Hi All,
I have DIY co2 laser(90W), using an MKS base DLC32 2.1 with grbl ESP installed and a S&A 5200TH chiller. When the chiller reaches it’s max set temp and begins active cooling (the fans turn on, it gets louder etc) My laser hangs, i.e. lightburn is forever in the sending data state, but there is no movement on the laser head, laser turns off. I have tried moving the USB to and from a hub(not an active hub though) and I have tried running the laser and water chiller from different sockets but nothing has resolved this. Is this EMI at play ? does anyone have suggestions on what I can try ?
All help appreciated
Absolutely!
The chiller compressor draws a huge current when it start, which either browns out the power supply for the laser controller, glitches its USB communications, or both.
Using the shortest and highest quality USB cable, run directly (not through a hub) from laser to laptop may help. Cable with ferrite slugs seem to have higher internal quality, although I’m not convinced the ferrite actually does anything.
This is a good test, but the outlets must be on different branch circuits with separate breakers back in the panel. If they’re on the same branch / breaker, there’s not much isolation between them.
Running either the laser or the chiller (but not both!) from a beefy UPS will probably provide enough isolation.
3 Likes
Thank you very much @ednisley ! Really appreciate the help. I will look into a suitable UPS. The sockets are on the same branch so that explains why that didn’t work.
I had an EMI issue on a go kart motor resetting a controller which I fixed by putting a couple of big caps(~2200uf) in parallel between vin and ground, would something similar be worth a shot in your opinion? Thanks again for the help !
The real problem happens on the AC line, where energy storage caps aren’t possible.
As an intermediate step, though, you could add a somewhat smaller 'lytic cap across the controller’s power supply to help it through the brownout. I think you’d end up soldering it to the power jack pins on the bottom of the board, so things could get ugly.
Also, verify the controller power supply isn’t underqualified for the job. Add up the peak current for all the stepper motors, double that number, and round up to the next even number of amps. If the power supply is smaller than that, you need MOAH POWAH!
(That may seem to be a surprisingly high number, but switching supplies have terrible transient response near their upper current limits: nothing succeeds like excess.)
1 Like
I’m a fan of simple hacks like this. I hadn’t considered just putting a capacitor on the DC input for the controller to fill in the low spot or trap the noise. A replacement socket, a power plug and a capacitor might be an easy plug-in upgrade / accessory and it would be nearly universal. I’d probably also throw a Zener diode into the mix to clamp high-transients that would otherwise pass through. We don’t know what those power-bricks emit under duress and power loss.
There’s typically a 'lytic cap on the board, but it’s the smallest / cheapest one that didn’t let the board fail very often while they were getting it running. Adding more capacitance rarely goes wrong, although at some point the additional inrush current can cause problems with the wall wart / power brick / whatever.
It ain’t pretty:
We have each (laser & chiller) on their own circuit. Never had an issue.
This topic was automatically closed 30 days after the last reply. New replies are no longer allowed.