[ 3 / biz / cgl / ck / diy / fa / ic / jp / lit / sci / vr / vt ] [ index / top / reports ] [ become a patron ] [ status ]
2023-11: Warosu is now out of extended maintenance.

/diy/ - Do It Yourself

View post   
File: 20 KB, 348x270, LX650G.jpg [View same] [iqdb] [saucenao] [google]
22075 No.22075 [Reply] [Original]

I need to drive a circuit at 30VDC and about 10.2A.

Problem is I need this to be a very small driver (not including the heat sink, which is what the casing of my device will be used for) and since I am apparently too weak in my google-fu to find one (closest I get is a 100w) it looks like I'm going to have to build it myself. Problem, I suck at electronics.

Any halp?

>> No.22505

30V is pretty uncommon. Maybe you can use two 15V/150W drivers in series? Also: 300W is a lot of power. Even with good drivers you will end up with about 30W of waste heat. I doubt that passive cooling will suffice...

>> No.22633

>>22505

You cannot run drivers in series, and I'm assuming OP is talking about a single input.

>> No.22685

You can connect linear regulators in parallel as long as each has it's own decoupling caps and a diode to stop them driving each other backwards. It's a bit quick and dirty, and you'll need a fair bit of heat-sinking but it's a pretty easy / cheap solution.

You'll need to get yourself a transformer you can rectify to >30v (a significantly over-spec'ed 24v one could probably pull it off)

>> No.22963

>>22685

Let's say I took three of what's in my OP pic.

How would I go about hooking those up? Could I do this directly on the DC output/return?

>> No.22984
File: 5 KB, 241x200, linear reg.jpg [View same] [iqdb] [saucenao] [google]
22984

I was thinking more like these thing, strapping 10 or so to a (substantial) heat-sink.

>> No.23035

>>22984

Gonna need 10A worth.

Why not an LT1038, OP?

>> No.23067

For the brave/stupid, what about modding a PC power supply?

>> No.23126

>>22633
>>22685
There is correctness in both of these (seemingly) contradictory posts. Note that BJTs are very vulnerable to thermal runaway because they conduct more current at higher temperatures.

If you had two or more drivers that are attempting to share the load and they are using BJTs instead of MOSFETs, the one bearing the most load will heat up, and in turn conduct more current, which makes it heat up more, which equals fire.

What>>22685
said is true, but the diodes won't prevent one from bearing more load than the others, which could lead to a shorter lifetime for MOSFETs, and fire for BJTs.

You can build current and voltage regulators, they aren't too difficult. I've also tracked down some fairly high amperage logic level MOSFETs for cheap - getting one that'll handle your current load isn't hard. You will need a large heatsink, though.

>>23067
PC power supplies can easily deliver enough current, but don't put out more than 12 volts. If you wanted a greater voltage, you'd have to change the transformer.

>> No.23153

>>23126

Ballasting the Vregs with low value (<1 ohm) resistors would help equalize the load, at the expense of accurate voltage output.

>> No.23195

>>23153
Too bad OP didn't tell us what he was doing, eh?

>> No.23218

>>23195

Nothing that requires dead-on accurate voltage. I can be +/- 2V, I'm just making a huge LED screen for my ceiling and need power.

>> No.23248

>>23218

You couldn't change the design to run off an ATX PSU?

>> No.23255

>>23248

Nope, specific layout requires ~30V, clamped if possible, little headroom if not.

And the wires on a computer PSU aren't long enough, I'm not mounting a PSU to the ceiling.

>> No.23297

>>23067
Yep, for that kind of output you should build a switched PSU (or modify a PC-PSU).
It's nothing simple anyway, and quite dangerous to experiment.
http://en.wikipedia.org/wiki/Switched-mode_power_supply
http://en.wikipedia.org/wiki/Buck_converter
http://en.wikipedia.org/wiki/Boost_converter

>> No.23333

Take a look at some switching voltage regulators. Don't skip on isolating from mains! 2 matched wall wart transformers end to end is all it takes.

>> No.23336

>>23255

The problem with the Vreg solution is that your going to need to find quite a substantial transformer from somewhere (10 amps is quite a lot of current) if you live somewhere with 230 / 240v AC you could maybe run a microwave oven transformer backwards (they are about 10:1 IIRC)