HE280 temperatures swing wildly

[Stimpy]

I am seeing issues on my Rostock MAX v2 which I upgraded with an HE280 years ago. It worked fine for a while after the HE280 upgrade (once I got a replacement HE280 due to accelerometer problems on the original). It sat idle for a couple of years due to life circumstances changing and me not having time to work on/with it.

Now I have an odd problem. When idling at temp (I seem to see it more often when at temp of 220-240) and more often when printing at temp, I will get some wild temperature swings occasionally to the point where the thermal fuse cuts to prevent meltdown. I have attached a graph of the temperatures from my latest print attempt where this occurred although I canceled the print prior to the thermal fuse cutting. The odd thing is that sometimes the temperature will stabilize and stay fine for hours, other times it will start fluctuating within 10 minutes of heating up, sometimes it will fluctuate when the hotend is not in motion other times I can only reproduce the issue while printing. I have checked the resistance of the hotend thermistor cooled and with room temp around 25-26C and it is showing as around 80k ohm. It is not the newer thermistor design on the HE280 with the plug in connector. It is the one with the copper compound holding it in the hole on the heat block with a screw into the heat block for wire retention and soldered on connections to the accelerometer board. I strongly suspect that these temperature swings were causing "def" errors that I was seeing prior to setting up octoprint on this printer yesterday so that I could graph the temperatures. Now with octoprint I don't see the def errors as I suspect the safety shutdown process in octoprint is resetting the printer too quick for me to get to the printer to see "def" on the LCD. The "def" errors were always preceded by significant temperature swings prior to setting up octoprint, my theory is that those swings eventually were getting high enough that the thermal fuse was disconnecting the hotend and causing the "def" a power cycle of the printer quickly after seeing "def" would show significantly higher than target temperatures and I could smell and hear the filament as though it were nearly burning.

About the attached graph:
The red line is the target temp, this drops after an hour of being stable due to me disconnecting from the printer for a moment before starting the print. Where it returns back to 235C is when I started preheating for the print, the time prior to this was just a temp stability test and shows that the hotend can maintain a stable temp for a while (sometimes).
The purple line is the max recorded temp from the hotend
The yellow line is the min recorded temp from the hotend
The values are grouped at 20 second timed intervals which is why there is a min and max value instead of just an actual value for each point.
The tallest peak is the purple hitting 278C (remember the set point was 235).
The shortest valley after starting the print is the yellow dropping to 202C.

Is 80k ohm normal for a warm room temp on this thermistor?

Is it likely a thermistor issue when as you can see in the graph I managed to hold temp fairly well for over an hour with no issues?

Should I be looking at something else as the culprit?

I have already re-flashed the firmware with the latest from SeeMeCNC following their instructions to the letter. I found it a little odd that the thermistor type in Configuration.h defaults to 97 but I didn't change it as other posts indicate that this is correct even though it is a 104-GT thermistor.

Also after re-flashing the firmware I did run 2 or 3 PID tunes with setpoints near or at my current target of 235C and witnessed the EEPROM values changing after each.

[geneb]

I would suspect a dodgy connection in the thermistor wiring. It could be that there's a wire working its way loose out of the whip connector, or the connector itself is coming partially unseated - although that wouldn't explain the swings when idle. Check the RAMBo end of the wiring as well.

g.

[Stimpy]

Will check more of the wiring later today, thank you for the suggestion on checking the RAMBO end as well. Something else I have noticed is that at room temp (around 25C) the thermistor is reading 80k ohms this seems a little on the low side wonder if that is the problem?

Also I have only noticed the wild swings while idle after a little while (10min-1hr) of having the hot end set higher than around 230C (235 or 240 seem to reproduce pretty reliably). I wonder if the thermal fuse is faulty and cutting the connection intermittently causing the drops in temp and then the RAMBO is overcompensating causing the peaks when the fuse reconnects? Any way to test the fuse? I have continuity across it at room temp obviously not sure of the best way to watch for a possible disconnect that is brief while at temp though.

[geneb]

I actually didn't know the fuse would reset. I thought they were single-use only.

You could try jumping the fuse for a print to see if that makes a difference, just take care to ensure that the jumper doesn't come into contact with the cooling fins.

It's also not a good idea to leave a hot end hot for a long time without printing with it.

g.

[Stimpy]

After working with SeeMeCNC Support (thank you Steve!) we found that the issue was with a faulty hotend thermistor connection. One of the two thermistor wires feeding into the 8 pin connector only had a couple of conductor strands actually in contact with the conductor in the connector. This was found by hooking up a multimeter on tone mode to the disconnected RAMBO side plug on one end then onto the small hole on the 8 pin. When moving the wire whip around the tone would occasionally stop indicating a broken connection. I tested both thermistor wires and the outer most one of the 8 pin connector was the culprit. Apparently when I installed it I broke a few strands of the conductor on the wire resulting in a bad connection that later failed due to movement stresses. The fix was to trim the problematic wire, strip it again (it had enough slack) and reinsert the fresh conductor to the connector.

[geneb]

\o/