Help with dual red setup.

I have been making a dual red setup with a PBS. I have done what I thought was a perfect alignment. The beam was real nice for about 15min then the LOC on the straight through axis went LED on me. I replaced the LOC and within 5min it was a LED too.

I searched through the forum and found many references to back-reflection. I believe this is the case here.

In a couple of posts it was stated to turn the PBS slightly to make the back reflected beams not return to the LOCs. No where was I able to find how much to turn the PBS.

When adjusting this dual setup, do you make it so the reflected beam does not return to the lens at all being just outside the lens area. Or is it ok to have the reflected beams enter the lens as long as it is not directly centered in the lens?

I don't want to loose any more diodes due to back-reflection.

Thanks!

Magman116

are the diodes isolated from each other?

Quote:

---

Originally Posted by andy_con

are the diodes isolated from each other?

---

Yes, the diodes are isolated in their mounts using 5mil Kapton tape.

I use Aixis modules with AR coated glass collimator lenses.

I am running them in parallel using a Flex Mod P3 with a 1ohm balancing resistor on each diode.

Drive current to each LOC is set @ 400mA.

back reflection does happen but its pretty rare

Quote:

---

Originally Posted by andy_con

back reflection does happen but its pretty rare

---

Back reflection is the only explanation I have. Current to each diode is good, isolation is good. temperatures at good. Power supply is very stable.

I replaced the last blown LOC and had both LOCs running since last night (not through the PBS) and they are still running great right now.

I am afraid, if I put them through the PBS then the through-axis LOC will blow again.

cod due to back reflection is indeed uncommon. Although it's more likely to happen when fiddling around with optics. Suddenly, for one split pico second all conditions are right, you create an external cavity, and poof !
Due to fabrication tolerances, if you replace your diode your laser-die will never be in the same place twice, and point in the same direction. So a double backreflection death is extremely unlikely.

Don't rule out esd, it's not always an instant killer, propagated delayed defects will kill eventually too. Short your diodes as soon as you receive them, if not already. And unshort them when connected to a lasorb.

Swap your drivers to rule out one of them causing havoc.

The diodes you are using (LOC) are rated for 240mA CW. You drive them with 400mA this makes them very sensitive to for backreflection, and is only possible with good optics.

Try running them at 300mA and if they still die replace the psu.

Quote:

---

Originally Posted by -bart-

cod due to back reflection is indeed uncommon. Although it's more likely to happen when fiddling around with optics. Suddenly, for one split pico second all conditions are right, you create an external cavity, and poof !
Due to fabrication tolerances, if you replace your diode your laser-die will never be in the same place twice, and point in the same direction. So a double backreflection death is extremely unlikely.

Don't rule out esd, it's not always an instant killer, propagated delayed defects will kill eventually too. Short your diodes as soon as you receive them, if not already. And unshort them when connected to a lasorb.

Swap your drivers to rule out one of them causing havoc.

---

Each time I replaced the diodes, I used a new Aixis module.

My setup is on an ESD protected workbench designed for electronics manufacturing. I am an electrical engineer and have a pretty full electronics lab, taking into consideration ESD.

I have two Flexmods. I guess I could run each LOC off its own driver. I would think that running them for 14+ hrs continuously on the same flexmod would indicated there wasn't an issue with the driver.

So, Bart, should I not worry about the laser alignment to the PBS and get them as square as possible? Or should I factor in a small turn to the PBS to slightly offset the reflected beams?

I just inspected one of the LOCs under the microscope. I saw the junction wire-bonds were still intact. Then slowly brought the voltage and current up to about 300ma. This was more than enough to illuminate the die.

Everything looked normal until I zoomed in. Once zoomed in, it looked like the very front of the die had a very small burn mark on it. It looked like this burn mark was preventing the light from emitting from the front of the die.

This looks to be back-reflection damage. I will see if I can't scrape the burn spot and see if it will lase.

Scraping was not a good thing. The die started to crumble where the burn mark was. If I chipped a small portion of the die away, I could get some lasing, but small. I tried to do more, but broke the die.

Something still burned the end of the die.

Quote:

---

Originally Posted by mccarrot

The diodes you are using (LOC) are rated for 240mA CW. You drive them with 400mA this makes them very sensitive to for backreflection, and is only possible with good optics.

Try running them at 300mA and if they still die replace the psu.

---

Been running them at 300mA. I have the PBS offset slightly with good near-field and far-field alignment. No problems yet. I do notice the through-axis LOC is drawing about 1.3mA more than the other, but that isn't a significant amount.

Curious. Just checking, you do have the LOC's polarisation set perfectly? And the PBS is the right way around right? If you mask the straight through LOC with a piece of black card do you see the any light from the other LOC/PBS hit the card? Just a thought.

In my quad, I am running 4 LOC at 450mA and thus far not a single LOC failure.

Quote:

---

Originally Posted by dnar

Curious. Just checking, you do have the LOC's polarisation set perfectly? And the PBS is the right way around right? If you mask the straight through LOC with a piece of black card do you see the any light from the other LOC/PBS hit the card? Just a thought.

In my quad, I am running 4 LOC at 450mA and thus far not a single LOC failure.

---

dnar,

I did as you said. I rotated each diode till I had maximum output from the PBS. The output forms an almost perfect "+", but oval.

I placed a black card in front of the through axis LOC and there was no sign of beam reflection from the other LOC.

I placed the card in front of the 90-degree axis LOC and there was a very small reflection from the through-axis LOC.

I really don't see how the through-axis LOC could reflect to the 90 degree axis?

Could the through-axis LOC be slightly off in its polarization? Maybe I don't have it turned enough?

Thanks for all the inputs!

I blew another LOC and I found out why.

While playing back a show, I monitored the voltage drop across each 1-ohm resistor. This would show me the current each LOC was getting. A series of patterns was displayed and I noticed the current to each LOC was over 650mA!!

I am using Spaghetti and set the min and max output in the projector settings and the current was within limits. I set the upper limit at 300mA and the lower at 14mA.

I am still not sure why the current went so high while displaying multiple patterns. i repeated the pattern and the current went high again and the through axis LOC blew.

I ordered 3 more Flexmod P3s.

-magman116

good to know, will built my quad red soon and will keep this in mind.
Michel

A couple of questions.

What voltage are you running the flex mods at?
Are you sure they are 1 ohm? might want to bump it up a little 1.5 ohm maybe?
How good is the PBS? Is it ar coated?

And double check the polarization. It will be arc seconds rotation to get it dead nuts. Do you have a power meter? Or can you rig a Si photodiode with a ND filter to get a relitive reading? you should optimize each diode rotation with a meter if possible.
Do you have lasorbs? If not you should have them just for peace of mind, and they work.

chad

Quote:

---

Originally Posted by magman116

I blew another LOC and I found out why.

While playing back a show, I monitored the voltage drop across each 1-ohm resistor. This would show me the current each LOC was getting. A series of patterns was displayed and I noticed the current to each LOC was over 650mA!!

I am using Spaghetti and set the min and max output in the projector settings and the current was within limits. I set the upper limit at 300mA and the lower at 14mA.

I am still not sure why the current went so high while displaying multiple patterns. i repeated the pattern and the current went high again and the through axis LOC blew.

I ordered 3 more Flexmod P3s.

-magman116

---

Right, I have heard a few issues with P3 killing diodes when modulated.

Tonight I will setup a test P3, with parallel test loads (1n4004's) and a function generator. Will report back.

I want to ensure I don't blow my quad once I get my projector running!

BTW, I recall Heroic was involved sorting a similar issue for some one else, I recall they fitted some C to the LD's but they were in series from memory.

Quote:

---

Originally Posted by chad

A couple of questions.

What voltage are you running the flex mods at?
Are you sure they are 1 ohm? might want to bump it up a little 1.5 ohm maybe?
How good is the PBS? Is it ar coated?

And double check the polarization. It will be arc seconds rotation to get it dead nuts. Do you have a power meter? Or can you rig a Si photodiode with a ND filter to get a relitive reading? you should optimize each diode rotation with a meter if possible.
Do you have lasorbs? If not you should have them just for peace of mind, and they work.

chad

---

Hi Chad,

I have a 9V 5A regulated switching power supply that runs my 2 LOCs, a 1W 445nm turned way down, and a 200mW 532nm. It is very stable with very little ac ripply.

The 1ohm resistors are 1% 3W.

The PBS is from LASERWAVE and is AR coated or it is supposed to be.

Many years ago, I made a laser power meter using a photo diode. I compared it with a power meter we used at work and noted the delta. While not as accurate, it is pretty close using the offset. I mainly use it as a general reference when tuning or adjusting the driver.

I did not use it to setup the polarization through the PBS; good idea.

I have thought about using LASORBS for the LOCs, but only have them on the 445nm and the 532nm. Seeing how I have already blown 5 LOCs, it would make sense to use them. I will be getting some 642nm 150mw diodes from Francisco to replace my LOCs soon.

Quote:

---

Originally Posted by dnar

Right, I have heard a few issues with P3 killing diodes when modulated.

Tonight I will setup a test P3, with parallel test loads (1n4004's) and a function generator. Will report back.

I want to ensure I don't blow my quad once I get my projector running!

BTW, I recall Heroic was involved sorting a similar issue for some one else, I recall they fitted some C to the LD's but they were in series from memory.

---

It will be interesting to here your findings.

I did some testing here with an o-scope and different software. I was observing the ILDA modulation line for each color, one at a time, and the output from the driver.

It was interesting to note that NLS, NE-Laserscan, and Spaghetti all have different pulse width outputs. The voltage was 5V for each waveform, but the on-time was different. NLS had the longest on-time, and Spaghetti had the shortest.

This could account for different current draw while running different programs. With Spaghetti, when playing back complex scenes, the pulse width would increase quite a bit.

I created a waveform scene all red ilda file, then played in the different programs. In NLS the current to the LOCs wound run about 380mA, but in Spaghetti is was around 205mA. Why?

Quote:

---

Originally Posted by magman116

It will be interesting to here your findings.

I did some testing here with an o-scope and different software. I was observing the ILDA modulation line for each color, one at a time, and the output from the driver.

It was interesting to note that NLS, NE-Laserscan, and Spaghetti all have different pulse width outputs. The voltage was 5V for each waveform, but the on-time was different. NLS had the longest on-time, and Spaghetti had the shortest.

This could account for different current draw while running different programs. With Spaghetti, when playing back complex scenes, the pulse width would increase quite a bit.

I created a waveform scene all red ilda file, then played in the different programs. In NLS the current to the LOCs wound run about 380mA, but in Spaghetti is was around 205mA. Why?

---

Curious. Did you apply a "colour change" effect to the scene? If so, Spaghetti alters the chroma but maintains the luminance value. The result is that when you change the colour of a frame, you drop brightness.

Quote:

---

Originally Posted by magman116

I blew another LOC and I found out why.

While playing back a show, I monitored the voltage drop across each 1-ohm resistor. This would show me the current each LOC was getting. A series of patterns was displayed and I noticed the current to each LOC was over 650mA!!

I am using Spaghetti and set the min and max output in the projector settings and the current was within limits. I set the upper limit at 300mA and the lower at 14mA.

I am still not sure why the current went so high while displaying multiple patterns. i repeated the pattern and the current went high again and the through axis LOC blew.

I ordered 3 more Flexmod P3s.

-magman116

---

Did you setup the max current of the flexmod with an external +5VDC on the mod input?

No matter what any DAC is outputting with whatever software is used, if the out current of the flexmod exceeds the set max current then this is obviously a problem with the current limiting of the flexmod. Probably an HF resonance thing going on here. If you’re using lasorbs try not using lasorbs.

I just read up on the lasorb and I couldn’t find any information on multi diode setups with a single driver. But from what I gather on the Lasorb functionality, it doesn’t seem to be a good Idea to use these in Parallel configurations. If a Lasorb acts out its protective function on one diode the other could get a peak current due to the first being shorted and shortly after the second lasorb becoming active causing instability in the circuit and start a resonance between the Lasorbs and driver.


Reading the datasheets of the lasorb it appears these are laid out for single diode protection with single drivers.

Quote:

---

am using Spaghetti and set the min and max output in the projector settings and the current was within limits. I set the upper limit at 300mA and the lower at 14mA

---

I keep reading this and canot make sense of it.
Are you saying you are using the software to limit the min and max current to the diode?

If so I would say this might be something eliminate from the chain first - unless you can be 100% certain that the max output can't be exceeded by the software.

What is the FM3 max output mA set to with 5V applied to mod?

Quote:

---

Originally Posted by norty303

I keep reading this and canot make sense of it.
Are you saying you are using the software to limit the min and max current to the diode?

If so I would say this might be something eliminate from the chain first - unless you can be 100% certain that the max output can't be exceeded by the software.

What is the FM3 max output mA set to with 5V applied to mod?

---

When I setup the flexmod, I set the max current at 840mA. I did have some issues with this FM3 after I first got it. It would not allow me to set the max current even though I could set the bias current. I put it aside for a couple of days and tried again and it worked that time. Could be a defective FM3 that is why I ordered 3 more.

So surely, having max current at 5V set to 840mA, you are asking for trouble if the software has any bugs or ways to output high power beams that bypass the max output.

At this point, I would think that if you actually set the output at 5V mod to 300mA on the FM3, you might stop blowing things up

Quote:

---

Originally Posted by norty303

So surely, having max current at 5V set to 840mA, you are asking for trouble if the software has any bugs or ways to output high power beams that bypass the max output.

At this point, I would think that if you actually set the output at 5V mod to 300mA on the FM3, you might stop blowing things up

---

He's running the diodes in parallel so 700-800mA could be ok.. But the FM3 max current has to be setup with 5VDC on the mod in of the FM3 to assure that the driver limiting is correct. If you’re still getting 600+mA per diode when modulating with a signal, then you’ve got a driver and/or resonance (if lasorb’s are used) problem.

Ah, ok, but I guess the sentiment is there - using software to limit max current is not a good approach, and introduces an unnecessary failure point. Better set the absolute max current required on FM3, then use software to reduce as necessary.

Of course, if the max current couldn't be set, this raises issues with the hardware.

I suppose the next question is:
Was the max current setup done as per the intructions, and what were the effects of not being able to set it? e.g. did the current just not increase/decrease when the pot was turned? What was its default value on switch on?

Maybe a thought:
A typical color modulation could be square wave, lets say 1Khz square wave.
Such square wave can have many harmonics resulting in high frequencies.

A diode also got a certain capacity, the higher the frequency the more current will flow trough a capacitor.

Maybe there is a difference between diodes causing one diode to get the full current when modulated at high frequencies.

But maybe it's better to not go this risky route just to save 25,- for a 2nd driver....

When the Fm3 was first powered up, as per the instructions, I was able to set the bias pot using no modulation. Then when I connected 5V to the modulation input, nothing happened. I believe the current was lower than when setting the bias. There was now adjustment when turning the pot. Then I tried it all over again with the same results. So I decided to put it to the side and work with another driver I had. A couple of days later I was putting together some more LOC modules and wanted to test the output and grabbed the FM3 again. This time I was able to adjust the bias and the max current. I set it at 400mA. Then I shorted the output and connect the LOC. It worked no problem. That is when I put it in the system and adjusted the current for 2 LOCs. Both LOCs powered up just fine.

Everything seemed to functional normally, but kept blowing the through Axis LOC. That is when I decided to start monitoring the current across the 1 ohm resistor.

Today, I added TECs to the LOCs and the driver heatsink. The driver heatsink was getting very hot. Too hot to touch for more than a second.

All 3 FM3s are mounted to a 4"x1-1/2"x1/4" aluminum jig plate.

After adding the TECs, it doesn't get very hot at all and the LOCs are cold to the touch. I ran the same frames that caused the current jump before and now there is no big current jump, there is an increase of about 50mA over its average range, but is staying within the max current setting of the FM3.

I think I will keep the TECs in to help with the heat. I am running them at 2A each. I still have not put TECs on the 445nm or the 532nm modules. I have about 20 40mmX40mmx4mm 50W TECs.

As your failures are always with the through-cube diode, I can't help but think you have something else wrong, like a faulty resistor, joint etc. When running parallel, any diode with dodgy wiring or resistoris going to damage the other diode(s).

Sorry for the double post. I setup a Flexmod P3 test today, using adjustable DC lab supply, function generator and a dummy load comprising 3 x IN4004 diodes + 1ohm resistor in series. Ignore the Mod signal going negative slightly, I used an AC generator and used a C/diode circuit to clamp/shift the signal.

The following DSO shots are measuring the following:
Trace 1 (Yellow): Modulation signal.
Trace 2 (Blue): LD current measured across the 1ohm resistor. 1mV = 1mW.

Setup procedure was as follows:
BIAS set with MOD at GND to 10mA
MOD @ 5V LD current set to 1000mA

PSU: 12V Freq: 2kHz Peak LD current: 1.02A
Attachment 23010

PSU: 12V Freq: 20kHz Peak LD current: 1.00A
Attachment 23011

PSU: 12V Freq: 200kHz Peak LD current: 0.776A
Attachment 23012

PSU: 12V Freq: 2kHz Mod: > 5V Peak LD current: 1.00A
Attachment 23013

PSU: 12V Freq: 20kHz Mod: >5V fast rise time Peak LD current: 0.968A
Attachment 23014

PSU: 12V Freq: 20kHz Mod: >5V Peak LD current: 1.00A
Attachment 23015

PSU: 24V Freq: 20kHz Mod: >5V Peak LD current: 0.992A
Attachment 23016

No signs of the LD current exceeding the configured setting! No signs of oscillation, ringing etc.

What is curious, is the phase shift between modulation and LD current at higher modulation frequencies.

Also, note the peak LD current when the PSU is raised to 24V. Not perfectly flat as with 12V supply.

I am hoping Dr Lava will add comment.

EDIT: And here is why I don't like running the P3 at 24V when driving a single LD or LD's in parallel:

PSU: 12V LD: 400mA
Attachment 23017

PSU: 24V LD:400mA
Attachment 23018

See how the current waveform is slewed in the rising edge?

dnar,

Great work!

I must say my waveforms do not look like yours at all. I will try to get some scope shots. Looks like you have a digital o-scope. Nice.

My waveforms look more like a PWM signal. It could be the DAC I am using and not the FM3s.

I am using the FSX2.0 DAC. http://www.ctrnd.com/NormLaserShow/FSX20.htm

dnar, what DAC do you use?

The tests above are not using a DAC. I am using a function generator. Purely testing the P3's. The max configured current is never exceeded. Pretty sure your problems lay elsewhere.

Quote:

---

Originally Posted by dnar

The tests above are not using a DAC. I am using a function generator. Purely testing the P3's. The max configured current is never exceeded. Pretty sure your problems lay elsewhere.

---

While I still haven't taken scope shots yet, I did the testing with a function generator. 5V signal, 10Khz modulation. I tested sin, square, and sawtooth waveforms. Each had a different result. Over all, the results were very comparable to what you got.

I then connected the DAC and displayed a test pattern. Looking at the modulation input and the driver output, the waveform is definitely a PWM waveform, both modulation and driver output waveforms were the same. There was a lot of noise and the output really started to look bad above 50khz.

dnar, would you please check your DAC output and tell me what you get?

Ok, I'll retest with my AUDIOLASE DAC. What DAC are you using?

Quote:

---

Originally Posted by dnar

Ok, I'll retest with my AUDIOLASE DAC. What DAC are you using?

---

I am using a Frame Screamer Express 2.0, FSX2.0

Nice scope shots, there. I especially like this one
http://www.photonlexicon.com/forums/...3&d=1295087813
that shows how the protection clips nicely when the input is overdriven.

With the 24V capture, there is a much larger discrepancy between the power supply voltage and the laser diode voltage, so the hump you see is the driver being more 'careful' at the peak to not to let the 24V supply spike the diode.

When driving LDOCs with 300ma or more, it's important to ensure the die is kept clean during assembly. any small dirt or smoke particle can lead to failure. What lenses are you using? Some can generate small back reflections.

Dr Lava,

I have added another FMP3 so each LOC has its own. I am not seeing the current spikes and all is well.

I am using O-like AR glass lenses for red.