Answered Power supply or overheating?

I got my first T-glase print to work and it was more than half way through when the lights went out on the 3D. The fans were still running, and the MatterControl console did not report a disconnect. After about 5 minutes I called support, and while I was waiting to connect, the lights turned back on. However, the print was labeled "Done".

I took the bottom cover off and it smelled hot, but everything I touched was not hot. I had the bed temp at 70 and the extruder at 238 and the final result looks really good. I also had 100% infill because I'm interested in the optical properties of the solids.

Has anyone seen this before? Did I over tax the power supply? I can reduce bed temp and slow down the printing even further (it was 16 mm/sec), but since no parts were actually warm I doubt that was the problem.

The print was the MatterControl test ring. Hard to beat for simple, but it should not have taxed the printer.

So I tried again after checking the fuse, and it ran for about 40 minutes, the LED's on the side started to flicker, and then they went out and the servos stopped moving. Again, no disconnect, and I could hear a fan still running. After 2 minutes it started fluctuating, and then the LED's came back on.

My Robo is not happy. At least it still under warrantee!

 

Makes sense on the USB, I figured the microprocessor was powered via that input if it is available.

 

Well I am really bummed - the second power supply failed the same way as the first one, but it took a lot of printing before it happened.
I have requested another power supply, but I suspect there is some other root cause. The temp of both bed and nozzle are being held at +/- 1 degree C, the motors sound fine and the supply just faints. After a couple of minutes it recovers, but of course the print is lost because the driver (Octoprint in this case) has lost control of the temp and gives up. Since the USB is powered, the Marlin is still reporting the temperatures, so in theory it could recover. But it should not fail in the first place.

I can attach a meter and monitor the current, but I'm wondering if there is some other thing to monitor or check. Any ideas appreciated!

 

Thank you! I was wondering about that fan - I will suggest that to support. If this is a common problem since others have seen it, then a much bigger fan or more holes in the case may be necessary. I don't think dirt is an issue, but air flow certainly could be - especially if that fan cuts out.
Mike

Edit: I took the cover off and listened to the fan - it was ticking. So then I looked carefully and noticed that one of the support arms was broken and the fan itself was rubbing on the housing. There is a connector directly above the fan it looks like there was some violent contact some time in this Robo's history. I will ask for a replacement - thanks again for pointing me in the right direction!

 

I had a 12V 2.5" fan from a broken 48 V supply. It moves a lot more air even at less than 1/2 speed, and it does not have a cracked frame. We'll see what happens over the next month. Thanks for pointers - that fan was definitely a problem!
Dr. mike

 

The printer has been running for 3.5 hours with 1.5 hours to go. I've never gotten this far before, either due to bugs in my gcode or hardware failure - usually it would fail after 40 or so minutes.. So thank you!! I will get a spare from Robo since it is under warranty, but I hope to never need again.