Revo heater core, incorrect temperature?

bug2k22

I've been changing the print head on my modified FT5, it's now running a micro swiss ng extruder with a revo nozzle and revo 60w heater core connected to a roto toolboard to a HC6.

The reprap configurator suggests that E3D use the 104-gt2 thermistor, but this does not give the correct temperature as reported on the web interface. My workshop was at about 10degC and it was displaying about -0.5degC on the web interface. I checked the resistance at the connector and it reads about 208K ohms, which is about correct when checked against the datasheet on the Semitec 104NT-4-R025H42G which is the thermistor shown on the E3D website.

I've used this line to configure the temperature sensor, which is plugged in to temp0 on the roto toolboard.

M308 S1 P"20.temp0" Y"thermistor" A"Nozzle" T100000 B4388 C7.06e-8

Phaedrux

M308 S1 P"20.temp0" Y"thermistor" T100000 B4725 C7.06e-8

Your B value is incorrect. Try this.

bug2k22

Thanks for that, the value does not significantly change the temperature displayed, maybe a degree or so. If I change the B value to something like 8000+ or so the ambient temperature displayed is closer to reality, but I would think that would mess with the scaling at higher temperatures.

jay_s_uk

@bug2k22 I would get in touch with where you brought it from as that resistance doesn't look right

bug2k22

@jay_s_uk I did check on the datasheet for the 104NT-4-R025H42G and this shows that the expected resistance at 10degC is 208.8Kohms which is about correct with what I measured. I purchased this from E3D.

gloomyandy

@bug2k22 It looks like the roto board has the ability to load calibration data from NVM, I wonder if the values it is picking up are correct? Might be best to wait for one of the Duet folks to comment further before trying any recalibration though.

https://docs.duet3d.com/User_manual/Connecting_hardware/Temperature_connecting_thermistors_PT1000#semi-automatic-calibration

elmoret

With a 2.2kohm pull up and cold temperatures resulting in 208kohms of resistance, you're asking the ADC to accurately measure about 0.03 volts relative to the 3.3v rail.

3D printer thermistors/boards are designed for accuracy around 200C, (at this point, thermistor and pull-up are similar so there's loads of accuracy/resolution), they don't have much accuracy at low temperatures. I wouldn't sweat it.

bug2k22

@elmoret Thanks, yes provided that accuracy is better in the operating termperature range. I just had not seen such a lower temperature accuracy issue on the 3 other printers I built using duet boards, this is the first tool board I've used so perhaps it's related.

dc42

@bug2k22 our automated test equipment checks the reading from the TEMP pins at three simulated temperatures. The lowest of these is 25C and the maximum error we allow at that temperature after calibration is 2C. At 10C the error will be higher, but I would not expect it to be as high as 10C.

The ADC reading code was changed at firmware version 3.5.3 and this affected the calibration (see https://github.com/Duet3D/RepRapFirmware/wiki/Changelog-RRF-3.x#reprapfirmware-353); therefore if your Roto tool board was shipped with firmware earlier than 3.5.3 but has subsequently been upgraded to 3.5.3 or later then it likely requires re-calibration.

I suggest you run the calibration procedure and see whether that gives you a more accurate reading at low temperatures. See the semi-automatic procedure at https://docs.duet3d.com/en/User_manual/Connecting_hardware/Temperature_connecting_thermistors_PT1000#semi-automatic-calibration.

bug2k22

@dc42 Thanks, that seems to have fixed the issue, it now reads within 1 degree of what my infra red thermometer reads.