Nozzle temp drops during print

[DIY3D]

Hi,

I've print a lot hours without any problem. During our monthly meeting I saw for the first time the nozzle temperature dropping during a print.
That was last month and since that time I've made several nice prints.... till now. I start the print and after +/- 10 minutes the temperature is dropping, I hear the gear trying to push the filament through the nozzle. When I stop the print > clean the heated bed and start again, it reached the right temperature, the print starts en after a while I've got the same problem.

I would appreciate some advise on troubleshooting and solve this issue.

Kind regards
Marco

[Dale Reed]

Marco,

My first-blush guess is that one of the harness wires for the hot end heater has been flexed too much and has one or more broken strands, and so can't carry enough current to the hot end to keep up with demand. I'd start checking the harness flex points for hot spots or visibly blackened / broken wires.

Also, if you're in the northern hemisphere, winter is approaching and ambient temps are lower. It may be that the PID maximum output in the firmware was sufficient when ambient temps were warmer but is not sufficient how. You could try running the PID autotune sequence to see if that helps.

But my money is on the first item. The PID control should have plenty of overhead available to keep the hot end hot despite ambient being just a few degrees lower.
Dale

[DIY3D]

Hi Dale,

Thanks for the quick reply

I guess you'r right, think I've to search for a broken wire, I've read it in a few topics now. I'll start checking the cables especially at the ty-rap connections.

Kind regards,
Marco

[DIY3D]

Hi,

Well, I didn't find any damage in the wiring. Thinking it must be in the software, I tried to run the PID auto tune a few times, but strange enough it failed (Below some rules from the logfile)
This morning I sliced again my object in a previous version of Cura (the print that failed was sliced in the latest version of Cura). Now it's printing fine, the temperature is stable.

Any idea what caused the PID auto tune failed?

######################################################
22:13:17.585 : N64 M303 E0 S200 C8 *126
22:13:17.585 : PID Autotune start
......
22:46:40.738 : bias: 296 d: 103 min: 196.68 max: 201.32
.........
22:27:20.673 : N357 M105 *6
22:27:21.063 : Error:Line Number is not Last Line Number+1, Last Line:73
22:27:21.063 : Resend:74
22:27:21.079 : Resend: N74 M105 *52
22:27:21.079 : Error:Line Number is not Last Line Number+1, Last Line:73
22:27:21.079 : Resend:74
22:27:21.094 : Error:Line Number is not Last Line Number+1, Last Line:73
22:27:21.094 : Resend:74
22:27:21.094 : Resend: N74 M105 *52
22:27:21.157 : Error:Line Number is not Last Line Number+1, Last Line:74
22:27:21.157 : Resend:75
.............
22:46:40.738 : bias: 296 d: 103 min: 196.68 max: 201.32
.............
22:56:56.471 : N678 M105 *14
22:56:56.502 : Error:checksum mismatch, Last Line:388
22:56:56.502 : Resend:389
etc..
#####################################################

Regards,
Marco

[Dale Reed]

Hmmm. The sequences with line number mismatches and checksum/CRC error message look more like a general communication problem and not a problem with the PID Autotune. Perhaps an odd character or some noise got on the line and got the comms "out of sync".

Try again, and be very careful about syntax of commands. Probably nothing you did wrong.... Do you have another machine control program, like S3D perhaps, in which to try the command sequence for autotune?

Dale

[insta]

That happens with the PID autotune specifically. The autotune takes several minutes, during which point Marlin is unresponsive to inputs from the host software. The host software will continue to send "pings" to get things like hotend temperature and such, but they will get ignored. The ignoring isn't such a problem, it's that Marlin & the hosts have implemented checksumming and line-counters.

If you take the command "G0 X100 Y100", Marlin understands that, but that command offers no safety. There's nothing to ensure it arrives as-is, and in the order it was sent, except a high probability based on how the serial protocol works. The host softwares will actually send something more like "N239 G0 X100 Y100 *44" (I made those numbers up). The "N239" part is the 239th command sent, and the *44 is a checksum based on adding up the value of all the characters and dividing by 255. This will ensure correct-order delivery and that characters won't get garbled in transmission.

These two things finally come to a head when doing autotune. Marlin receives the "M303" command, and records that as the 13th (or whatever) command it's received. It then receives, and ignores, the 14th, 15th, 27th, commands for "get temperature". When the autotune is finally finished, Marlin goes back into the main processing loop, expecting the 14th command -- since as far as it knows, 13 was the last that was received. The host software sends "N31 ...", and Marlin goes "wait, 31 is bigger than 14. ERROR", and there's your message.

[DIY3D]

Well that is a clear explanation
I've noticed that the temperature drops and remain stable at 170 C. After rebooting I could do a fresh PID tune and stored the results in Eeprom
Now I've printed some test objects and everything looks fine. The temperature stays stable at 200 C

Thanks Dale and insta for reading and helping!

Kind regards,
Marco