PT100 vs PT1000



  • As I get closer to starting my "forever printer" corexy kit, I'm obsessing pretty heavily over all the details, which is part of the fun for me.

    I've come across this thread on the matter:

    https://forum.duet3d.com/topic/13344/pt100-vs-pt1000?_=1594460364837

    On my previous build I went with the E3D PT100 in 4 wire mode and the daughter board, but for the sake of efficiency (2 wires/no daughter board) I'm considering a PT1000 this time. One for the hot end, 1 for the enclosure, and I'll use the built in thermistor in the Keenovo silicone heater mat (unless someone has a reason why I shouldn't).

    I actually like the look of the Triangle Labs units, as I thought the E3D PT100 wires looked a bit flimsy. In fairness I've had no problem with it either.

    Here's their PT1000, good to 400degC which will certainly do me.

    https://www.aliexpress.com/item/32859917151.html?spm=a2g0o.detail.1000023.23.7c893064DyI2vX

    Only problem is that it mentions changing the "pull up" resistor on the mother board for best performance, and while I've got soldering gear and reasonable skills in that area I'm not so keen on doing that, especially without using air soldering gear.

    They even show pics of it being done on the Duet board, is this worthwhile?

    My reasons for doing it would be that I like the idea of 2 wires right down to the board and no noise interferance, but +/- 5degC accuracy would not be acceptable for me. I've never needed a plug inline on my other printer as it's never failed and I can just loosen the grub screw and take it out of the block if I want to remove the hot end.

    They also have the PT100, and recommend 4 wires like I did on my other printer:

    https://www.aliexpress.com/item/32828593452.html?spm=2114.12010615.8148356.31.72c62166kwq6N0

    I'm just not so keen on doing the 4 wire thing again, although if I do I'll use a 4 core shielded cable as a preventative measure. I forget how I set up the 4 wires and it's in the cable mesh wrapping, I'll have to have a look.

    I will say the the PT100 on my other machine has worked well and seems very precise, only within a degree of the bed heater thermistor at room temp, and both have worked for years.

    Any thoughts or recommendations?



  • @Corexy said in PT100 vs PT1000:

    Only problem is that it mentions changing the "pull up" resistor on the mother board for best performance, and while I've got soldering gear and reasonable skills in that area I'm not so keen on doing that, especially without using air soldering gear.

    Maestro and Duet 3 already have this change. (edit 2k2 instead of 4k7 iirc)



  • @Corexy Personally, I wouldn't bother with PT100s again. You might get a more accurate temperature reading but that's debatable. But does it matter? We tend to set the hot end temperature to whatever works best for any particular filament, so does it ally matter if the displayed temperature is a couple of degrees different to the absolute value? My opinion is that you need to buy an additional daughter board which will give you a more accurate reading, which you'll largely ignore and simply set the hot end to whatever temperature works best. So personally, I'd spend the cost of the daughter board on something else.



  • @deckingman said in PT100 vs PT1000:

    so does it ally matter if the displayed temperature is a couple of degrees different to the absolute value

    couldn't agree more, at least for my own hobbist use case.

    on a larger scale it would make more sense to have both accuracy and precision in the temperature readings across multiple printers.



  • @bearer said in PT100 vs PT1000:

    @deckingman said in PT100 vs PT1000:

    so does it ally matter if the displayed temperature is a couple of degrees different to the absolute value

    couldn't agree more, at least for my own hobbist use case.

    on a larger scale it would make more sense to have both accuracy and precision in the temperature readings across multiple printers.

    In principle, yes. But even then it's debatable whether the typical variation between thermistors would be significant. For one thing, those multiple printers would need to have identical hot ends, and part cooling solutions. For another thing, the temperature sensor doesn't actually measure the temperature of the filament as it exits the nozzle. It just gives an indication of the temperature at a particular location on the hot block, some distance away from the nozzle tip and generally shielded from any part cooling air which might be blowing across, or deflected towards, that nozzle.



  • OK,

    So I'm using the Duet wifi, and I actually already have a daughter board here from years ago, brand new in the packet.

    I just don't like the idea of soldering the 4 wires and having a joint in the line, or using a daughter board for that matter, but I'd really prefer a temp reading that was accurate.

    Are there any real pros/cons either way, or is it 6 to one, half a dozen to the other?


  • administrators

    PT100 gives higher resolution than PT1000 and (subject to the next point) greater accuracy. It's more affected by the resistance of the cable than PT1000, but you can avoid that by using a 4-wire connection.



  • @Corexy said in PT100 vs PT1000:

    So I'm using the Duet wifi,

    i'd say its relatively easy to solder another resistance on top of the existing one to reduce the resulting parallell resistance. another 4k7 will get you close to the Maestro and Duet3, or a 1k3 value gets you close to the 1k referenced with respect to pt1000.

    thats imo perfectly doable with tweezers and a soldering iron, no need for hot air.


  • administrators

    The PT1000 works reasonably well on the Duet 2 using the standard 4K7 resistor. It's the older 8-bit boards with 10 bit ADCs (Duet 2 is 12 bit) that really need lower value resistors. However, my calculations indicate that 1K is too low, and 2K2 is about optimum.



  • @dc42 said in PT100 vs PT1000:

    The PT1000 works reasonably well on the Duet 2 using the standard 4K7 resistor. It's the older 8-bit boards with 10 bit ADCs (Duet 2 is 12 bit) that really need lower value resistors. However, my calculations indicate that 1K is too low, and 2K2 is about optimum.

    It sounds like my OCD will only tolerate the PT100, as reasonably well is not going to be good enough for me unfortunately. And if I'm breaking out the soldering iron, I'd rather solder wires than mess with my motherboard.

    So if my PT100 sensor is 2 wires, at what point do I convert it to 4 wires? Do I make the 4 wires as long as possible?



  • @Corexy said in PT100 vs PT1000:

    Do I make the 4 wires as long as possible?

    yeah, as close to the sensor as possible.



  • @bearer said in PT100 vs PT1000:

    @Corexy said in PT100 vs PT1000:

    Do I make the 4 wires as long as possible?

    yeah, as close to the sensor as possible.

    It's so strange that it makes a difference...





  • @bearer said in PT100 vs PT1000:

    some background https://en.wikipedia.org/wiki/Four-terminal_sensing

    Copy that, thank you.

    Is there a size/type of shielded 4 core cable that's most suitable? Last time I ran 4 separate pieces of wire, and I didn't have any problems, but I understand noise can be an issue and I'd like the wires all in a single cable this time if possible.



  • i'd choose twisted pairs over shielded, but if you can have both go for it.



  • @bearer said in PT100 vs PT1000:

    i'd choose twisted pairs over shielded, but if you can have both go for it.

    More likely I've got twisted pairs here already in my box of RC plane bits. Thank you


  • administrators

    I use 4-core 7/0.2 unshielded cable.



  • @dc42 said in PT100 vs PT1000:

    I use 4-core 7/0.2 unshielded cable.

    Cheers David.

    Is that 7 strand you are referring to? Any idea of AWG?


  • administrators

    7/0.2 means 7 strands each 0.2mm diameter. https://www.canford.co.uk/TechZone/Article/MetricAWGWireSizeEquivalents may be helpful.



  • @dc42 said in PT100 vs PT1000:

    7/0.2 means 7 strands each 0.2mm diameter. https://www.canford.co.uk/TechZone/Article/MetricAWGWireSizeEquivalents may be helpful.

    So this might roll my PT100, heat break and cooling fan all into the one insulated cable? Twisted pairs too, so everyone's helpful advice has been heeded!

    https://www.jaycar.com.au/cat-5-8-core-stranded-network-cable-sold-per-metre/p/WB2020


  • administrators

    @Corexy said in PT100 vs PT1000:

    @dc42 said in PT100 vs PT1000:

    7/0.2 means 7 strands each 0.2mm diameter. https://www.canford.co.uk/TechZone/Article/MetricAWGWireSizeEquivalents may be helpful.

    So this might roll my PT100, heat break and cooling fan all into the one insulated cable? Twisted pairs too, so everyone's helpful advice has been heeded!

    https://www.jaycar.com.au/cat-5-8-core-stranded-network-cable-sold-per-metre/p/WB2020

    I would not mix temperature sensor or endstop wires in the same cable as motors, fans or heaters.

    My delta uses one 8-core 7/0.2 cable to connect the PT100 (4 wires) and the 4 wires that connect the built-in probe of the Smart Effector. It uses another 8-core 7/0.2 cable to connect the hot end heater and 2 fans. The hot end heater uses 2 wires in parallel to each end, to better handle the heater current.



  • @dc42 said in PT100 vs PT1000:

    @Corexy said in PT100 vs PT1000:

    @dc42 said in PT100 vs PT1000:

    7/0.2 means 7 strands each 0.2mm diameter. https://www.canford.co.uk/TechZone/Article/MetricAWGWireSizeEquivalents may be helpful.

    So this might roll my PT100, heat break and cooling fan all into the one insulated cable? Twisted pairs too, so everyone's helpful advice has been heeded!

    https://www.jaycar.com.au/cat-5-8-core-stranded-network-cable-sold-per-metre/p/WB2020

    I would not mix temperature sensor or endstop wires in the same cable as motors, fans or heaters.

    My delta uses one 8-core 7/0.2 cable to connect the PT100 (4 wires) and the 4 wires that connect the built-in probe of the Smart Effector. It uses another 8-core 7/0.2 cable to connect the hot end heater and 2 fans. The hot end heater uses 2 wires in parallel to each end, to better handle the heater current.

    Nice one. Especially about the heater wires in parallel, that's a handy way to keep it all in the multicore cable.

    It's another whole discussion, but I was considering using sensorless homing on the XY axis's, for the very reason of reducing the number of wires.



  • I mentioned it in my other post, but what's the advantage of an PT100/1000 over a thermistor?

    I see slice use a thermistor in their Mosquito hot end. Is there any reason I'd go with a PT100/1000 in preference to that?



  • @Corexy said in PT100 vs PT1000:

    I mentioned it in my other post, but what's the advantage of an RTD over a thermistor?

    I see slice use a thermistor in their Mosquito hot end. Is there any reason I'd go with a PT100/1000 in preference to that?

    TBH, I don't understand why Slice Engineering chose to use a high temperature thermistor rather than a PT 100 or 1000. It's inaccurate at ambient temperatures, not that it matters. Maybe they couldn't source a high temperature Prt in the USA? They do like to use locally made stuff, rather than imports.



  • Bugger it, I bought all of them. They cost bugger all from Triangle Labs, so I grabbed an assortment of PT100's, 1000's and other bits for the bits box. Was no point getting their standard thermistor cartridges, as they only rated to 280degC and if I'm buying the ludicrously expensive Mosquito/BMG combo, I at least want the option to heat it right up, even if I never do.

    Plus a nice member here sent me a spare Slice thermistor he had laying around (cheers Deckers!!), so I'm now spoilt for choices when the time comes.

    Now if only it all gets here before the bombs start dropping or the next exotic disease....


Log in to reply