Bed heater tuning dead time question



  • I just ran a successful bed heater tuning cycle on my new build (auto tune PWM 0.5 completed 3273 seconds, max temp reached 123.6deg C) with a warning that the heater may be too powerful, predicted to reach 218deg C which I don't care about too much as I have a thermal cut out ready to install. The values I ended up with were: gain 198.7, time constant 1205.5, dead time 4.8 (which is surprisingly low) so that's what I put in my config.g i.e. (M 307 H0 A198.7 C1205.5 D4.8 B0).

    It's quite a big big plate 400mm x 400mm x 10mm thick aluminium. In hindsight, this is much too thick and will be changed when I get the time and money. The temperature sensor has been moved from between the heater and plate, to a hole in the side of the plate, 40mm deep and close to the top surface. So there is initially a big lag between the heater coming on and the sensor seeing a change. If I try using the values above, after about 15 to 20 seconds or so I get a heater fault error - temperature rising much more slowly than expected. I'm a bit concerned about the dead time at 4.8 seconds. It's nothing like I observe when looking at the web interface. I can't really see any change in temperature for the first 30 seconds after starting the heater so I'm not sure where the 4.8 seconds came from.

    Anyway, I've tried upping the dead time from the 4,8 seconds I got doing the tuning to 30 seconds and that seems to have cured the slow temperature rise error. The question is, is that an OK thing to do or will it screw something else up?

    Thanks
    Ian


  • administrators

    Increasing the dead time is OK, at worst it will lengthen the PID response time. Reducing the dead time could lead to unstable PID.



  • Thanks David,

    I'll set it to the minimum that doesn't cause a slow temperature rise fault.

    Ian


 

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