High Current Laser Diode PSU / Projects / Forums

Forums

4hv.org :: Forums :: Projects

« Previous topic | Next topic »

High Current Laser Diode PSU

previous 1 2 3 4 next last

Move Thread

LAN_403

EEYORE

Sat Aug 09 2008, 07:42AM

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

Ok, I am having some trouble. I got my new EL5244CN op-amps and tried them out. I cant get them to control current. I am now using a 5volt flatpac since these are 5 volt op-amps. The current starts at 1.8amps and can only be varied to 2.040amps. Any ideas what may be the cause of this? (Still using the trimpot/resistor divider Vref. Are some op-amps not going to work for this? (I just went looking for a Rail-Rail op-amp with lowest min. operating voltage. I just dont understand why the TL082 works just fine, but others do not even come close to working at all?

My original vref was a 2k trimpot and a 10k resistor. The trimpot failed somehow so I switched to a 50ktrimpot and 470k resistor. Thinking the current was too low, I switched to a 500ohm trimpot and a 1kresistor. Nothing works with these new op-amps.

Also, I noticed that my circuit operates differently from the explanation. It seems that the current rises as the difference between Vref and input voltage increases. When Vref=voltage on the mosfets, no current flows. As I lower Vref, current begins to flow. This is with the TL082op-amp. Im so confused

Matt

Steve Conner

Sat Aug 09 2008, 01:04PM

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

Remember that stuff I mentioned about input common-mode range?

EEYORE

Sat Aug 09 2008, 06:59PM

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

Dr. Conner wrote ...

Remember that stuff I mentioned about input common-mode range?

Hello,
I believe so...Is that the -5volt supply you mentioned? Is it possible to make them work on just a +5volt supply?
I got another op-amp from maxim to try. The MAX4167. It is a 5volt rail-rail op-amp. It worked a little better than the other new one, but would not go past 2.2amps. After 0.5volts difference between Vref and the 5 volt power supply, current would no longer increase. It would go from 0-2.2amps nice and smooth and then stop. Is this 0.5volt difference related to the "common-mode" range? Or would the limit in current be a function of the 5volt supply and the higher in resistance of the mosfets along with the op-amp? (0.5ohms each for the mosfets).

I was probing around the circuit and found one of my sense resistors getting hot! (I checked it as I noticed current was rising bad). The other one is cold. Upon measuring voltages across them, I found the hot one had all the current through it while the other one had nothing? This happens with all the op-amps used now. For some reason, before I moved houses neither resistors would warm at all! Even with 5amps through them. (0.1ohms, 5amps should be 2.5watts right? I^2R?) Perhaps I need a look at the mosfets now...

Hahahaha....this power supply has got my tail between my legs. I better go get a book on op-amps...
Thanks for the help guys!
Matt

Electroholic

Sat Aug 09 2008, 09:38PM

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

from the epcb layout, seems like you are doing high side current sense and using pfets.
The constant current dummy load was using nfets with lowside current so everything was referenced to ground.

Now i looked at the el5244 datasheet, seems like Vcm can only go up to 3.5V with a 5V supply.

Voltage feedback from your high side current sense will range from 5 to 4.5V with respect to 0-5A. notice thats inverted.

Voltage from Vref will range from 0 to 1.25V(if you are using a 1.25V chip)

So, there are two problems,
1, you are trying to use an opamp to keep the two voltages the same, but in reality the will never even come close to each other.

2, feed back is totally out of your opamps Vcm range. If you look up high side current sense, you will see they usually call for differential amps, not opamps.

EEYORE

Sun Aug 10 2008, 06:18AM

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

Thanks for the reply. It looks as if I went way off the deep end with these other op-amps. I switched out the pfet for an nfet as well as configured the sense resistor as is done in the dummy load circuit. Still no luck.

I dont have a vref chip just yet, so maybe that is the deal? Id love for all these problems to be simply the darned vref being a crappy resistor divider

I think I will switch back to the TL-082 op-amp since I had some success with it. 12 volts may cause more heating, but it works.

Now after re-doing the supply for the TL-082, it seems as if one transistor is hogging the current. One gets SMOKING hot as does its sense resistor. This NEVER happend before even when left on for hours. Any ideas? Just bad luck that finally kicked in here? So odd! It was working well minus the heating. (Fan mounted on heatsinked helped). Maybe God is telling me to give up?

Matt

Electroholic

Sun Aug 10 2008, 06:29AM

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

i would probably add the 12V and tide the grounds together. so 12V for the opamp and 5V for the high current bus.

some capacitance at the opamp inputs might help stabilize things a bit.

u sure you didn't just fry half of that opamp?

EEYORE

Sun Aug 10 2008, 07:50AM

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

Thank you for the tips! I would be stuck with a battery and resistor without you guys helping out! So I swapped out some copper wires for nickel plated wires and things seem to be back to normal with the tl082 version. Seems copper oxidation was at fault for the strange behaviour. That and the breadboard sucking bad! (I need to get this circuit soldered down!)

So what kind of capacitance are we talking here? 0.1uF or something small on the - and + inputs to the op-amps? How about on the output? Might this help with the "current walking"? It likes to climb up and down a bit from where it is set. Could this be due to the crappy Vref I am using? (trimpot/resistor divider). It usually either climbs some 10mA-30mA or so and settles or drops some 10-30mA and settles out a bit. (Fine for high current laser diodes that are not being run near their max, but BAD for anything else!)
Matt

Electroholic

Mon Aug 11 2008, 05:50PM

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

not sure how well the 5V supply is regulated, but any small voltage drift will definitely result in "current walking" if you are only using a resistive divider. i reckon a bandgap voltage reference will fix that right up.

EEYORE

Tue Aug 12 2008, 04:55AM

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

Electroholic wrote ...

not sure how well the 5V supply is regulated, but any small voltage drift will definitely result in "current walking" if you are only using a resistive divider. i reckon a bandgap voltage reference will fix that right up.

I am using a vicor flatpac. I beleive it is supposed to be very decently regulated, so Im unsure as to the problem here. I dont see how the original maker of this circuit has met with such sucess? (He IS using an LM723 as the vref).

Here is what stumps me most. All of you guys who know this stuff say that the op-amp works by comparing Vref to the feedback voltage and adjusting output until the feedback is the same as Vref. Right? So why is mine not doing this? Here is what mine does. Both Vref and voltage input start at 11.99volts. (Went back to the 12volt supply). Zero current flows at this point. As I adjust the resistor divider such that the Vref is decreasing, current through the fets goes up. This may sound alright, but it makes making a solid Vref very difficult! In order for things to work out right, Vref needs to be regulated, right? AND the voltage in MUST start at Vref's max voltage. If there is ANY difference, current begins to flow. So, I havent seen any regulating circuit that can go from 0-voltage in. Max regulated voltage is always some voltage less than voltage in. Thus, I cannot get current to start from zero.

Is this because of the op-amp I am using? Or is it due to the way I am operating it?
I watched the voltage on Vref while running 5amps. It DID change by 10mV or so...But it went back down. I still dont understand the walking I am seeing. Sometimes it will just keep on climbing and not ever stop. It will steadily rise to 100mA greater than what I set it at. Could it be to the breadboard connections? I do need to solder it all down, but plan to use Expresspcb. (My homemade days are over). Id hate to pop out 60$ with some kind of fatal error...
Thanks for you help guys!
Matt

Electroholic

Tue Aug 12 2008, 06:37AM

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

sounds like you are back to high side current sense.

as i said, voltage from feed back will swing from 12 to 11.5 when your current goes from 0 to 5. So you want your control to be in this range. just use a bandgap voltage reference and hook one end to the positive rail and set up a divider across this reference. this way, you can get a stable voltage to the opamp as control.

previous 1 2 3 4 next last

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