Signs of a failing Thermistor?

Hi, I'm printing now for a few weeks with the G3D Budaschnozzle and I've had good results.

Now since a few days I have some weird effect in my Repetier Host, the temperature goes up less quick than before and eventually when it gets around 210º, suddenly in a split second it jumps to 189º and back to 233º, not in a linear fashion, it just displays those three values withint a few seconds.

Since a temperature can't possibly fluctuate that fast, I guess there must be something else wrong. It's either the firmware interpreting the values weird, or it is the thermistor beginning to fail.

When my hotend reaches target temperature: 227º, it doesn't fluctuate that much anymore, although, since the fluctuations have started, the temp sometimes drops to 204º and then quickly climbs back up. I don't know where to look!

I am thinking, it can be: thermistor, ceramic heating element, the mosfet that supplies power to the hotend or wiring. I have checked wiring and haven't done anything funky to it.

Running a Sanguino with Marlin.

I can videotape the fluctuations, but I think it's clear already

Hmm, I have very similar symptoms, with quite different setup, so it may not be firmware or electronics related. I will try a spare hotend and see what happens.

Thermistor is a voltage sensor.

Very susceptible to noise from high-power sources. (like a heating resistor)

This is the mostly likely reason you are seeing the temperature jump around.

I don't think so, because it wasn't doing it before, and the firmware does filter out high-frequency noise on the ADC. It also happens when the extruder is heating up and is 100% on, there is no noise likely to be generated by the resistor, but it could be picked up from somewhere else I guess.

I might expect a small DC offset when the power drops when the heatbed turns on.

The failure mode is that the temperature abruptly changes for several seconds, then flips to another value. I suspect it is thermal stress in the thermistor which is changing the value. When the hotend reaches a stable temp, the temp reading also stabilises.

I will also take a look at the signal with the scope.

Which thermistor and thermistor table have you got selected?

The fact that it's jumping around could mean that you've got the wrong table selected, or that the table is not calibrated well (gaps between points are too high at the higher temps).

One thing that can definitely cause temp problems is if you have any sort of electrical connection between the hot end element (resistor or cartridge type heater) and the thermistor leads. Specifically, some of the ceramic glues that tend to get used on hot ends can conduct enough to throw off readings (or in severe cases, enough to fry the thermistor input).

As such, I tend to avoid covering the ends of the resistor with the ceramic glue I have (which was from MakerGear) so as to avoid this issue, and I'm very careful with the thermistor as well. If either are held in with a grub screw instead of glue, just make sure to not over-tighten them, particularly the thermistor.

If the thermistor cracks, it'll most likely fail 100%, but I've not seen a lot of failure modes of thermistors to be able to comment on that.

Note: I've done a bit of testing of some of the ceramic glues (active component is usually sodium silicate), and this effect lessens over time and use, but it never truly goes away. And while the resistance may be something of the order 10Mohm, this is still a reasonable amount compared to the average 100k of a thermistor, and worse if you take into account that the hot end heater runs (usually) off 12V.

The whole thing is that my printer didn't do this before, so thats rule number one.

Have checked all my wiring, there are no shorts and I have no idea how that could possibly happen since with a Budaschnozzle, the hotend wires are so incredibly well kept in place by the plates above the aluminum heater block..

These temperature fluctuations are logically not a power fail, since they are going off 40 degrees celsius within one second, so it's more likely to be a read failure than a heater cartridge failure. Temperatures don't fluctuate that fast. You cannot 'force' it to instantly cool down a 40º celsius by disconnecting the power source. 40 degrees in a second is impossible. So I am clueless, but am thinking towards the hotend thermistor..

Also I ordered something from QU-BD to insulate the heater block, would be here after 8 days, but now after two weeks, still nothing

If you suspect its not from thermistor, then could disconnect the thermistor completelly, and check if there are still fluctuations then perhaps then its noise from adc comparators or something similar. Could be noise picked up by thermistor wiring, in which case may separate parts of it to test separatelly, ex. stepper outputs and/or heated bed.

I'm starting to suspect something is weird on my Sanguino. Today after replacing all the wiring, my Z-axis moved up when I told it to look for home... After restarting the printer a few times, and playing with the wiring, it worked again. I still don't suspect any wire problems, I have measured the connections that are giving me trouble, even wiggled them while measuring and I couldn't find any disconnections.

I notice my hotend is also heating up slower than before now. With half degree increments instead of a whole to two degrees increments. Maybe it's the ceramic cartridge failing, but I can't believe it has anything to do with it since the quick temperature shifts are impossible to happen that quick. Also checked the connections for the ceramic cartridge, it's completely insulated, doesn't make any shorts, not even on the hotend. I really hope my QU-BD order comes in one of these days. I've been waiting almost 14 days now

Hi,

I have a very similar problem. however Im running on a ramps 1.4 nad if Im also using the G3D nozle.
Although this is an old topic, has a solution been found?

In my case the temp reading drops instantaneously and is held there (or jumps around at a faulty temp) for a random amount of time and then the actual temp value is read, in some instances (more often than I should like) the actual hotend reaches a temp value which is too high and then the firmware auto-shutsdown the print. This leads me to believe that the board still recieves the temp signal and corrects according only that the reading mechanism must be the faulty part.

I should add that I also changed the pin to which the hot end thermostat was connected to and I am still getting this weird drop.

---

I thought what I'd do was

Check that you haven't got a resistance between the hot end and the thermistor somehow. Shorts are bad, but a high resistance (say 10K-10Mohm) is even harder to debug. Unfortunately many glues suitable for use on hot ends can be somewhat conductive. Where possible, avoid applying any to pins on either the hot end heater (eg: resistor leads, etc) or the thermistor leads.


The thermistor is (usually) anything up to 100Kohm at low temperatures. If you have any resistance between the hot end (which is 12V normally unless you're running a higher supply) and the thermistor, this will feed into your thermistor signal line, skewing the temps. If that voltage on the thermistor signal line gets too high, you can actually damage your electronics (most don't have over-voltage protection on the thermistor pins).

hi I'm not sure what you mean with the resistor being between the thermos tor and the hot end. I've connected them both to the RAMPS 1.4 board directly haven't soldered on any resistor, so unless the board comes with it pre soldered l wouldn't know.

As for glues I haven't used any anything has been soldered on, I don't know about the hotend heater it came attached to its cables.

And I have set my voltage supply to 13 Volts (I used to have it on 12 volts chwnyiy voltage had no effect on this phenomenon)

Not an actual resistor. Some high-temp glues can act like a resistor if they are in contact with the metal leads of either the hot end heater or the thermistor.

Use a multimeter, and measure the resistance between one of the thermistor leads and one side of your hot end heating element (whether it's a resistor, nichrome wire or a heating cartridge). If that is >100Mohm, try measuring the other side of the thermistor.

If you have a link to the type of hot end you are using, I might be able to explain it better (using it as the example)?

Edited 1 time(s). Last edit at 02/28/2014 01:40AM by Cefiar.

Hi,

The hotend Im using is one purchased from 3dgadgets: Budaschnozzle 1.1
with a ceramic cartridge heater : ceramic heater

I measured the resistance in the following manner:
whilst powered down and disconnected from the power source:
Result = as low as 1.45 MOhms (although the value slowly decreased, and I tried this between both thermistor pins and both hot end power connections on the ramps board (so neg to neg, pos to pos, neg to pos, pos to neg))

when the power supply is connected this value drops to about 1000 Ohms

when connected and both hot end and heat bed engaged in heating the values drop (to about 1.40-1.20 MOhms) as it reaches the desired temp at which point it shuts of.

---

I thought what I'd do was