2v 28A high power diode laser supply based on old PC AT PSU / General Science and Electronics / Forums

Forums

4hv.org :: Forums :: General Science and Electronics

« Previous topic | Next topic »

2v 28A high power diode laser supply based on old PC AT PSU

1 2 next

Move Thread

LAN_403

Carbon_Rod

Fri Aug 17 2007, 03:02AM

Registered Member #65 Joined: Thu Feb 09 2006, 06:43AM
Location:
Posts: 1155

The ATX mod thread is very extensive, however the 20A @ 5v is not enough to maintain the needed threshold current. (ATX power ratings often share with the 3.3v and 5v)

The old DELL PC 230watt AT based supply is rated for 32A @ 5v, and I plan to drive it around 20A +- 3A at 2v +- 0.3v (depending on temperature).

Originally I had opted to try modifying the KA339A comparator input to drop the 5v output voltage (extend the range of +- 0.3v). However, the supply had a failure and continuously blew the fuses even after the comparator mod was undone. Obviously fried something, but as I have a stockpile of these things it was easier to try another design then a repair.

So I decided to try something like this schematic (ignore the 3.3v line, diode count, and R1 may change depending on operation data.)

Has anyone tried this arrangement for testing high power diode lasers?
(will the output voltage remain stable upon powering on with the Power Good signal wired on.)

[Edit: Width=400]

...

Fri Aug 17 2007, 03:24AM

Registered Member #56 Joined: Thu Feb 09 2006, 05:02AM
Location: Southern Califorina, USA
Posts: 2445

I don't quite see the point in using diodes to drop voltage when you have a resistor available to do it (last time I checked, nichrome was way cheaper than silicon...), and your resistor is way to big. I would expect the diode to run at at least 2v (when the resistance of the wire and connections are factored in, and the operating voltage drop of the diode--laser diodes definantly do not follow the 60mv old wives tale) ... With that in mind, your .1r resistor should be just about perfect to draw juice strait off the 5v bus, assuming that the diode will drop 2v and your supply is putting out exactly 5v (will probably sag a bit especially if you are using the stock 16awg wire), that leaves 3v to drop across the diode, which will take 30a.

Although I personally wouldn't trust a laser diode that I paid much money for (then again I paid about $20 a pop for my 50w bars) againts a freebie atx psu, you should be fine. You might try adding a nice fat, low ESR, cap directly across the diode, after the resistor, that is permanantly, 100% absolutley never going to be disconnected from the laser during operation (or you will find yourself the owner of a DarkEmittingDiode laser), to help round off the inevitable spike as the supply tried to cope with the sudden 30a draw.

Carbon_Rod

Fri Aug 17 2007, 04:44AM

Registered Member #65 Joined: Thu Feb 09 2006, 06:43AM
Location:
Posts: 1155

My guess-timate is its going to be about 23A @ 2.2v before adjusting the PSU output. The OEM does not publish the SOA curves so any example docs would be great (they really want you to buy the $6k supply with the TEC control.)

The current and voltage need to be regulated, voltage will be held constant, and R will be varied or upgraded for constant use if it operates as expected.

ESD protection is a moot point for diode bars ("Dark" emitters can be repaired sometimes by severing the offending units lead from the array if you are skilled with a razor.)

As for the cheap PSU, I rarely talk about my projects here... The short answer is I may require 5 driver units.

Regarding the YAG we were discussing in the other thread, you may find this interesting.

Cheers,

...

Fri Aug 17 2007, 05:27PM

Registered Member #56 Joined: Thu Feb 09 2006, 05:02AM
Location: Southern Califorina, USA
Posts: 2445

You still need to be extreemly careful about ESD and rogue cap discharges with large bars, as it is still very possible to fry the whole bar. Even if you only manage to take out one emitter, you can't just cut the wirebond going to the offending emitter, as the GaAs substraite is quite conductive, to 'repair' a dead emitter it is necessary to grind through (keeping in mind that you are working with a bar that is about 50% bioavailable arsenic) the entire bar and remove the emitter completly. Or unsolder the bar, cleave the ends next to the dead emitter, and then solder it back onto the block.

Those 'donut' arrays are nice, but they cost more than you can imagine. I have yet to see one come up surplus ever, but the linear arrays of 5-20 bars go for thousands even surplus. Cooling is also very tricky with them, you can't just bolt them down to a waterblock and hope for the best.

Carbon_Rod

Sat Aug 18 2007, 12:47PM

Registered Member #65 Joined: Thu Feb 09 2006, 06:43AM
Location:
Posts: 1155

The impedance of these kinds of lasers is not linear, however even the CW designated units are profiled using pulse mode testing (negligible heating.)

Driving these things without a TEC PID controller is easier. The problem is going to be with the fundamental output wavelength that will vary unpredictably during operation.

Additionally, I don't think we have the same style of device. The unit in question has several gold micro-wire jumpers from a ceramic board to each of the micro lens coupled output clamped diodes. Removing the arrays' lens would cause misalignment and incident rays that could cause more damage.

I am also looking into a classic crowbar protection scheme on the supply's output to shunt any spikes (should be fast enough.) Yet its unlikely going to be necessary

...

Sat Aug 18 2007, 04:18PM

Registered Member #56 Joined: Thu Feb 09 2006, 05:02AM
Location: Southern Califorina, USA
Posts: 2445

Carbon_Rod wrote ...

Additionally, I don't think we have the same style of device. The unit in question has several gold micro-wire jumpers from a ceramic board to each of the micro lens coupled output clamped diodes. Removing the arrays' lens would cause misalignment and incident rays that could cause more damage.

I am pretty sure that the laser bar itself is the same as the ones I have, does your laser have a single bar about 1cm long (which has multiple emitters in it), or a bunch of individual emitters?

Carbon_Rod

Sun Aug 19 2007, 08:16AM

Registered Member #65 Joined: Thu Feb 09 2006, 06:43AM
Location:
Posts: 1155

"does your laser have a single bar about 1cm long (which has multiple emitters in it), or a bunch of individual emitters?"

The unit in question is an odd one as it has bonded leads and is relatively new looking (see drawing to compare with a known module.)

In order to confirm physical isolation it would need destructive testing. However from a strictly non-invasive perspective there appears to be a physical separation cut in every few wires.

It is unlikely going to endure the driver design process, but it can easily be replaced with one of the newer 19 fiber newport units.

...

Sun Aug 19 2007, 03:26PM

Registered Member #56 Joined: Thu Feb 09 2006, 05:02AM
Location: Southern Califorina, USA
Posts: 2445

The picture on the right is s standard bar soldered to a block, and the previous comments about the emitters being connected apply. If the one on the left ha 19 emitters, you can be 98% sure that it is also a normal bar (as they all have 19 emitters).

Steve Conner

Mon Aug 20 2007, 08:34AM

Registered Member #30 Joined: Fri Feb 03 2006, 10:52AM
Location: Glasgow, Scotland
Posts: 6706

carbon_rod wrote ...
It is unlikely going to endure the driver design process

Wouldn't it be a smart move to test your creation on a big regular diode first?

You can use a scope to detect any spikes you are making, before you actually hook it to a laser. Preferably a fast DSO, but any scope would be better than nothing.

I used methods like this when designing laser and TEC drivers, and I am proud of the fact that in the whole 5 years I worked on them, I only ever killed one laser (and that was an out-of-spec one we kept around as a guinea pig for new designs)

A big wirewound resistor would probably be better than a string of diodes for dropping the voltage. The laser drivers I designed used synchronous buck converters with average current control, but if you just want to test things, a resistor is a lot easier :P

Carbon_Rod

Mon Aug 20 2007, 01:15PM

Registered Member #65 Joined: Thu Feb 09 2006, 06:43AM
Location:
Posts: 1155

I plan to move the current regulator to the front of the series as the required zener values were not in stock. However, the diode string seemed like a good idea for several reasons:

1.) It has been documented by other people to work with bench-top supplies.

2.) The large diodes have internal capacitance (about 15 nF each). I thought perhaps it may reject any turn-on spike with a few ns of blocking... But, a fast cap and MOV could make sure... I'd also hoped that the setup may act a bit like a primitive soft-start device as they are comparatively slow.

3.) The lasers max current is greater than the supply can offer... so as long as Vmax is kept stable it should survive for awhile.

4.) I do not need tuning capabilities for this design yet, but I had considered augmenting the output stage of a $3 precision driver/regulator chip.

5.) Adjustable current regulation means it can support smaller laser diodes of the same make for testing various designs. I would wager if the use of R is limited it should scale up nicely without PTC or large L issues.

Steve, do you know if the safety clip on Newport diode bricks are still present in their block modules?

=)

Edit:
Wow, found someone who did something similar and sells 19watt systems on ebay =/
http://4hv.org/e107_files/public/1187937087_65_FT1630_b33a_1.jpg
http://4hv.org/e107_files/public/1187937106_65_FT1630_b65b_1_.jpg
http://4hv.org/e107_files/public/1187937125_65_FT1630_b88e_1_.jpg
http://4hv.org/e107_files/public/1187937146_65_FT1630_bac4_1_.jpg

Looks like a kid built it, but shows it can be done =]

1 2 next

Moderator(s): Chris Russell, Noelle, Alex, Tesladownunder, Dave Marshall, Dave Billington, Bjørn, Steve Conner, Wolfram, Kizmo, Mads Barnkob