Grounding PT100 to eliminate heater faults



  • I am experiencing occasional heater faults while printing with my Delta printer.

    The other day (4) hours into a print, my print failed due to a heater fault. I looked at the temperature graph and it appeared to be a spike in temperature.

    I'm assuming that this spike was due to interference in the PT100 cable (however I am just guessing).

    I'm hoping to resolve the problem by re-wiring the PT100/Z-probe connection to the Smart Effector using a shielded cable. However I don't know how to properly ground the cable?

    Question #1:
    Will replacing the PT100/Z-Probe cable with a shielded cable help resolve the issues with temp. spikes?

    Question 2:
    How do you ground the cable to the PT-100 sensor? Does anyone have a picture of how they did this?

    Question 3:
    Can someone recommend a 4-conductor cable to use? Most of the shielded cables I've seen have a solid copper wire for grounding, that would be to inflexible to use on a 3D printer.



  • Probably not due to interference, probably due to poor connection somewhere. Increased resistance is interpreted as increase in temperature.



  • A spike in temperature indicates a sudden increase in resistance. While it could be due to an induced voltage or current, why wouldn't it be there all the time and then suddenly appear at random times? It is more likely that the problem is an intermittent connection, either in the connector(s) on the cable or in the sensor itself.

    Inspect the connectors at both ends of the cable. Check the cable itself by shorting one end and read resistance at the other while flexing the cable. Temperature sensors often come with crimped connectors that work fine in theory, but in my experience, the crimping is often poorly done and fails intermittently. Try soldering the cable to the sensor lead.

    A shield is used for reducing noise that is capacitively coupled into the wires inside the shield. Grounding a shielded cable is best done at one end of the cable only to avoid creating ground loops.

    A shield does nothing for currents induced by magnetic fields. Protecting from magnetic field induced current is done by twisting the wires in the cable, and keeping them away from other wires that carry large currents (motor and heater wires, for example, which should also be twisted to reduce the chances of them inducing currents into nearby wires). A 4-wire shielded cable will usually have two twisted pairs. You want the sensor signal to ride on a twisted pair, so if you have a 4-wire PT100, make sure that you put one connection from each side of the sensor into each of the twisted pairs.

    Alarm and microphone cables often have four wires in a shield and are always flexible. Alarm wire is cheap, microphone cable can be expensive.



  • To ground a shielded cable that has no drain wire, strip the outer casing of the cable, and then carefully unravel the shield that is exposed. Bundle this into a wire and put a crimp terminal on. You can now attach the shield to a ground point.

    Only do this on one end!



  • Thanks to everyone for all the advice. I greatly appreciate it!

    I have tried to recreate the heat spike by wiggling the 8-pin PT100/Z-probe cable and connection, the PT100 sensor and the heater cartridge/fan cable with no luck.

    I also made a new heater cartridge/fan cable as well as double checked the PT-100 daughterboard connections.

    I have noticed that the heat spikes occur when running a bed calibration (Oddly enough I cannot recreate these spikes by moving the X, Y, or Z axes through Duet Web Control). I have also noticed that I sometimes have an issue when running bed calibrations in which the Z Probe is triggered before it probes the bed. I thought that slowing down the calibration procedure might help, however this made the false triggering much worse.

    Since the Z-Probe wires and the PT-100 sensor wires are all connected to the same 8-pin connector and run in the same cable wrap do you think there signals are interfering with each other?0_1553721040918_2019-03-27 16.59.44.jpg Would simply rewiring this connection with twisted pairs solve the problem or is it possibly something else?



  • Hi,
    I've had a great deal of problems with spiking temperature readings when I first installed PT100 sensors in my machine. they would only occur when a print had started and/or the heater was at above a certain temperature level. In my thread you can look at the temperature graphs. I solved this by using a temperature sensor with shielded leads (the shield provides mechanical shielding as well) and grounding the shield at one side.


  • administrators

    @vacalos said in Grounding PT100 to eliminate heater faults:

    I have also noticed that I sometimes have an issue when running bed calibrations in which the Z Probe is triggered before it probes the bed.

    The most common causes of false triggering of the Smart Effector are:

    1. Magnetic fields generated by fans;

    2. Some bed heaters generate large magnetic fields, and the switch on/switch off of the bed heater may induce a voltage in the strain gauge amplifier. A workaround for that is to use the B1 parameter in the M558 command.



  • Thank You,

    I tried using the B1 parameter and it didn't have any effect.

    I plan on rewiring my PT100/Z-Probe and Heater/Fan cables using (4) 4 conductor twisted pairs to see if this helps.
    But I don't know if any of the cables should be shielded and grounded. And if they should be grounded I don't know how.

    Can anyone advise on the grounding issue?



  • Also,

    I am having a difficult time finding (4) conductor twisted pairs that are not shielded. Would it cause any problems to use shielded cables without grounding them?



  • @vacalos regular UTP, ethernet cable, it would probably be best to cut one of the short ones made to connect to the devices because as far as I know they are made with stranded wire, so it's more flexible. I don't know how durable they are in a constantly moving printer.



  • Cat5, Cat5e, and Cat6 ethernet cable are typically solid core wire, and are not intended for constantly moving applications. That said, they would probably be fine for a number of years without failure, but it's likely that some place in the cable it would eventually fail.



  • @supraguy said in Grounding PT100 to eliminate heater faults:

    Cat5, Cat5e, and Cat6 ethernet cable are typically solid core wire, and are not intended for constantly moving applications. That said, they would probably be fine for a number of years without failure, but it's likely that some place in the cable it would eventually fail.

    Both Cat5 and Cat6 are made as both solid core and stranded. Cable used for installation is typically solid core, and cable used for patching is stranded.


  • administrators

    @vacalos said in Grounding PT100 to eliminate heater faults:

    Also,

    I am having a difficult time finding (4) conductor twisted pairs that are not shielded. Would it cause any problems to use shielded cables without grounding them?

    You could solder a wire to the shield and ground it to the ground pin of a spare endstop connector.

    We'll try to include a ground pin on the next version of the PT100 daughter board.



  • @vacalos

    Hi, as dc42 pointed already out, I can - at least for me of course - confirm that using GND(dc) instaed of PE(from AC) was in my case better. To avoid a catastrophal event if that ground should ever touch something it should not, I put a resistor in the cable that runs from GND to the PT100-shield

    If you are looking for TP-cables besides ethernet-cables for moving applications have a look at (sorry as somebody living in germany these are the things google threw up when googling in german language...):

    • unitronic FD CP (TP) plus, EDIT: this has 1 overall shield
    • metrofunk kabelunion Li2Y(C)H-C11Y 2x2 AWG24, EDIT: this has the benefit of shields around the twisted-pairs + an overall shield
    • "single pair ethernet" for moving applications with two of them in one cable -> 2x 2x ... TP

    If you do not need those kind of cables, alarm-wires are often twisted-pair and seem to be cheaper then the often twisted microphone-cables I read in this forum at some other place... some industry-cable companies also sell ethernet cables for moving applications if searching the web and or electronic-reseller-websites you will find lots of those...

    Other things that might help:

    Question to the whole forum: Could anybody give me a hint how to come up with a macro that optimizes the fan(s)- & heater-frequency to something that has the least possibility to interfere with the PT100-frequency?


  • administrators

    @lb said in Grounding PT100 to eliminate heater faults:

    Question to the whole forum: Could anybody give me a hint how to come up with a macro that optimizes the fan(s)- & heater-frequency to something that has the least possibility to interfere with the PT100-frequency?

    See my reply in the other thread.


Log in to reply