Octoprint Unhandled Communication Error

[zemlin]

Capture.JPG (18.98 KiB) Viewed 12994 times

Been getting these. Googling turns up two schools of thought.
1) Thermistor condition or connection problem.
2) USB communication problem.

I saw lots of reports that a flaky power source for the PI can account for 90% of USB communication issues, so I moved the PI from an old phone charger to a 2.1a USB port on my UPS. I also replaced the cable and the USB cable between the PI and the printer.
The Extruder 1 thermistor was going bad and I saw a big dip in that plot at one of the shut downs, so I replaced that thermistor. I am only using Extruder 0 for these prints. The temp plots for that extruder look fine.

I will replace the 0 thermistor tonight, but I am skeptical about that doing anything.

If that doesn't do it I figure it's time to start replacing boards. PI at about $30. RAMBO at $170. The RAMBO live inside my heated enclosure, so I figure that may actually be a better candidate for failure, although 40C shouldn't be too severe for 'lectronical things.

Also thinking about moving the web cam and printer to a powered USB hub to cut the current draw from the PI.

Thoughts?
Theories?
War Stories?


Also, right now I just need to run some parts. Thinking if I can print from an SD card I might bypass the issue, but I've never tried printing from the card. Laptop with S3D -> upload to SD -> print from SD. Can S3D be shut down/disconnected at that point? Also Octoprint has the option to upload to SD. Would that allow the machine to keep printing if it's a USB error?

[ednisley]

an old phone charger to a 2.1a USB port on my UPS

That's your problem, right there. Get a 2 A power supply and watch the clouds vanish from your skies.

It need not be this fancy, but avoid dirt-cheap eBay junk:

https://www.raspberrypi.org/products/ra ... er-supply/

The power supplies included in RPi kits from Canakit work well, even if they're not Official RPi Wall Warts:

https://www.amazon.com/CanaKit-Raspberr ... B00MARDJZ4

Bonus: UL listed!

The Pi draws a low average current, so it sorta-kinda works with marginal supplies. It has a surprisingly high peak current which causes erratic (and baffling) failures from those same supplies when the voltage briefly drops below the Pi's brownout level.

Clean power is worth a couple of extra bucks ...

[zemlin]

an old phone charger to a 2.1a USB port on my UPS

That's your problem, right there.

So you don't think the 5V out on an APC UPS would be worthy? I tried it powered from a Dynex (Best Buy brand) powered USB hub as well.

[ednisley]

So you don't think the 5V out on an APC UPS would be worthy?

By definition, a compliant USB port will be limited to 500 mA before enumeration. Nonstandard ports, such as the one on the UPS, may provide more power, but that's not assured and, in any event, 2 A will put a strain on it.

Powered USB hubs have the same limitations, plus they're almost certainly not capable of delivering clean power at 2 A even after negotiation.

The RPi wall warts simply provide high current to the USB jack on the Pi, without the overhead involved in negotiating a "real" USB connection, which is what you want in this situation. The wart also gets rid of the USB data cable, which may have little teeny wires incapable of carrying moderate currents:

https://softsolder.com/2012/08/07/usb-w ... er-notice/

Just do it and report back ... [grin]

[sthone]

I have 5-6 of the Canakit ones and I've never had any problem on any of my RPi's.. well worth the $10 (just make sure you get the 2.5amp ones.)

-Steve

[insta]

MINTEMP comes from RAMBo, full stop. There's nothing to do with a Pi in that error.

[zemlin]

fyi

Send: N476432 G1 X3.398 Y182.547 E4.8347*98
Recv: Error:1
Recv: : Extruder switched off. MINTEMP triggered !
Changing monitoring state from 'Printing' to 'Error: 1: Extruder switched off. MINTEMP triggered !\x0a'
Recv: Error:Printer stopped due to errors. Fix the error and use M999 to restart. (Temperature is reset. Set it after restarting)
Recv: Error:No Checksum with line number, Last Line: 476431
Recv: Resend: 476432
Recv: ok

I'll have a new power supply tomorrow. Going out now with my DeOxit contact cleaner and will hit all the thermistor contacts - although based on the above it seems the root may be a checksum.

edit - Contact cleaner didn't solve any problems. My last shutdown did not display the Checksum error messages.

[ednisley]

the root may be a checksum.

Which smells like a USB communications failure, because USB packets have error detection wrapped around the emulated serial connection. When the serial link has problems, something else has gone horribly wrong and, on a Pi, that's always the power supply.

This is somewhat out of date, but covers the territory:

http://nordicgroup.us/rpi/power/

If there's also genuine MINTEMP error buried in that mess, then always suspect a failure in the hair-fine wires in the thermistor bead. We've seen a few cable breaks along the way, but that's not the way to bet. By comparison, the contacts in the connectors never, ever go bad.

When you hear hoofbeats, expect horses, not zebras ... but software failures have a remarkable tendency to poot out zebras, so ya never know.

[zemlin]

Didn't make it through the first layer with the new power supply. Here are the last few lines from the Terminal window. I don't see any temperature issue on that last report.
fyi

Send: N810 M105*46
Recv: ok T:244.9 /245.0 B:99.3 /100.0 T0:244.9 /245.0 T1:40.8 /0.0 @:59 B@:127
Send: N811 G1 X-13.429 Y122.750 E32.4020*126
Recv: ok
Send: N812 G1 X-13.429 Y127.235 E32.6456*121
Recv: Error:1
Recv: : Extruder switched off. MINTEMP triggered !Changing monitoring state from 'Printing' to 'Error: 1: Extruder switched off.
MINTEMP triggered !\x0a'
Recv: Error:Printer stopped due to errors. Fix the error and use M999 to restart. (Temperature is reset. Set it after restarting)
Recv: ok

I'm going to replace the thermistor on the left extruder (the one I'm running), although I see no sign of trouble from it, and I don't see any reason to doubt the PI either. Starting to feel like I need to get a Rambo board on order.

Update: replacing the thermistor didn't fix it.

[ednisley]

I don't see any temperature issue on that last report

Starting with the first "Error:1" line, everything stems from the MINTEMP failure.

Improving the power supply eliminated the spurious checksum error, which is definitely progress.

replace the thermistor on the left extruder (the one I'm running)

Because Marlin doesn't know which extruders you'll use during the print, I'd be unsurprised to discover a thermistor failure on the "inactive" extruder will also trigger a shutdown. Alas, the error message doesn't identify which thermistor produced the spurious reading.

I'd replace the thermistor on the right extruder with a known-good one, as it's the most likely remaining zebra.

If you don't have a new spare extruder, replace it with the one you just removed from the left extruder, because (most likely) that one is good.