, I want to extend a inch or too for more volume. I could get away with maybe 2 or 3 but the more heat means more consistent temperature. Plastic is an insulator so the more the better and I don't care for having a hotter nozzle.

>>1549722
I though id be safer than having live go directly to the heating element, but I guess having the neutral go through a switch instead means ground is less likely to be "used" in a functional sense.

>>1549726
I was thinking about doing that to a terminal box with the heating elements hooked to live and neutral. That actually why I left a spare terminal box on the pic by accident.

Thanks for talking me through this guys! I feel much better.

>>1549726
>you might want to think twice about a 110V version of this project unless you have outlets rated for the current you'll be pulling.

yea i might have to reduce the heating elements to 3, I just don't have 220 plugs where I can work only 110

>>1549748
It's not the resistance that matters per-say, but if you want the same power output you'll need different elements. The original diagram has 4 190W elements, each one designed for 220V so that's about 255Ω each. Simplify this to a single element and it's 760W of heating with a total effective resistance of 64Ω. To get 760W of heating with 110V, that's 16Ω effective (actually a quarter of the resistance, oops). This 16Ω could be split up among 4 64Ω 190W elements, or any other arrangement, and there's nothing particularly wrong with using a different power output, but I'm assuming something between 1kW and 500W is what you're after, assuming your SSR (and plugs) can take it.

Stupid question;
If a circuit has no passives (resistors, capacitors, inductors, etc), does that mean it's a bad/incorrect circuit?
I ask because the circuit I designed uses an AD595A (pic related), a thermocouple, an SSR, and a Teensy μC with appropriate wires hooked up to each component. It works for what I need it to do (oven controller), but it slightly bothers me by how simple and straightforward the circuit is.

>>1549796
I looked at the datasheet for that IC.
Aside from things like decoupling caps that you probably should have (but that a circuit will work without regardless) I see no problem. The AD595A is designed to not need external passives. Many ICs are like that actually, it's one of the big advantages of modern ICs, you can synthesize so many passives right on the die and end up with a better overall design.

>>1549797
I guess it makes sense otherwise not needing external passives for ICs. Just feels weird to me is all.
>Decoupling capacitors
Right, I forgot about those guys, for filtering out noise on the input voltage right?
I can throw them in there, my circuit feels too naked anyway. Is there a way to calculate how many and what values you need? IIRC; the more you put in parallel, in general will mean the smoother your output voltage (after the capacitors) will be?

>>1549804
Bigger does not mean better in this case. Higher capacity usually implies higher ESR and ESL, so beefy electrolytics are useless against high frequency switching noise. It is common to put a single aluminum cap of several microfarads on your voltage rail for low frequency bypass and one ceramic of 0.1 microfarad per each IC right next to them, to reduce the traces length between IC and bypass (and thus, inductance).

>>1549804
The aim is to put the caps as close to the rails as possible, the actual value you want is no more than a few nF in most low-power cases, so there's no reason to go for multiple caps if you're doing THT design.

Also do you actually have a CDIP package? If so that's cool beans.

>>1549804
> (after the capacitors)
A capacitor is not a bridge rectifier. It does not have "before" and "after" pins. Its damping effect goes in all directions, limited only by equivalent resistance and inductance of the traces, which is quite significant at high frequencies.

>>1549815
>Higher capacity usually implies higher ESR and ESL
Are we talking about an individual capacitor or an
I realize every capacitor has it's own ESR and ESL, but hooking them up a multiple in parallel will increase those values?
> It is common to put a single aluminum cap of several microfarads on your voltage rail for low frequency bypass
Sorry for asking you to spoonfeed me here, but I just want to understand what's going on here. The large capacitor smooths out AC noise coming from your DC source, correct? How do we know how to size this particular cap? Several μf, sure, but is there some sort of calculation or just a general rule of thumb? Would I just have to experiment with capacitor values until I see a "smooth enough" voltage output on my scope?
> and one ceramic of 0.1 microfarad per each IC right next to them, to reduce the traces length between IC and bypass (and thus, inductance)
Again, I'm asking for a spoonfeed. Is there a reason why it needs to be 0.1 μf? How is it reducing the inductance between IC and the bypass?

>>1548254
It works beautifully.
I hope i get a promotion and a raise.

>>1549796
>>1549804
For those interested, here's a crude sketch of my circuit that I'm using. Currently just have it wired up to a perf board. Eventually want to depart to a PCB

>>1549831
>Are we talking about an individual capacitor or an
Wow. Didn't even finish my thought;
Are we talking about an individual capacitor or a bunch of capacitors hooked up in parallel?

can you recommend a good crimping tool from chinks on ali?
they have lots of them but they have useless shit like screw shortening and such, i want one that has support for as many terminal types as possible

>>1549831
I recommend watching Dave EEVBlog's video on the subject, and also looking through the comments to find that one guy saying "you can't just put caps in parallel and expect it to work flawlessly".

>>1549837
Yeah I'd add a cap onto the ADA's Vcc as PWMing the SSR might have a tendency to drop the rail for a fraction of a ms. The Teensy presumably has its own onboard caps.

>>1549834
Cool shit anon, always nice to see someone finish their project. But you'd better ensure it's nice and secure, don't want anyone getting into it like you hear happens to IOT devices occasionally.

>>1549818
It is a primarily THT design, that much is true. Though I guess I'm having trouble understanding how these capacitors are doing their job and why their values are selected as such.
I think it only comes in a CDIP package as far as I'm aware. Bloody thing was expensive too.
>>1549819
So if I were to probe right between the source and the bypass cap, I would still see the smoothing effect?

>>1549845
>don't want anyone getting into it
it is almost impossible to hack because it is not connected to the internet so any hacker would actually hop on a plane and fly over here and then manage to break WPA2 encryption to get in, so that is secure enough for a thing literally just pushes a button

>>1549796
yup, we've gone from cookbook circuits to just-add-water circuits, thanks to the magic of excellent transistor matching on a single chip

>>1549804
CMOS circuits will draw great spikes of current when switching, and devices that deal with millivolts (like thermocouple sources) will see ground bounce
datasheets will usually recommend an appropriate decoupling scheme. since that DS doesn't, the customary 100nF is probably enough, but you might further consider running a ground trace back to the analog ground of your MCU that's separate from any digital or power ground

>>1549842
if you can find a clone of the Molex 63811-1000 service grade crimper, you'll be able to crimp lots of small connectors
otherwise, you'll want one with changeable dies, and possibly the ability to buy additional dies (for like $6 each) to crimp a new style/series of terminal
that said, it's not necessarily easy to find a *good* die for "Dupont" terminals, for which I use a separate Plato SN-01BM, while not being absolutely sure it's the "recommended" crimper, it does give pretty good results once dialed in

>>1549849
It's on a private network? In that case you're probably golden.

>>1549848
If you're THT then there's no point in adding really small caps as their series inductance will be too high for what they do. Watch the video https://youtu.be/BcJ6UdDx1vg, I'm an amateur compared to him. Also the look for the comment by "hugo schmeisser" which proves he's also a little amateurish.

>So if I were to probe right between the source and the bypass cap, I would still see the smoothing effect?
At that point your probe wires might start making trouble, though I'm not sure. It would be tantamount to trying to measure the inductance of the wire from the source to the bypass cap, which I doubt you'd have much luck with.

>>1549852
>It's on a private network?
kind of.
the device it self acts as a router to which a client connects. basically think of it as a router with no wan connected

>>1549173
I'm more interested in cheap prices than MADE IN AMERICA. So I'll pick chinkshit. Local economy is always overpriced.

>>1549851
>you might further consider running a ground trace back to the analog ground of your MCU that's separate from any digital or power ground
I never quite understood this concept. How do you run a separate ground trace for analog? To me, that just sounds like a floating ground, essentially a loose wire hanging off analog gnd. If I'm tying all my other grounds together (digital, power, etc), where the heck is my analog ground going?

>>1549862
A wire going out the window, tied to a metal rod hammered into the ground.

>>1549862
it gets tied in with the other grounds, at a specific point, preferably at the power inlet of the board or some other point that doesn't bounce much. it's the same principle as a Kelvin connection: one connection for current, one for voltage. power ground carries all the current in the system, AGND ideally carries none. here's a shitty sketch. the idea is that current flowing to other loads through the ground plane, whose resistance is small but not zero, won't skew the reference zero input of the TC amp with respect to the reference zero input of the MCU, and therefore won't skew the measurement
(assuming the MCU has a differential input and/or a voltage reference that is set with respect to AGND)

>>1549862
or, think of it as breaking a ground loop

With Easy EDA being pretty decent, is there any reason to download Eagle or KiCad, the 5.7gb behemoth?

>>1550026
> With Easy EDA being pretty decent
Except that it isn't if you're going to do anything more complex than a blinking LED. The worst part is the component libraries. They are of inferior quality.

>>1550026
>5.7gb
>behemoth
What is this, 2001?

>>1549606
thank you

>>1550052
False


>>1550066
Eagle 7.6 is less than 300 mb, 9 is under a gb (why is it 3 times larger than 7.6?)

>>1550096
>bluepill and ESP-01
why

>>1550122
Because cheap + wifi-capability courtesy of ESP3D

>>1549861
For small stuff oshpark is often pretty close to chink prices.

>>1550194
mainly because chinks generally charge one price for anything under 5x5cm, and also shipping is separate. pretty sure chinks are cheaper and faster for anything significant over two layers

>>1550096
>why is it 3 times larger than 7.6?
libraries, maybe
also, Autodesk

>>1550026
if EasyEDA handles your use case, go for it

>>1550244
The library folder in 8.3 is 5.29mb, in 7.6 it's 164mb

>>1550244
>and also shipping is separate. pretty sure chinks are cheaper and faster for anything significant over two layers
>tfw shipping within China for 4.2kg worth of PCB's is only $9
>tfw removing just a single component in a pcb design yields a $12 cost reduction for 1000 boards assembled by PCBway

Tomorrow I'm going to suggest the "nip the capacitors until it stops working" technique to a customer and see how much money he'll save in total

>>1550253
then definitely Autodesk

>>1550254
based Earl Muntz
>$0.012 each
did you try eliminating BOM lines rather than placements? it costs goo money to have Lee check out that extra reel from stock

>>1550289
I'd assume that too, but it doesn't change the price. Maybe something is wrong with their pricing system.

>>1550320
hmm, could be that they bury BOM line price in the placement price, or just have enough margin that they dgaf

>>1550244
>if EasyEDA handles your use case, go for it
It's pretty great, actually. The workflow is like:

>pick part from LCSC
>copy the LCSC part number into EasyEDA library search
>there's magically a part with the correct footprint and price
>insert it into the schematic
>update layout and there it is

Like magic. In Eagle I have to look at the datasheet, import a footprint and do everything manually. I'll finish maybe one or two projects in Eagle that I still have to, and then uninstall it. Plus, EasyEDA lets you draw bigger boards than the free version of Eagle.

>>1550333
oh, I wouldn't recommend EAGLE to anyone. KiCAD drank their milkshake

So noob here, I've been thinking about my next project. pls give me your 2 cents
>A H-bridge with some beefy mosfets that can handle a lot of current and hundreds of volts (I have several of them from scrap)
>duty cycle control and frequency control (idk how many decades yet)
>inputs for optional external control of the mosfets
>Probably using a uC, so besides control I could record some of the output/input data
>used to test future projects, stuff like power supplies, motors, some capacitive loads
Anyone has something like this, or ever found need of something similar? What are some caveats and tips I should be aware? I still have to decide if I put in a power supply with it or make a small board that is to be used with a bench supply.

>>1550371
Just controlling without feedback would be a piece of cake, but what sort of feedback would you want such that you could record data? Current or voltage feedback?