PT1000 inaccurate readings at higher temps



  • I have a K type for the heat bed and a pt1000 for the hotend. At room temp they measure the temp within .4 degs of each other and are also within about 1 deg of an external k type thermometer.

    The problem arises when heating up to printing temps. The pt1000 reads 240C, but when measuring the temp with the k type thermometer at the heater block I'm getting 260C. Compared to using a pt100 where I measured 230C at the heater block when the pt100 was reading 240C. This is also evident during printing. Swapping between the pt100 and pt1000 both reading 240C, the parts printed with the pt1000 show significant signs of being printed too hot.

    I have seen that you can tune the pt1000 with the M305 L and parameter, but that is for tuning readings at low temps. If I modify the L parameter to read correctly at 240C then the room temp reading is off.

    Any idea what is going on?


  • administrators

    @jckray said in PT1000 inaccurate readings at higher temps:

    Compared to using a pt100 where I measured 230C at the heater block when the pt100 was reading 240C.

    Please explain what you mean.



  • @dc42 I used a separate thermometer to measure the temperature on the heater block right where the nozzle on a V6 joins the heater block. Now I realize that this might give a different reading than the cartridge sensor embedded in the block, but if both the pt100 and pt1000 are reporting the same temperature, I would think that the thermometer would measure reasonable close to the same temperature for both.

    However, with a set temp of 240C for both the pt100 and pt100 I measured drastically different temps with the thermometer. The pt100 would read 240C but the thermometer measuring the temp on the outside of the heater block would measure 230C. Not, necessarily a problem, I wouldn't expect them to read the same thing, but they should be fairly close.

    The same test with the pt1000 yielded temps of 260C measured with the thermometer on the outside of the heater block.

    What I'm saying I'm experiencing a 30C variance in the measured temperature and the actual temperature between the pt100 and pt1000.

    Sorry if this is a little confusing.


  • administrators

    Which type of Duet is this?



  • @jckray said in PT1000 inaccurate readings at higher temps:

    Now I realize that this might give a different reading than the cartridge sensor embedded in the block,

    Offhand, I'd want to insulate the PT100 on the outside as a good deal of it's surface area is exposed to air, unlike the embedded sensor. If you had some of that cotton batting & Kapton tape you could temporarily wrap it around the outside of the whole thing. I think you might get a more accurate reading.

    That said, I have a PT1000, and when I first set it up I used a K-Type bead hooked to a bench multimeter (pretty accurate) and found some differences as well. PID Tuning handled it.

    The discussion was here: https://forum.duet3d.com/topic/9445/temp-woes-continue
    (the "continue" part was I had a config problem the day before that was strictly my fault)



  • @dc42 I'm using the Duet 2 WiFi rev 1.04b.

    @PuterPro Thanks, I will check out that thread.

    Since my last post, I was able to achieve reasonable results using the M305 L and H parameters. Not sure if this is the best solution, but if it works... I had to use L10 H-120.


  • administrators

    @jckray said in PT1000 inaccurate readings at higher temps:

    @dc42 I'm using the Duet 2 WiFi rev 1.04b.

    @PuterPro Thanks, I will check out that thread.

    Since my last post, I was able to achieve reasonable results using the M305 L and H parameters. Not sure if this is the best solution, but if it works... I had to use L10 H-120.

    Are you able to do some tests using fixed resistors in place of the PT1000, and compare the firmware readings against the values in a PT1000 resistance vs. temperature table? The H and L parameters are normally needed only on the older Duets; and H-120 is an extreme value.



  • @dc42 I certainly can. Do you happen to know where I can find said table?


  • administrators

    @jckray, a web search will turn up lots of them. Here's one https://www.sterlingsensors.co.uk/pt1000-resistance-table.


  • administrators

    PS - I just did some tests on the Duet WiFi on my bench:

    1K resistor reads -2.7C to -3.0C. It should be 0.0C, so if my resistors are accurate then it is under-reading by almost 3C.

    2K resistor reads 266.9 to 267.6C. It should be 266.3C, so if my resistor is accurate then it is over-reading by about 0.5C.

    These are with zero L and H correction. I then tried the same resistors on a Duet Maestro:

    1K gave -0.8 to -1.0C (under reading by 0.9C)
    2K gave 270.3 to 270.7C (over reading by 4.2C)

    The Maestro can measure PT1000 sensors more accurately. I think I may be able to improve on the error at low temperatures on the Duet WiFi.


 

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