Diagnosing a galvo problem

LightBurn Hardware Compatibility Galvo
1

Well, I was pretty adventurous and got a JPT M7 that had a 300W MOPA source. Off AliExpress, I couldn’t find US vendors for that wattage and the US vendors wanted more than I paid for that for much lower wattage machines. It was definitely a different case and the source+controller box weighed in at 70kg.

I got it working under EZCAD2 and then Lightburn, but after a few hrs of trying it out I heard a pretty loud high pitch whine from the galvo head and it wasn’t cutting right. Both the red dot tracing and cut were smeared around. I pulled the lens and found when I touched the back of the Y galvo mirror, the noise stopped. I powered it down and the mirror does feel like it’s turning freely.

I cycled power with the USB unplugged and the whine noise remained. So, I can exclude anything to do with LB and the computer.

I came back the next day and it didn’t have a problem at first, but then ran a fill at 3000mm/s and after only like tens of seconds of that it started whining again and stopped scanning correctly. Again, the noise stayed while powered on, even while idle.

So, it seems like it will go away if left powered off for awhile. If it powers on without making the whine, it doesn’t seem like it trigger the problem while left idle. But it triggers quickly with any high speed motion. Once triggered, it will stay stuck in that problem.

Seems like the problem could be triggered by heat in the motor or driver, and the driver’s PID goes into crazy oscillations that create enough heat to sustain the problem and will only cool down if powered off. All speculation.

I messaged the seller and haven’t gotten a reply yet. I don’t know how much support I can get, if any, but I’m hesitant to crack it open as breaking those warranty seals could shut down my options of help from the vendor.

At least it doesn’t seem like any prob with the MOPA source. I checked the DB25 cable going into the galvo head and it looked fine and reseating it changed nothing. So this could be galvo head, the bjjcz controller card, or even power supply.

Or it could be as simple as a gain pot for a PID on the bjjcz card being set too high. I don’t know if there’s even any adjustment like that, though.

I found one guide online for a similar galvo system suggesting ground loops were a problem and the first thing to try would be bolting a new ground wire from the galvo to the controller and the controller to the 3rd pin. I’ll probably try that first once I know if the seller isn’t going to help and I’m ok with breaking the warranty seals.

Anybody familiar with troubleshooting these?

Can someone confirm my assumption that LB isn’t programming any config things onto the bjjcz that would survive a power cycle?

2

Perhaps a loose motor mount?

It’s been a looong time since I had anything to do with laser galvo mirrors (*), but I recall keeping everything absolutely motionless posed something of a challenge, even in a lab environment atop a couple of tons of air-suspended granite table.

Stipulated: all the tech has changed since then.

I wonder if the galvo motor mount has a slightly loose screw or not enough knockoff Glyptal. Might be easy to try without leaving much evidence.

(*) It was a one-dimensional setup; I was writing the track-following code to nail a 1 µm laser spot onto a 1 µm line while it danced ±100 µm. The optics guys sweated bullets just coaxing the spot into existence.

3

I am strongly suspecting the ground on the motor driver. It’s a differential analog voltage. I found references online to this being a problem.

Differential has noise immunity but if the ground is loose, then when high current pulses create a voltage drop on the ground the diff signal can go out of range of the receiver stage. This makes sense.

The seller did get back to me and is sending a new head. I’m begging them to give me permission to break the warranty seal on the head and just reseat the cable inside.

Initially I suspected one of the trimmers on the driver for PID parameters was set too high, but reading into that, the main consequence of that would be overshoot/undershoot or ringing. Outright on/off oscillation doesn’t really line up with those functions

4

Which would confirm a good debugging rule: It’s always the connector. :grin:

If you can keep the deader, definitely poke around inside!

1 Like
5

Look up the specs on that particular galvonometer. Some of the high power galvos are not designed to run that fast.
Just curious, did you try slower? I looked at one that maxed at 2000mm/s

6

Do you have a model number for the galvo head? Or post a picture of your specific one.

7

It was an SG7110, rated 5000mm/s marking/8000mm/s travel.

they sent a replacement head and it works

I have to send the original back with the warranty seal unbroken

I do see there’s a “high speed” version, the SG7210. 10,000mm/s marking speed. It would be a $450 upgrade, but doubling the machine speed would be worth a lot.

The M7 MOPA can run pulses as high as 4000KHz, so even at 10,000 mm/s that’s 0.0025mm spatial period. The pulse itself is only a fraction of that, but the spot size on a typical lens is considerably larger than that so the performance should still be fine.