PT100 sensor giving wrong readings



  • I have been running my printer with a PT100 (E3D PT100 cartridge running in 2 wire mode in E3D v6 hot-end) sensor for a while now, everything working fine. When I heated the hot-end today to start a print, I received a bunch of heater errors that the heater is heating too slowly. I checked both the heater core and the sensor to ensure they are properly connected to the heater block and they were.

    Trying again, to heat up, I continued getting the heater faults, which I then dismissed as it seemed fine, but as I set the print to start (at only 210C for PLA), I saw something like smoke pouring out of the nozzle and immediately performed emergency stop (and retracted the filament). This was the first point where I realized something is wrong.

    A while later, after everything cooled down, I tried a PID Tune and measured the temperature using my multimeter with its k-type thermocouple and it showed 300+C while the firmware read temperature was leveling off at about 200C. Seeing that I immediately emergency stopped once again.

    I then disconnected the sensor and it showed copper oxidation all over its surface (normally it have a nice copper surface, this one was now black). Using a bit of diluted HCl I cleaned the sensor, then neutralized the acid with a base and finally washed it in distilled water before drying it off.

    I have inserted it back into the printer to run another PID Tune, this time more closely monitoring the temperature. At first it started with a good steep curve, which suddenly leveled off at about 90-100 before continuing on a more flat curve; my measurement with the multimeter was 200C when it displayed only about 140C.

    The multimeter temperature measurements where done in the same cartridge hole used by the sensor. The PT100 readings are stable - minimal fluctuations (less than 0.5C).

    Do anyone have any suggestions to get it working again? I have a spare PT100 sensor, but it would be my 3rd PT100 sensor in only a few months (previous one got stuck in a heater block covered in burnt PLA and its wire broke at the sensor) and the sensors are not available in my country (importing is expensive). How do you prevent the sensors from oxidizing when exposed to temperatures close to 300C (which is where PT100s are recommended)?


  • administrators

    If the readings from the PT100 are correct at room temperature, then they should be correct at higher temperatures unless one of the following happens:

    1. There is a bad connection, which increases the resistance. Using a 2-wire connection, this will cause the temperature to over-read.

    2. The PT100 develops an internal short. I don't know the construction of the E3DE sensor, so I don't know if this is even possible. If that happens then you should see a sudden drop in temperature when the short occurs.

    The sensor could develop a short between the element and the casing, but if that happens then I would expect the converter IC on the daughter board to detect and report an error.

    I get it's possible that the converter chip has failed in a way that reports an incorrect reading. You could try using the other channel of the daughter board. But I suggest you get a new PT100 sensor too.

    It's good to know that the heater fault detection was working properly!



  • The reading at room temperature is about 3C lower than my heatbed thermistor (17.5 vs 20.8); my multimeter reads a 21C (can only display full degrees). It should be noted that my heatbed thermistor might not be exact (it is not a common thermistor, but got the values from the datasheet, thanks RS Components for having datasheets on everything).

    Heated it up to 100C, and the multimeter reads 118C. It seems to follow a curve like I have seen when using a thermistor with incorrect Beta or other constants - it starts very close to the actual, but the further from ambient it goes, the greater the difference. After cooling, it once again cooled to below the heatbed temperature.

    I don't find any sudden drops in temperature readings - always a natural curve.

    I will open the machine a little bit later and then do some of the changes I have been putting off for a while now. and then test the second Pt100 channel. I have a new PT100 sensor, but they are quite expensive to import and that will be my 3rd PT100 sensor in 4 or 5 months (every time a spare sensor is inbound, either just before or just after my other sensor is having issues) on the same machine.



  • I swapped the PT100 channels and now it is reading slightly higher than the thermistor for ambient, but that is fine. I was able to perform a PID tune and it is quick like it should be.


Log in to reply