High Current Laser Diode PSU / Projects / Forums

Forums

4hv.org :: Forums :: Projects

« Previous topic | Next topic »

High Current Laser Diode PSU

first previous 6 7 8 9 next last

Move Thread

LAN_403

Electroholic

Thu Feb 26 2009, 08:11PM

Registered Member #191 Joined: Fri Feb 17 2006, 02:01AM
Location: Esbjerg Denmark
Posts: 720

Hey matt, how is your flatpack hooked up on the input side, I managed to fry two smaller 75W models by turning them on too fast.

If you know the input is 5V and you know the current going thru the circuit is I, input power is then 5 * I.
Now if you know the voltage drop across the LD is Vd, diode input power can be expressed as Vd * I.
Which means the rest of the circuit i.e. FETs + sense resistors have to drop 5 - Vd that many volts.

Current sharing between the 3 channels would be as good as your sense resistors.

...

Thu Feb 26 2009, 08:34PM

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

I would agree that you will just take a kind of convervation of energy aproach to the mess
you know you have 5v and 40a going in, feeding a load that is 2v, so there is 3v at 40a being lost to friction (or resistance whatever you want to call it) or about 120w of power.

Internally in the unit, you have 2 voltage drops, one across the sens resistors and one across the mosfets. The sens resistors are dropping .05r*40a = 2v (ie, 80w) which leaves 1v left for the mosfets (40w). To get the power for each individual device just divide by the number of devices (3) to get about 25w in each resistor and 12w in the fets. Actual power dissipation will be a little bit lower because of the resistance in the traces/wire, in fact if you have more than about .02ohms of resistance in your wiring you will need to up the voltage a little (lucky the flat packs have a nice easy way to do that) or use smaller sens resistors. Alternatively, you can lower the power dissipated in the mosfets by turning the flatpack voltage down a little so that the mosfets are just at the edge of saturation

I any case, yes the linear supply route is probably the way to go for driving diodes, and when proper choice of supply you can actually get it to be quite efficient. But its expensive and bulky, so I am thinking that instead of varying the resistance in series with the supply, just adding a fixed minimal resistance (primarily a small current sens resistor and the wires from the supply to the diode), and varying the voltage coming of the the supply. My plan is that it will let you run at like 5-10w dissipated in the resistance, and have a nice 80-90% efficient switching power supply doing the grunt work.

aww
BTW, now that I am at a uni and have access to the shop doing heatsink work is a breeze for me. If you know the shape you want and don't mine waiting little while I could trade you a nice looking heatsink for a driver board or something. Another thing to consider would be to design the board so that it fits on a CPU cooler, since they are dirt cheep and can dissipate tons of power.

EEYORE

Thu Feb 26 2009, 08:35PM

Registered Member #99 Joined: Thu Feb 09 2006, 06:10PM
Location: florida, usa
Posts: 637

Electroholic wrote ...

Hey matt, how is your flatpack hooked up on the input side, I managed to fry two smaller 75W models by turning them on too fast.

If you know the input is 5V and you know the current going thru the circuit is I, input power is then 5 * I.
Now if you know the voltage drop across the LD is Vd, diode input power can be expressed as Vd * I.
Which means the rest of the circuit i.e. FETs + sense resistors have to drop 5 - Vd that many volts.

Current sharing between the 3 channels would be as good as your sense resistors.

Hello,
My flatpack is just connected to the mains (hot, neutral). How did you fry them by turning them on too fast?
So it is safe to assume 200watts will need to be dissipated when run at full power then...LD will see about 80watts or so. Current sense resistors will see about 13A^2*0.05 each...
Thats leaving 94watts for the three mosfets...Bummer...That will all require a hefty as hell heatsink for the mosfets/resistors. I was hoping not to have to trim the voltage down, cuz I need to run 5volt fans to keep the heatsink cool. (would be wonderful to power everything off the one flatpac

)

I better back up, take a breath, and plan a little more before ordering more parts. I need to find out how low all my stuff can go. (voltage).
Thanks!
Matt

Electroholic

Fri Feb 27 2009, 12:17AM

Registered Member #191 Joined: Fri Feb 17 2006, 02:01AM
Location: Esbjerg Denmark
Posts: 720

Mine was VI -200 75 W It was connected to a low impedance Li-poly pack thru a relay. IIRC lt needs a few uS to start up properly.
Can you down trim your flat Pac 10%? It would help with the heat.

EDIT:

NOTE: Do not allow the rate of change of the
input voltage to exceed 10 V/Î¼s for any input
voltage deviation.

Quoted from vicor app notes.

EEYORE

Fri Feb 27 2009, 06:42PM

Registered Member #99 Joined: Thu Feb 09 2006, 06:10PM
Location: florida, usa
Posts: 637

Electroholic wrote ...

Mine was VI -200 75 W It was connected to a low impedance Li-poly pack thru a relay. IIRC lt needs a few uS to start up properly.
Can you down trim your flat Pac 10%? It would help with the heat.

EDIT:

NOTE: Do not allow the rate of change of the
input voltage to exceed 10 V/Î¼s for any input
voltage deviation.

Quoted from vicor app notes.

Mine is a VI-LU0-EU. The specs say it can be trimmed down 10%, but the engineer said it can be trimmed down to 2.5volts. I dont know why they would differ so largely with each other, and I have found that the flatpac WILL shut off when trimmed down a few volts. (I cant remember if I had a load on it or not).

I have no idea how to control the input to the flatpac. Seems that would be Vicor's responsibility. I just plug the thing into the wall and let it rip! Before the current controller, I used just a 0.1ohm 100watt resistor to limit current, and I had a 1.5ohm resistor acting as a load across the output. Laser diode would get slammed with 28A or so all at once. There is a slight delay though from plugging in and actual power output. (I think Vicor speced a 10mS delay). I dont guess this is a soft start? I couldnt get much out of the engineer...

I managed to get two largish CPU heatsinks with fans on ebay for $1 each. We will see how these pan out. I may use them both even though one is large enough to PHYSICALLY fit all the parts on. (Dont know about thermally).

Matt

EEYORE

Fri Mar 06 2009, 04:13AM

Registered Member #99 Joined: Thu Feb 09 2006, 06:10PM
Location: florida, usa
Posts: 637

While still waiting for the heatsinks to arrive, I decided to play around with some TEC cooling since I havent overcome the cooling issue.

I used a TEC from an FAP 800 module. (Has four in series which requires 48volts 7A max). I sandwitched it between a laser diode and a copper heatsink with fan. I then cranked up the current to 22A and watched the resistance on the thermister. With 12volts applied to the single TEC, the laser got cold within minutes even though it was running 22A!

This is very good news for me. The folks at alt.lasers seemed to think I was an idiot (when I asked about doing such). They all agreed there would be no way to remove the heat with a single TEC. Well, I just had to turn the TEC off because the laser was getting very cold to the touch!

The water cooling diding to crap. The temperature began to rise quite quickly with the CPU cooling kit I got awhile back. That was a waste of 130$!!!!!!!!! (Maybe I can sell it on ebay and recover a little).

So now it looks as if the TEC route is the best. It may deffinately NEED more power to MOVE the waste, but the laser diode is whats important here! That and space/ease.

Looks like I will need to design some crude way of controlling the TEC to ensure it doesnt cause condensation on the laser...Any ideas? Im thinking an op-amp to control a mosfet. Kind of like a thermostat.

I still have to try this out with my coherent 40watt laser diode. (I have a "new" 18watt fiber coupled laser diode from OptoPower Corp.). I imagine the coherent will likely be happy with the TEC as well.

Matt

...

Fri Mar 06 2009, 04:36AM

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

Tecs do do awesome things, althogh I prefer a combination of tecs and a decent waterblock (although the new heatpipe based heatsinks will beat many watercooling rigs!), but to control them i would recomend getting a used pid controller off ebay. They make them for running ovens and the like, and you basically just hook it up to a big relay in series with your tec and it will regulate the temperature down to a few degrees :)

seems like a pretty sweet deal

lets see some high powers lasers already!

EEYORE

Fri Mar 06 2009, 06:29AM

Registered Member #99 Joined: Thu Feb 09 2006, 06:10PM
Location: florida, usa
Posts: 637

... wrote ...

Tecs do do awesome things, althogh I prefer a combination of tecs and a decent waterblock (although the new heatpipe based heatsinks will beat many watercooling rigs!), but to control them i would recomend getting a used pid controller off ebay. They make them for running ovens and the like, and you basically just hook it up to a big relay in series with your tec and it will regulate the temperature down to a few degrees :)

seems like a pretty sweet deal

lets see some high powers lasers already!

Wow, pretty advanced. I will snipe it!

Yea, TECs are inefficient but compact and easy to use. Its nice not to have the temperature rise quickly as before and not to have to have a 5lb heatsink...

Pictures will come soon. I have to find the charger for the battery!

Matt

EEYORE

Sat Mar 14 2009, 07:25PM

Registered Member #99 Joined: Thu Feb 09 2006, 06:10PM
Location: florida, usa
Posts: 637

Hello,
I am having some trouble with the new power supply. I was testing it out and have found a strange problem. When I trim the voltage down (at all) the dang thing refuses to act nice. (Flatpac)

I cant get the current to go from zero to max. It will go from zero to some random level based on the voltage input. (much less than the max current of the flatpac) When trimmed down to just 4.5volts (5 volts supply), it will not reach 20A. The further I trim down the voltage, the less current it will reach.

Another funny thing is that the supply voltage goes down when current goes up. (Even with no trimming). As current goes up, voltage drops down. (5 volts down to around 4.5 volts when drawing 25A or so) I put a 2.5ohm power resistor across the output of the flatpac and this actually decreased the max current level (when trimmed at all). Before, I cant trim the voltage down past 4volts at all without some load across the output. With no trimming, the load resistor has no effect on voltage or current.

The flatpac also has a high pitched squeal when running around 10A or so. It turns to a hiss when running more current.

I tried lowering the load resistance (for more current through the load across the outputs) and it just hurt performance even more.

This shouldnt be going on at all. The op-amp is speced for 2.7volts as its min supply voltage. The mosfets ought to work fine around 3volts, as should the voltage reference. I feel the stuff on my end ought be ruled out, but I could be wrong. Seems something aint right with this flatpac. The engineer told me it would work fine when trimmed down to 2.5volts (from 5volts). He didnt mention anything about a load either. Seems it isnt regulating at all when drawing current. This thing wont work right even at 5volts. (The current goes to max, but the output voltage drops considerably rather than STAYING at the friggen speced voltage of 5 volts).

Any ideas? Anything need clarification? I am now going to re-check all the parts specs to make sure they WILL indeed run from 3 volts or so.
Matt

...

Sat Mar 14 2009, 11:07PM

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

It sounds like oyu are having some kind of odd oscillation in the driver, have you tried poking at it with a scope to rule that out?

Also, how are you getting the flatpac to trim down below like 4.6v? I believe we have the same one (from the electronic goldmine, I forget the exact p/n) and even grounding the 'trim' pin I don't think I managed to get it to drop below 4.5v.

Good luck!

first previous 6 7 8 9 next last

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