Printer resetting -- how to diagnose
Hi folks! My Prusa i3 Mk3S (purchased August 2019) resets itself about 10 minutes into a print, and I don't know why yet.
The symptom is that the display flickers a few times, and then the printer resets completely.
The last thing I did before this started happening was to replace the nozzle. To do this, I partially disassembled the extruder per instructions. To rule out the possibility that I damaged high-current wires while doing this, I wiggled all the wires while the printer was running, but was unable to reproduce the fault.
My house wiring may be at fault, since I live in an old house, but none of the other appliances on the same 120VAC circuit had any trouble when the printer reset.
Is there a way to get error logs or otherwise to get some post-mortem of what happened? How should I diagnose this?
Kind regards,
Ihab Awad
The printer logs crashes and power outages but this sounds more like a power glitch or surge. First make certain that all cables are properly connected. If you have the proper equipment to monitor your power supply use that but I guess you have not ... Find an AM radio, use mains power and tune it to a relatively quiet station - talk radio rather than heavy metal, then listen as you print.
Mains glitches should be audible as interference - do any coincide with your resets?
If so check all your appliances, if something that takes a heavy load is running out of spec then it might be the culprit.
Cheerio,
RE: Printer resetting -- how to diagnose
Thank you for your help! I plugged the printer into a known-good 120VAC circuit (kitchen -- operates an air fryer without trouble) and observed the problematic behavior. This time I have it on video. 🙂 First, this is the flickering that occurs during startup:
https://photos.app.goo.gl/GcpsnXPu7AnH3ViEA
I then started a print, and it failed about 3.5 minutes into the job. The following video starts as soon as the axes start moving, but you can fast-forward to the last few seconds to observe the failure:
https://photos.app.goo.gl/g8AsxjLak3bkxVJw6
Does this still look like a 120VAC power input problem to you?
My printer is mounted in an IKEA LACK enclosure with the power supply outside the box. The arrangement has worked just fine for a long time. Since I had to disconnect the power to change my nozzle, it is reasonable to suspect that I re-connected the power incorrectly. However, I have confirmed that the power connections are secure -- I disconnected and re-secured the screw tabs and verified that the flickering persists. The following are photos of my enclosure:
https://photos.app.goo.gl/UhmWofgEcjWfzVcM7
https://photos.app.goo.gl/hDmDU584CewPNhye8
Kind regards,
Ihab
RE: Printer resetting -- how to diagnose
One more note. Upon unplugging the printer, I noticed that the wiring it was plugged into included this failed wall plate:
https://photos.app.goo.gl/1s9h3ic9DB1XVTiP6
I have no idea when this failure occurred, and will rip open the wall plate to fix what seems to be a poor connection. My circuit breaker did not trip, hence it is not clear that there was a short -- more likely just poor wiring and arcing. Most importantly, when I was setting up the printer after the nozzle change, I did not hear or smell any evidence of burning.
It seems strange that the two things (printer resetting itself and wall plate arcing) would occur together without a common cause. But I also cannot formulate a hypothesis where one would lead to the other. If you have suggestions, I would be very grateful!
Ihab
RE: Printer resetting -- how to diagnose
Your video shows like multiple resets at startup but you also mentioned failure at about 10 min (heat related?)
1)
If you have a voltmeter you might check that the the PSU powers up properly and has the right voltage present on its output at failure time.
No PSU voltage at failure time could be a PSU failure or overcurrent overheating that forces it to shutdown. If the PSU voltage is OK even after the failure you have to look at the mainboard and its connections.
2)
I would try to disconnect the "power panic wires" to the PSU.
3)
If that does not work a review of the PSU to the controller for a loose connector, bad solder joint or signs of over heating at the connector.
4)
Finally (remove power to the wall socket) and examine closely the inside wiring in the socket enclosure.
If all fail, in desperation, use an extension from a different socket, just in case....
Hope that helps.
REPAIR, RENEW, REUSE, RECYCLE, REBUILD, REDUCE, RECOVER, REPURPOSE, RESTORE
It does look like you need to check out your connections. I would like you to try the radio experiment and also check your mains voltage is up to spec.
Cheerio,
RE: Printer resetting -- how to diagnose
Hi folks,
Sorry for the delay -- I have been preoccupied with other aspects of my projects. 🙂 Thank you for all your feedback so far.
I unplugged the "panic" cable to the PSU, but that did not fix the problem.
I then instrumented both PSU power connections during the failure. I saw "sawtooth" transients in the voltage, ranging between 24.102V at the peak and 23.376 volts at the valley. These occurred frequently during operation and did not seem to cause any problems. I could not see any distinctive features in the PSU voltage at the moment of failure.
I currently have a single GCODE file loaded on my SD card and am running it over and over again, and the printer fails at exactly the same point in the print every time.
Ihab
RE: Printer resetting -- how to diagnose
Some more data points.
1. Previously, I have been using my printer in standalone mode. Now, in an attempt to figure out what it's doing, I installed OctoPrint on my Linux laptop and connected to the printer via USB. The printer consistently seems to lose connectivity with OctoPrint.
2. I updated the firmware on the printer. Now, when it fails, the screen freezes up with the LCD blinking quickly, but the temperature readouts continue to read correctly and it still seems to be controlling the print head and bed temperatures.
3. Recall that the failure happens in the same position in my test print. After the firmware update, it still fails in exactly the same place.
4. I scrolled through the GCODE file and was not able to find any obviously corrupted segments -- it all looks like valid data which "should" not be able to crash the printer firmware! The file is here, for reference:
https://drive.google.com/file/d/1VgpNySDl4b1FyIhyVfiAqXrkcu81x_P3/view?usp=sharing
Ihab
So Octoprint glitches at the same places, that figures.
What was the result of the radio test?
Cheerio,
RE: Printer resetting -- how to diagnose
Hey again! Well, so I had to look around the house to finally find a mains-powered AM radio; it was an old clock radio that my wife was about to donate to the local Goodwill! But I ran the test. The result is this video:
https://photos.app.goo.gl/ZdmzVGzP5A8aRjyZ6
And finally, my printer displayed an error code I can search on: "TMC DRIVER OVERTEMP". With no other obvious problems visible, I decided to order and install a new Einsy Rambo board. Wish me luck.
Ihab
RE: Printer resetting -- how to diagnose
Hey everyone,
I think I found the problem! I installed the new Einsy Rambo board and it (of course...) started to do the setup and self-test. During the self-test, I noticed that the print fan test caused the printer to crash, as observed previously. I looked at the print fan wiring and -- lo and behold - it was frayed, separated, and intermittently shorting:
https://photos.app.goo.gl/mbHPsTS44H2Bc99CA
So basically, at the point in the print where the print fan was turning on, there was a short circuit that brought down the whole printer!
The lesson here is: Before doing anything, run the PRUSA self-test!! 🙂
Ihab