50A CC Linear Driver?

[Wolfman29]

Hey everyone -

So I am in the works of getting the 40W 808nm diode bar working... and basically, I need a 50A linear driver. Heatsinking won't be an issue, and I will have a stable voltage just above the linear driver's threshold voltage.

BUT. I can't seem to find anything that will run more than 10A or so. I think that I could use a bypass transistor in an LM317 type circuit, but my sims show it won't work :thinking:

Anyway, I'd appreciate any help any of you guys can offer.

 

[Meatball]

I've got the thread for you right here buddy. Its a switching design, but it should cost less than $20 since you can get the module as a free sample from TI. The schematics are somewhere in this thread - its only 2 pages

http://laserpointerforums.com/f42/pt-121-testing-73397.html

 

[Sigurthr]

You could parallel 50 LM317T's XD. That would be a sight to see!

 

[Cyparagon]

A typical op-amp current sink similar to this is almost infinitely scalable

 

[Meatball]

^^Are those circuits better with MOSFETs or darlingtons?

 

[Hiemal]

^^Are those circuits better with MOSFETs or darlingtons?

With a mosfet it would have very very low dropout voltage, since a mosfet doesn't have a voltage drop, but an on state resistance instead. Plus MOSFET's are also easier to find in higher current capability ratings.

 

[Cyparagon]

^They only behave like that when driven in saturation, which offers no control over the current and isn't applicable here.

Sure, you can find mosfets in much higher currents, but, again, it doesn't apply here. You'll need a volt or three of headroom for either, and so the limiting factor is power dissipation. This value often depends on the case type rather than the transistor type.

Mosfet would be slightly more efficient because there's no drive current while BJT is more stable, but you'd have to build a pass bank no matter your choice.

 

[Wolfman29]

I think I am going to go with the circuit Meatball linked. Thanks!

 

[Benm]

As far as the opamp circuit being scalable: It is, but it does require the use of an additional transistor (in a darlington construction) to have sufficient gain. With plenty of supply voltage it can be used directly by replacing the BD139 with a (bank of) suitable mosfets.

There is the practical problem of the sense resistor though: you should scale that way down along with the rest of things. If you are looking at 50 amps, the replacement for the resistor would probably have to be a length of wire rather than something you can buy as a component.

 

[Hiemal]

50A Current Shunt For AMP Meter Gauge PROBE [K304] | eBay

Something like this would work, I believe.

 

[Cyparagon]

75mV might not be high enough of a signal to regulate properly. A difference of 500mA through it would only be a difference of 750µV for example which is lower than the noise you might see in the circuit.