[HELP]High Power 808nm driver

[Icephoenix]

Hey guys,
I've bought a 8W 808nm c-mount laser diode and some protection goggles for a pretty cheap price, but i cannot find any driver suitable of outputting 8A at 2 volts as the datasheet of the diode says.
As i always wanted to build electric devices on my own i just made this little schematic of a constant current driver which should work if i remember my physics lesson correctly.

Ok, i tried to improve the driver according to your feedback, this is what came out.

Im still having the problem to decide which mosfet to use for pwm modulation, it has to be a logic level mosfet which is pretty fast.
I'm also curious how clean the output of the driver will be.
If i calculated correctly, the Transistor regulating the voltage has to have an amplification factor of 40 (200mA Base current and 8A output). As i want to feed it with a 3.3V power supply (2V output) i cant use a darlington transistor
Should i use 2 Resistors to get the reference voltage (as a PC power supply is already stabilized
Does anyone know a single AND gate IC?

Ohh, and because i failed with EAGLE again, R11 needs to be 0.0875 Ohms

 

[Cyparagon]

Yeah, that'll work. Except the relay isn't needed. Q1 will turn it off. You have your relay backwards, by the way.

You should label all transistors with Q, as T usually designates transformer.

What is the purpose of D2? Reverse protection?

T2 will need to be BIG and will also need a LOT of heat sinking. A 5V >10A power supply should be fine as the power source.

 

[Icephoenix]

well, im using the relay because my microcontroller cant set its pwm output to 0, thats why i thought of using a relay to turn it off
the diode D2 is a zener diode setting the output voltage. The maximum output voltage should be the zener voltage of D2 - 0.7V (T2) - 0.7V (R11)

Edit: I will be using a PC power supply so i have both 5V and 3.3V if i find a driver which is more efficient

Edit 2: Another important reason for the relay is that if there is no modulation input the driver should be turned off

 

[Icephoenix]

Ok, i tried to improve the driver according to your feedback, this is what came out.

Im still having the problem to decide which mosfet to use for pwm modulation, it has to be a logic level mosfet which is pretty fast.
I'm also curious how clean the output of the driver will be.
If i calculated correctly, the Transistor regulating the voltage has to have an amplification factor of 40 (200mA Base current and 8A output). As i want to feed it with a 3.3V power supply (2V output) i cant use a darlington transistor
Should i use 2 Resistors to get the reference voltage (as a PC power supply is already stabilized
Does anyone know a single AND gate IC?

Ohh, and because i failed with EAGLE again, R11 needs to be 0.0875 Ohms

 

[Speedy78]

If you get this working I would like to pay you to build me one haha

 

[Icephoenix]

why "haha"?
do you think it doesnt work? if so tell me why it should not work

 

[Speedy78]

I just need one for a pump diode

 

[Cyparagon]

You've got your mosfet upside down. You'll get less than 120mA base current with that value of R10, not 200mA.
Vgs might not be high enough unless your microcontroller is operating at 5V and your driver at 3.3V.
You may want to look into sinking the base current away with the FET instead, or even switching to P-MOS.

well, im using the relay because my microcontroller cant set its pwm output to 0

Sure it can. You just haven't found the right code/command to use Yet.
Arduino for example would be a simple DigitalWrite(pin, LOW); Keep looking.

Another important reason for the relay is that if there is no modulation input the driver should be turned off

So put a pull-down resistor on the gate. Problem solved

 

[Icephoenix]

Well the manual that came with my microcontroller said you cant put pwm to zero. Microcontroller is running at 5V and i want to run the driver at 3.3V to keep the power loss over the resistor as low as possible.
I'm thinking about using a driver for the mosfet because I think the current might be to high for the microcontroller.

 

[phenol]

you could replace the pass NPN transistor with some mosfet like IRF3205 for instance, connect the gate via 1k /or more, or less, depending on how fast you want it to turn on/ to 12V and a low power n-ch mosfet (2n7002 or similar) from gate to ground. turning it on will short the gate of the big one to gnd and disable the laser. turning it off will bias the gate 4-5v more positive than the source. the separate gate supply voltage will let you use very low voltage at the input side (drain) for a very low dropout solution...

 

[Cyparagon]

^It won't be much different. That 0.7V on the sense resistor accounts for almost all of the drop in either case.

Well the manual that came with my microcontroller said you cant put pwm to zero.

So they're saying PWM_SetValue(0); doesn't work, But the pin output can surely be set to "off".

 

[Icephoenix]

Yeah, and if not i'll just buy another microcontroller... Tomorrow i'll remove the and gate. Ok, only Problem left is that i'm still missing a transistor with an amplification factor which is high enough. I think i'll just use an other method to get my reference voltage. If anything fails it will still run at 5V

 

[phenol]

dont bother using darlingtons or regular npn. first off, driving them is not as easy as driving a mosfet(provided you have enough voltage headroom to bias the gate). second, while it is true that the loss across the shunt will remain 0.6v, the mosfet can stay in linear mode with much lower voltage across DS than bipolar transistors. typical VCEsat is couple of volts. the bottom line is that it will not source 8A with 3.3v in and waste more power with 5v in.

 

[Icephoenix]

Yeah but how do i drive a mosfet with 3.3V? Or even 0.7V? Ok, lets say i use a +12V voltage just to power the mosfet how do i limit the voltage? With a npn the voltage cannot exceed the base voltage -0.7V (-1.4V for a darlington transistor)

 

[phenol]

something like this. you need 12v just to bias the gate with respect to the source, the mosfet sources current from the low voltage input. The 2n7002 mosfet grounds the gate if you apply 3-5V from the microcontroller port to its gate. No gate driver is needed. the down side is that the diode will be powered up while the mcu is initialising its ports. if you turn off the pass transistor too rapidly, some brief current undershoot may appear due to the finite reverse recovery time of its body diode...hence the 470ohm resistor to slow it down somewhat...
then again, this is not an awfully stable design due to VBE drift of the npn current sense transistor...i think it should be kept away from sources of heat like the pass transistor and the shunt resistor. you could use it as a negative thermal feedback--current will roll off as temp creeps up...

 

[Cyparagon]

What mister phoenix is concerned about (I think?) is the current that the gate will require to charge is too high for the controller. (M2 in your example) This is a valid concern since the gate is analogous to a capacitor and acts as a short for a few microseconds. A simple ~150 ohm resistor on the gate will lower the charge current to more manageable levels.