PT54 - 13A LDO Current Source

[justinjja]

Designed a PSU for running the PT54 Led at full power,
Here it is:

Important:

OP Amp - Needs to be rail to rail, or at least have a good voltage swing
it also needs to be able to operate properly at ~3.5v

Mosfet - Needs to have a low RDSon at ~3V Vgs

Resistor - the 0.02 Ohm resistor needs to be a big 5W resistor


The high current bypasses the switch in this circuit,
I'll be doing this in a maglite,
I will post some pictures as I work on this

 

[phenol]

Your schematic drawing skills are commendable, to say the very least ...

anyway, scope the output for oscillations. the op amps may not feel comfortable driving large capacitors (Cgs of mosfet) directly...i'd plug a 300r-500r resistor in series with op amp output

 

[Bionic-Badger]

I would really suggest looking for a DC-DC converter if you want to power those things. Having low Rds_on doesn't make a lick of difference when you're using the MOSFET in linear mode, as you're purposely utilizing the MOSFET as a resistance. I tried essentially what you're planning for one of those PT-54s and the MOSFET started smoking the board because it was generating so much heat, and that was at only 8A. Also, should you not have full control of the input voltage, the excess voltage will drop across that MOSFET and can cause the power dissipation could be much higher than you expect.

Try checking out some of the high-current power modules from Texas Instruments (e.g. PTH05010W or PTH08T220W). They're only about $20 on Mouser.com and have pretty good efficiency. You can probably set them up for constant-current mode by placing a tiny resistance in the path of the load, and then provide feedback via an amplifier to the analog input pin, or some other scheme.

On another note, that diagram is hard to decipher. Please try laying out the circuit in some circuit design software. Even some PCB design software like Eagle would work.

 

[Meatball]

^^That very thing was done here by this guy:
http://laserpointerforums.com/f42/pt-121-testing-73397.html

It does indeed put the module into an actual constant current mode of operation, and the module requires much less heatsinking than that giant block of 20W of heat.

I don't think MOSFETs are really all that handy in the triode mode. They get hot and are not super responsive as they are a somewhat capacitive load to whatever is driving the gate.

 

[phenol]

Any active device in linear mode will get hot provided it wastes some power. At any rate, the SOA of the device must not be violated. Also, the thermal contact between the tab and the heat sink becomes a critical factor when the part has to dissipate large amounts of power. Moreover, mosfets come in a big variety of packages, many of which have superior power handling properties than the widely used TO220

 

[justinjja]

Your schematic drawing skills are commendable, to say the very least ...

anyway, scope the output for oscillations. the op amps may not feel comfortable driving large capacitors (Cgs of mosfet) directly...i'd plug a 300r-500r resistor in series with op amp output

Not a bad idea, Ill probably throw in a current limiting resistor on the gate.

I would really suggest looking for a DC-DC converter if you want to power those things. Having low Rds_on doesn't make a lick of difference when you're using the MOSFET in linear mode, as you're purposely utilizing the MOSFET as a resistance. I tried essentially what you're planning for one of those PT-54s and the MOSFET started smoking the board because it was generating so much heat, and that was at only 8A. Also, should you not have full control of the input voltage, the excess voltage will drop across that MOSFET and can cause the power dissipation could be much higher than you expect.

Try checking out some of the high-current power modules from Texas Instruments (e.g. PTH05010W or PTH08T220W). They're only about $20 on Mouser.com and have pretty good efficiency. You can probably set them up for constant-current mode by placing a tiny resistance in the path of the load, and then provide feedback via an amplifier to the analog input pin, or some other scheme.

On another note, that diagram is hard to decipher. Please try laying out the circuit in some circuit design software. Even some PCB design software like Eagle would work.

Dc/Dc converters are too expensive for this build,
With DTR basicly giving away The pt54's it can make for a cheap and impressive build.

As for the mosfet overheating, It wont be an issue as long as I heatsink it properly
Even worst case senario with a fully charged battery,
it will still be under 10W, and mosfets are oftem designed to disipate 100's of watts.

^^That very thing was done here by this guy:
http://laserpointerforums.com/f42/pt-121-testing-73397.html

It does indeed put the module into an actual constant current mode of operation, and the module requires much less heatsinking than that giant block of 20W of heat.

I don't think MOSFETs are really all that handy in the triode mode. They get hot and are not super responsive as they are a somewhat capacitive load to whatever is driving the gate.

Try again lol, the 20W of heat is the LED,
it doesnt matter how it is driven it still produces lots of heat

In the link you posted the led is making 140W of heat.

 

[Meatball]

Right, and its on a damn huge heatsink that's got a damn huge fan! The point of that post is that the driver is not having to dissipate much at all. I'm sure it could help with the battery life of the build. Heck, I'll send you one of those boards myself if you want one. I've got 5 or so!

 

[justinjja]

Sweet, Got it working this morning!

Dont have a heatsink for the mosfet yet :eg:
After running it at 12A for about 30 sec one of the wires going to the mosfet desolder itself

Good to know the mosfet is tough as nails, re-atached the wire and its still working lol... time to go find a heatsink