Let's Make Robots!

I need advice for a power transistor

Hi all,

Since DanM posted his laser blog I was hooked to the idea of pulse driving laser diodes.

I have created my own pulsing hardware using an ATMega328 to generate a 62.5 ns pulse at 16 Mhz, driving the diode trough an ULN2803 (Toshiba made) which seems to be able to respond well to such short pulses. 

However the ULN2803 is limited in power output so the highest current I could sink is about 700mA (I know the ULN2803 is limited to 500mA yes) by parralelling outputs. At this level the ULN2803 is way over its limits and starts overheating.

So to proceed with higher currents (I'm positive the diode can take at least 1A in pulsed mode) I'm looking for a NPN transistor with at least 1A/12V rating, faster then 60 ns on/off switching times which can be driven with 5V TTL logic .

Has anyone suggestions? I'm sure there are lots of options out there but I have a hard time finding the "right" one, I need an expert :)


Later edit: Thank you all for the input, I'm settling on a MOSFET, I'm sure I can locate some in the electronic junk I have at home to do some testing. If the laser diode still holds up at 1 A I'll mount it on the Phoenix - see if cutting is possible which such a low power diode in pulsed mode :D

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.

I'd make sure the FET has a fast enough transistion time!

Digging trough my jung pile usually means:

  1. - find the part 
  2. - lookup datasheet 
  3. - if datasheet == found && specs == good => good part -> desolder :D 
  4. - else goto 1

So yeah time waisting ... but with enough junk it usually works ;)

Thank you all for the input.

 I think I'm going the MOSFET way maybe paired with an lower power transistor, I'll do some experimenting as soon as I can. I'll let you know how it goes. 


If you need to do this asap, I have some serious power mosfets floating around at home, that can handle hundreds of volts at many amps, more than perfect for this application. Let me know I can mail you one for free.

But there is no rush. It's just yet another project of mine, I'm just investigating the possibilities. No need to send me anything, but thank you for the thought :)

There are several to choose from. Here's a page with a few. One question would be which ones are readily available at a reasonable price. http://www2.renesas.com/discrete/en/products/transistor/function/pwtr/pt2.html However, since you will be driving the laser circuit from an arduino (which only drives about 40 mA per lead) driving large current pulses with a single transistor stage is not going to work out.

Thus, I think you should take a look at;

 1) cascaded transistors,

 2) a darlington transistor, (same thing as #1, but in a single case)

 3) a thyristor/SCR/SCS,*

 4) an avalanche transistor.

 As suggested by others, you can also find 5) MOSFETs that should work. I took a quick look at the data sheet for the one OddBot suggested (IRF820) and it looks like a good choice.

*Thyristors are notoriously slow, not known for pulsing in the nano-second range, but there are some available that will be quite satisfactory. (I think I added a circuit or two using scr's in my drawings a while back and some with avalanche transistors. Without going back and looking I actually forget which circuits I posted, but I have some of each available in my notes if they are needed.)

Also, here is some background info (a .pdf file) from Osram that might prove informative as to laser driving.



According to my understanding of a SCR, they won't "turn off" until the forward bias drops below a certain point, which won't work for quick timings correct? He could always use a 2N3055, a big NPN transistor.

That is correct. If you need 'turn off' capability as well, you could use an SCS, or SCS-equivalent circuit. However, the circuits I am thinking of charge a capacitor through a resistor/transistor, etc. and then an incoming pulse fires the SCR* to discharge the cap. through the laser diode array all at once in one fast pulse. The circuit is set up so the current through the resistor in the capacitor charging circuit is below the current needed to keep the SCR on. Consequently the pulse obtained through the laser is just that which is stored in the charge capacitor.

The SCR here could be replaced by a transistor, a MOSFET, cascaded transistors in darlington configuration (selected for fast enough turn on/turn off as needed and to allow proper current at the V++ voltage) or an avalanche transistor, etc. etc.

Here is a simplified version:

The following shows the charge storage capacitor I am talking about in both the laser driver circuits.

[ PLEASE Ignore the top circuit (light demod) that was already in the drawing. ]

.. take care of that?

SCR's are really slow anyways, a MOSFET or a bigger transistor would work fine. A Zetex Transistor would work too. They are beefy little to92 transistors that can handle an amp.