Adjust temperature from one printer to another



  • I have 3 duet3d all with pt100. I been testing the real temperature because I saw differences on each printer. Using a tempestik of 220C I was able to see a 10-12C difference from A-B-C printer.

    I would like to know if it’s possible to calibrate or readjust the temperature table. This way I would increase the print repetition and will remove the manual adjustments. Mind that I do much print repetition for my shop so repetition is key for me, and even 8C can change the gloss and finish of very small details. (This is critical for me)



  • Machine to machine calibration is tricky, but having the facility to allow for this in the firmware would be useful. A correction table approach could be used to adapt the user defined set points to control values and correct the raw values from the temperature sensor back to the value presented to the user. You could achieve the same result with a correction table similar to what you would get with a torque wrench or other bits of tested equipment.

    The other current thread that discusses GCode portability has very similar objectives to what you are asking for. Assuming you've got your x/y/z axis calibrated, square and corrected and your bed physically aligned as well as possible your sources of machine to machine variance would also include how tightly your extruder grabs your filament which could be a sod to calibrate.



  • ...measuring the actual temperature is an awkward thing to prescribe. What calibrated temperature sensor to use? Different advice for different hot ends? After all the temperature of the hot end isn't really the issue. It's the temperature (more accurately temperature profile across the nozzle area) the polymer reaches by the time it exits the nozzle for a given extrusion rate.

    Edit: I realise your use case is a little easier to control with the same machines I am just trying to stoke the conversation about how it can be made useful to others. Make the feature more generically useful and the chance of getting the request accepted increases dramatically! 🙂



  • @doctrucker

    I use a tempilstik 1% margin of error. I rub it on the same place on the 3 printers and I can literally see at which point each hotend reach the temperature. Since I don’t mind that the value is different than reality is ok, what I want is to adjust the 3 printers so they reach the point at the same time. Mind that ofc the 3 printers use the same heaters and sensors and hotends (um2 olsson block).


  • administrators

    If the difference is caused by differences in the exact resistance of the PT100 sensor, then you can correct for those using the R parameter in the M305 command. The default value is 430 ohms (the value of the reference resistors on the daughter boards). Reducing the R value will increase the temperature reading. For example, changing it to 426 ohms should increase the temperature reading by about 2.5C at room temperature and more at higher temperatures.

    If you are using 2-wire PT100 connections and the difference is caused by differences in the resistance of your wiring, then use 4-wire connections instead.



  • @dc42 very interesting! Was thinking about touching that but wasn’t sure how works. I have 3 wire pt100s



  • Out of interest what is the margin of error introduced by the duet components when reading PT100, PT1000, and the standard thermistors at around 200-250C?


  • administrators

    @doctrucker said in Adjust temperature from one printer to another:

    Out of interest what is the margin of error introduced by the duet components when reading PT100, PT1000, and the standard thermistors at around 200-250C?

    Using a 4-wire PT100 connection, the error introduced by the Duet components at those temperatures should be no more than 0.5C for the MAX31865 chip plus another 0.5C for the reference resistor tolerance.

    Using a PT1000, the maximum error will be greater. On the Duet WiFi/Ethernet it could in theory be as much as +/-18C because of ADC gain and offset errors, but in practice it seems to be much better than that - I have always seen better than +/-2C in my tests. The Duet Maestro is better optimised to use PT1000 sensors and it should generally be accurate to within about +/-2C.

    You can test the accuracy by substituting 0.1% tolerance fixed resistors for the sensor and using a PT100 or PT1000 resistance/temperature table to see what the reading should be.


Locked
 

Looks like your connection to Duet3D was lost, please wait while we try to reconnect.