Viasho Driver Question

[movax]

e2: Ooops, "this is not the my laser is broken" forum, I'm dumb. If a mod could move it over to the right place, that'd be great!

Hi guys,

Having some trouble with two 1W Viasho Green Lasers from CT Lasers.

The setup (important things that pertain to this):
- Custom circuit board that uses an Analog Devices AD53x5-series DAC. The output from this DAC feeds the Viasho Driver.
- The driver is I believe a VA-II-N series from Viasho. CT Lasers doesn't list the model numbers online and the laser is currently setup around 200 miles away. It's an external driver that takes in DC power and outputs to the laser head over a HD-15 umbilical cable.
- Power supplies for the Viashos are the supplied ones from CT Lasers
- Switches on the driver are set to 'ANG', which I believe should be Analog modulation. I assume TTL+ is active-high TTL and TTL- is active-low TTL. I helpfully did not receive any documentation with this laser

The issue:
The intent is for the brightness to be controlled over CAN bus based on input from a Kinect. In the software I wrote, I added a "manual" mode, which is basically a bunch of sliders that go from 0-255 for each color. In this mode, the lasers run fine and I can vary their brightness happily. Nothing breaks, and the Viashos happily go down to 0 (0V) and back up to 255 (5.0001V) with no issues.

As soon as Kinect mode goes on though, the green lasers very quickly cease to emit coherent light. The fan on the laser continues to run, but there is not output. The primary difference between these modes is that in Kinect mode, I am sending update messages every frame, or every 33.33 milliseconds. I see on CT Lasers website that the Analog Modulation is rated to 10kHz, which is a period of 100us. My resultant signalling rate should be well within the specifications of this input.

Even weirder, this seems to cause some kind of latch-up with regard to the driver (or triggers some safety cutout in the head). I have to power-cycle the lasers, ensuring that they power up with the analog modulation clamped to ground (the DAC in-contrast powers up in High-Z until I activate it), and then switch it back over to DAC control.

Questions/theories:
1. Anyone have a manual / datasheet for this thing?
2. Is there a warm-up period for the laser? Is this defined as power-on the laser and emit no light for a good 10 minutes before using it? In one case, the laser was powered on and immediately set to emit light, but remained at a constant brightness for a hour or so. Upon trying to switch to the faster control rate (via Kinect), they went down in a few seconds.
3. Input impedance of the driver may be lower than expected / DAC is drawing too much current with rapid updates.
4. 10kHz update rate is a lie, and I'm updating too fast. I don't think this is the case though, because I'm sure people use these in projectors.
5. I thought there was a grounding issue, but I ensured that the grounds of all the disparate power supplies are tied together, so it should have the same reference as the DAC.
6. I noticed that as the load increases (i.e. brightness) and therefore the current draw, the OEM PSUs respond by pretty aggressively over compensating and outputting ~13.5V (obvious as you can hear the fan get louder as its voltage increases) vs. expected voltage sagging/regulation to 12V.

For what it's worth, the other lasers are powered by FlexMods and identical copies of my custom board, and they've been abused for 10+ hours at the Kinect update speed, and they have never choked once. The boards themselves run at 40MHz, and the fastest control loop I had running took around 300 microseconds to run.

And to make my life even harder, these are part of an exhibit set-up two hundred miles away. Took a late night trip out there a few hours ago (2 hours of sleep, joy) after getting bug reports, and trying to debug the damn thing.

e: Oh, and the switch on the outside labeled 'Rt' / 'Curr' on the driver, does that select what voltage I get at the Test pins on the driver? I.E., I get some kind of current sense voltage, or a relative TEC voltage that correspond/scale up to real values?

[gottaluvlasers]

Please call me. We'll try to figure the problem out.
203-889-1319

[movax]

Please call me. We'll try to figure the problem out.
203-889-1319

Will do; not sure when you guys are open, so I'll call on Monday.

I searched around, and found some posts mentioning that they let the laser warm-up for a good 10 minutes before modulation, and that seemed to improve stability. Didn't appear to have an effect according to the team on-site though.

[movax]

We think it could be that the input impedance on the Viasho is in Chinese ohms, not real ohms (read: really really low). One of our guys will try building a quick unity-gain/voltage follower circuit to see if that helps alleviate the issue.

I think that in Kinect control mode, the modulation input could see some 0-255 steps that may be knocking out the DAC as current draw spikes.

[swamidog]

have you tried something like this: http://www.photonlexicon.com/forums/...407#post140407

We think it could be that the input impedance on the Viasho is in Chinese ohms, not real ohms (read: really really low). One of our guys will try building a quick unity-gain/voltage follower circuit to see if that helps alleviate the issue.

I think that in Kinect control mode, the modulation input could see some 0-255 steps that may be knocking out the DAC as current draw spikes.

[movax]

have you tried something like this: http://www.photonlexicon.com/forums/...407#post140407

That's the unity-gain/voltage follower/buffer/many other names op-amp circuit the other guy will try to build tonight, hopefully.

If the measured impedance turns out to be below 2kOhm, that's probably what's murdering the DAC.

[swamidog]

great. let us know if that solves the problem.

"color correction" boards, like this: http://shop.stanwaxlaser.co.uk/dz-co...oard-252-p.asp

will solve impedance issues and allow you to compensate for blanking delay issues, if you're running multicolor systems.

That's the unity-gain/voltage follower/buffer/many other names op-amp circuit the other guy will try to build tonight, hopefully.

If the measured impedance turns out to be below 2kOhm, that's probably what's murdering the DAC.

[movax]

Alright, so think we tracked it down now that the drivers/lasers are comfortably back at Mike's place. Unfortunately, didn't have a scope handy so these results are only 99.9% certain

Measured the input impedance of driver... 25kOhm. That's fine.
Measured current draw of modulation input...max, 210 microamps. Well shit, this isn't the problem.

Confirmed issue that our software/hardware allows a condition to exist where modulation input can see a very rapid 0V-5V (max) event happen (at our CAN bus speed + microcontroller speed, maybe 300-400 microseconds at most). This knocks out the driver and lasers remain off until power is reset.

Measured current-draw on 5V rail at steady-state, max brightness... 5.70A. Uh-oh, the PSUs that came with the Viasho are only rated for 6.00A.
Measured voltage on 5V rail at steady-state, max brightness... 4.65V. Uh-oh, that's quite a large voltage sag.

The DMM we had handy (Extech) didn't update fast enough to catch the transients, so couldn't get a number for the exact voltage sag. Was sitting there figuring out where to get a beefy 5V supply before recalling Mike is an electronics packrat, and went looking for caps.

Placed a probably decade old 16V, 2200uF electrolytic cap directly at the 5V supply input to the driver as a quick sanity check. Attempted to re-create condition...and lasers didn't die. Handled the jump just fine! 2200uF is probably overkill and I need to quantify with a scope, but in the end, the power supplies were undersized / had insufficient bulk capacitance. Transient / in-rush current draw of the driver sagged voltage enough for some type of latch/UVLO to trigger in the driver.

Now, there's probably a few things we can do here:
1. I need to update my embedded software to run a maximum 10kHz modulation rate for these drivers. Currently it just attempts to output whatever the PC commands as fast as possible; with a 30FPS Kinect input, this is still slower than 10kHz, but this can't hurt. Even then, doing 0-255 (0 to 5V) just once is enough to knock out the driver.
2 (*). Possibly enforce a ramping mechanism for certain deltas on modulation voltage. Can anyone comment on the health of the laser diode itself jumping from off to max brightness? I feel like this should be expected behaviour as in a projector, you kind of need this to happen, so I don't plan to try this out. So, I think these supplies should have been able to handle the laser jumping from 0 to 255.
3. Get a bigger supply, as described below:

My current plan is to replace the existing power supply infrastructure (2x OEM supplies that came with Viasho + random 12V brick for other lasers) with a single ATX PSU from Corsair or Antec in the 500W-600W range. I'll get ~24A on 5V, and in excess of 40A on the 12V rail, plus regulation that has to meet stringent ATX spec.

We'll probably implement a buffer for DAC outputs on all out future boards instead of depending on internal buffering in the DAC, but seems to be OK for now in that respect.