TLDR: How should the Varipower bed thermistor be described in the Duet WiFi's config.g file? I was trying to understand the somewhat arcane way the Duet WiFi does heater tuning. I started with the M305 command and I fell down a rabbit hole. I’ll show you my path and hopefully someone can correct my misunderstandings. I started with a simple question. What thermistor is in the BigBox Varipower heated bed? Let’s check with E3D! Hrumpf! E3D has removed the bed from the store indefinitely. But Filastruder still has the description of the bed even though it’s out of stock. https://www.filastruder.com/products/e3d-varipower-heated-bed-pack That says the bed thermistor is a Semitec 104GT2. Let’s go get the data sheet for that. Uh oh. It’s not on Semitec’s website… A little searching and we find Mouser still lists it, but as an obsolete (i.e., discontinued) part. http://www.mouser.com/ds/2/362/semitec usa corporation_gtthermistor-549366.pdf Mouser recommends we switch to the current Semitec part, the 104NT-4-R025H42G. Mouser is kind enough to give us that data sheet as well http://www.mouser.com/ds/2/362/NTCThermistor_NT-461415.pdf What’s the difference? Same temperature rating, dimensions are slightly different, but what about the “B” factor used by the Duet temperature control algorithm? 104GT-2-20201 B = 4300K 104NT-4-R025H42G B = 4267K There is also a close relative to the suggested replacement 104NT-4-R025H43G B = 4390K So they are pretty similar from the Duet’s point of view. But wait! Speaking of the Duet, here’s a helpful page on setting up the Duet: https://duet3d.com/wiki/Connecting_thermistors But it says use B = 4388 for the Semitec GT-104 Hmm… These 3 values only span about 3% so that’s not very much error. But wait we have one more entrant! Elmoret’s super helpful posting on how to adapt the BigBox to use the Duet WiFi https://forum.e3d-online.com/index.php?threads/bigbox-duetwifi-controller.1766/ He has this in his config.g file M305 P0 T100000 B4388 R4700 H30 L0 Aha! The value straight from Duet3D.com. It may not be right, but we can guess why he had that number in his configuration. What about the other parameters? P0 is the heater number. T100000 is resistance (ohms) at 25 celsius B4388 is the thermistor B value R4700 is the value of the series resistor on the controller electronics The R4700 parameter is “always 4700 ohms on Duet WiFi and Duet 0.8.5 controllers” per https://duet3d.com/wiki/Connecting_thermistors But that M305 command has those H30 and L0 parameters. According to the RepRap.org gcode guide these are Lnnn ADC low offset Hnnn ADC high offset Okay I have no idea what those should be! But I did find this https://duet3d.com/wiki/Changes_to_config_gcodes_and_firmware_behaviour “The M305 H and L parameters are still supported for the moment, but they should not be needed because the ADC is calibrated automatically.” So maybe we should be using something more like M305 P0 T100000 B4388 R4700 Although I’m still not sure which B value is right. It's a small effect though so I don’t suppose it matters much. Still it looks like we can simplify the M305 command by eliminating the L and H parameters. Is it okay to delete the M305 H and L parameters? But where did that particular B parameter come from? Were all the Varipower beds built with 104GT-2-20201 thermistors? If the new beds ever come in that Greg says are on order, what thermistors will be on them?
M305 H/L can be deleted. B parameter came from Duet recommendation. Yes, AFAIK. Unknown, also not sure why it is relevant? All the beta values span a very small range as you said - more error is likely to be introduced by things like the tolerance of the balancing resistor, etc. I'm also not entirely sure why it is relevant. If you set 80C because that's the recommended print temperature but the bed achieves 78C in reality while reporting 80C, I can't think of a material that would be so sensitive to matter? Worrying about the exact beta value seems like a fruitless venture to me. What about the way the Duet Wifi does heater tuning is arcane?
Thank you! That's some good information. As to why would I care, I guess that's a lifetime of engineering paranoia. When dealing with anything that has effectors that can hurt us, say ones that can set the house on fire, I believe there should be traceability as to why you configure things the way you do. Putting it another way, there should be a documentation trail as to why you believe things are configured the way they are. So, whenever I see discrepancies in numbers, it arouses my curiosity. So in a sense, I don't care what the B values are, I just want to know why they are. Many a disaster has been caused by typos. Sometimes it's the loss of a very expensive spacecraft, sometimes people die, usually we get lucky and get away with it. Again, thanks for the help!
The small range of betas - some come from datasheets, others from calibrated testing. Thermistors actually don't perfectly fir the Steinhart-Hart model, so the ideal "beta" varies wrt temperature. So Duet/David Crocker uses beta values that are a little more accurate in the 200-300C range as compared to datasheets.
