10A Driver Help Please?

[AMooMooCow]

Hi! I'm having electrical issues and I figured who knows constant current more than LPF.
Background:
- mechanical engineering student in design project group with 3 other MEs with no EEs. I've been charged with all the electrical duties because I know how to solder. Professor always disappointed in group because I'm building everything and nothing's done yet cuz everything's electrical.
Why:
- We have to drive solar cells to produce IR emission. Need ~10A. Tested today using 12v 50A voltage source and almost started a fire twice.
Vf=~0.5v. PV cells act just like big LEDs.
- I'm afraid that if I just get a 7.5A linear regulator that it won't be enough current and the professor will be much more disappointed in me.
- Can't just use resistor because our results are a function of PV cell temperature
- I have about 20A left to spare
- We have to spend the just about the whole budget on other things

How can I double the current from 5A regulator like an LM338?

Thank you so so much for your time

 

[Cyparagon]

Are you allowed to put PV cells in series? That would make your project WAY easier. If not, you can simply put two LM338 in parallel. Just be sure to power them with 5V (not 12V) and add a lot of heat sinking to each. Be sure the tabs of the two are isolated electrically.

Ideally, a bench power supply would be best for this job, but that's ~$100 if you don't already have one.

 

[RedCowboy]

For 5 amps I have used a 1084 for 10amps I have not used it but a 1038 or 1581 should work.

LT1038 - 10 Amp Positive Adjustable Voltage Regulator - Linear Technology

http://www.linear.com/product/LT1581

 

[AMooMooCow]

Are you allowed to put PV cells in series? That would make your project WAY easier. If not, you can simply put two LM338 in parallel. Just be sure to power them with 5V (not 12V) and add a lot of heat sinking to each. Be sure the tabs of the two are isolated electrically.

Ideally, a bench power supply would be best for this job, but that's ~$100 if you don't already have one.

Nope, just one cell at a time. I've read that LM devices can't be connected as such because they'd share current unevenly here. Could I just ignore it without starting a fire?
That also raises a concern. (12v-0.5v)10A. I'd need to spend $10 on a converter anyhow

Ideally, could someone just attach a bunch of LM317s in parallel with enough heatsinks? I'm just afraid to cause an upset

 

[Cyparagon]

I've read that LM devices can't be connected as such because they'd share current unevenly

Correct. However, that's when connected as a voltage regulator. You're using them as a current regulator. They will share the current.

Ideally, could someone just attach a bunch of LM317s in parallel with enough heatsinks?

No. That's certainly not IDEAL because it's so inefficient. It'll work fine otherwise, if that's what you're asking.

 

[AMooMooCow]

Correct. However, that's when connected as a voltage regulator. You're using them as a current regulator. They will share the current.

No. That's certainly not IDEAL because it's so inefficient. It'll work fine otherwise, if that's what you're asking.

Thank you so much. This is wonderful news to me

 

[Benm]

Current sources and sinks can indeed be used in paralel whenever you want to, the total current will just be the sum of all connected in paralel - identical to how you can connect voltage sources (such as batteries) in series without any problem.

In theory you could just put about 10 LM317 based 1 amp current sources in paralel, but as mentioned this would be quite inefficient and require a lot of heatsinking for each one. The heatsinks also need to be isolated from the chips and/or eachother making the whole thing a bit impractical.

With LM338's you only need 2, but they require proper heatsinking as well. Make sure to calculate the dissipation considering your innut voltage and heatsink accordingly.

Also, not that these regulators have a built in protection system to protect them from overheating, so measure the actual current to make sure the protection doesn't kick in lowering the output current. This often happens rather gradually and could go unnoticed yet ruin your results if you're not aware of it.

 

[AMooMooCow]

:beer:
Here was my solution cc module
The literature said someone had provided 6A from 22v somehow; the literature is just very sketchy so that's why I was so afraid. It also said someone "used a 10A power supply".
This 8A should be enough, but if it's not I could stick another current supply onto it. This was cheaper and less messy than getting a stepdown converter and the regulators and heatsinks. Neatness counts for us too
Edit: But this will probably be useful for lasers in the future too. I couldn't find any thread yet describing someone connecting LM's in parallel

 

[Cyparagon]

But this will probably be useful for lasers in the future too.

The one I tested was okay for battery charging, but the huge current spikes on startup made it useless for most other constant current operations. It would probably also be unstable or unusable under 1.25V like most step-down converters. I guess you'll find out the hard way :undecided:

edit: yeah, I see the DC output voltage is rated no lower than 1.2V. I'm fairly certain your 0.5V load will make it quite unhappy, if it starts up at all.

 

[AMooMooCow]

The one I tested was okay for battery charging, but the huge current spikes on startup made it useless for most other constant current operations. It would probably also be unstable or unusable under 1.25V like most step-down converters. I guess you'll find out the hard way :undecided:

edit: yeah, I see the DC output voltage is rated no lower than 1.2V. I'm fairly certain your 0.5V load will make it quite unhappy, if it starts up at all.

Oh well. If it's no good I can just buy a 50w rheostat. But the professor's disappointment level will be at an all new level. I'll post if it blows up. These solar cells can take a beating too though. I connected 40A by accident once and it still worked afterwards

 

[Benm]

If the minimum output voltage is a problem, you can solve that by putting resistors and/or diodes in series with the load.

In case it is 1.2 volts, 2 silicon diodes will drop that voltage for you nicely and you can supply the current you want down to zero if need be.

At 10 amps these diodes need to be pretty beefy as they will be dissipating in the order of 10 watts each. A bridge rectifier can be used as a simple 2-diodes-in-series component though, and those are often fully isolated from their case so easily mounted onto a heatsink or even against the wall of a metal case.

As for the solar cells, i presume the damage would be mostly from overheating. This should not be a huge problem: When these things are 10% efficient under normal operation, they are taking 90% of light as heat and dissipating that. The other way around, they should in theory take 10 times the nominal electric output power in reverse without burning out.

 

[Isaac Clarke]

are you building a solar beam death ray?.....i'm confused.....

 

[Benm]

Just heating up solar panels in the dark i guess - by the time they start producing any visible light it's probably about time to call the fire department

 

[AMooMooCow]

We're observing cracking in solar cells due to fatigue in varying temperatures and correlating it with efficiency, then comparing it to simulations run in Matlab and Comsol. Since the cells act like LEDs, and emit IR from about 900nm-1300nm, we can see this electroluminescence with a cheap modified CCD camera (the real deal SWIR camera would cost ~$20,000US)

In this video, you can see dark areas where electrical continuity is interrupted due to cracks.

... Speaking of solar cell deathrays, that's an actual thing (link).

 

[Benm]

That's quite interesting really, i expected the wavelength to be a fair bit longer.

How efficient are solar panels in producing 900-1300 nm light? Perhaps this has useful application for them as IR floodlights to be used with security cameras at night an such?

 

[AMooMooCow]

You'd need a reaaaally big lens, like a watermelon. But even then I gotta put 5W in to get emission at forward current. So fragile too. We were walking with one of these and it snapped under its own weight. Not saying no one should try it, it would just be really expensive.
Edit: peak emission at about 1100nm