Heated Bed Thermistor Calibration

[Mooselake]

Getting reasonably accurate heated bed and extruder temps seems to be an ongoing problem. Marlin makes it relatively easy to adjust it's thermistor tables (based on looking at the code, not actually doing it).

Does anybody know what thermistor(s) is/are used in the PB? It's not listed in the BOM

Would anybody with a temp gun like to measure their bed temps at several different temps 10 to 20 degrees apart, and post measured and Pronterface values? Is the error consistent or way different for different bots? How's your thermistor attached and insulated?

Not sure how you'd check the hot end temp, or if an IR thermometer would take an accurate reading at the nozzle. Does somebody know?

I'm a botless PB+ backer, so I can't do it myself. Wedding/reception commitments (should have had boys...) that are getting close enough it'll be late August (assuming my bot's here by then) to try them myself. Perhaps lincomatic or somebody else can use the data before then.

Thanks!
Kirk

[eddiema]

Getting reasonably accurate heated bed and extruder temps seems to be an ongoing problem. Marlin makes it relatively easy to adjust it's thermistor tables (based on looking at the code, not actually doing it).

Does anybody know what thermistor(s) is/are used in the PB? It's not listed in the BOM

Would anybody with a temp gun like to measure their bed temps at several different temps 10 to 20 degrees apart, and post measured and Pronterface values? Is the error consistent or way different for different bots? How's your thermistor attached and insulated?

Not sure how you'd check the hot end temp, or if an IR thermometer would take an accurate reading at the nozzle. Does somebody know?

I'm a botless PB+ backer, so I can't do it myself. Wedding/reception commitments (should have had boys...) that are getting close enough it'll be late August (assuming my bot's here by then) to try them myself. Perhaps lincomatic or somebody else can use the data before then.

Thanks!
Kirk

Bed temperature accuracy is the one thing I'd rate at a defect in my PB.
2 rough data points are 85 set = 105 on the bed.
I won't go there again but 100 hit 145 on one spot.

[CL1]

Marlin makes it relatively easy to adjust it's thermistor tables (based on looking at the code, not actually doing it).

Yes. Anybody familiar with microcontrollers will have no problem dialing theirs in better than stock. No problem in technique, that is! Getting the actual values right seems to take quite a bit of effort, based on what I observed at PBHQ.

Does anybody know what thermistor(s) is/are used in the PB? It's not listed in the BOM

It's 100k, but I don't remember the rest right now. Might be listed at the PrintrBoard schematic? If no one else chimes in, I can look it up. Laine Walker-Avina would know and hangs out at the Reprap IRC.

Would anybody with a temp gun like to measure their bed temps at several different temps 10 to 20 degrees apart, and post measured and Pronterface values? Is the error consistent or way different for different bots?

At PBHQ Botfarm the error was typically within a +10-20 degree range. So if 80 was set, expect a real temp of 90-100. More occasionally there might be one +30. Most end user reports I've seen fall into this range too.

How's your thermistor attached and insulated?

At PBHQ Botfarm the thermistors are Kapton taped under the bed in most cases with a couple soldered direct to the extended traces provided and Kapton covered.

Not sure how you'd check the hot end temp, or if an IR thermometer would take an accurate reading at the nozzle. Does somebody know?

I have both a cheap IR non-contact and a cheap contact version digital thermometer. I trust the contact version far more than the IR. Didn't take it to HQ, where they used an IR cheapie like mine that was clearly well off actual values too. But as with the range above, the error appears to be in offset rather than scaling, so after a few uses you can empirically dial in to what you need to know. But IME the IR always has more variable results than the contact type. Angle and direction and distance all need to be somewhat repeated to get somewhat repeatable results. Tempilstick or thermocouple would be best. My thermocouples are all too high range types made for kiln firing and foundry work.

CL1

[Dick Brewster]

How's your thermistor attached and insulated?

At PBHQ Botfarm the thermistors are Kapton taped under the bed in most cases with a couple soldered direct to the extended traces provided and Kapton covered.

CL1

Seems like if the thermistor is only kapton taped to the board with only a layer of kapton as insulation, that a good share of the 10 to 20 degree error could be because the thermistor is actually cooler than the board.

[Mooselake]

Seems like if the thermistor is only kapton taped to the board with only a layer of kapton as insulation, that a good share of the 10 to 20 degree error could be because the thermistor is actually cooler than the board.

That's a good point. Also a layer of glass would do the same. Hard to IR the top.

I was planning to try a cardboard insulator, always around and looked good on the insulation value tables posted a while back. It could be a problem with how flat the bed is.

@CL1 - thanks, as always. Were the botfarm beds close to the same, or do the thermistors vary a lot? No feel for how consistent cheap thermistors will be.

I was looking at thermocouple thermometers on the bay, but have self-imposed a freeze on parts/tooling spending. Don't want to spend more than the PB before it arrives.

Kirk

[Michael@Oz]

I would be good if someone can find out the manuafacturer & part number of the thermistors (if the bed & hot-end are different).

Thus we can get a datasheet.

[eddiema]

How's your thermistor attached and insulated?

At PBHQ Botfarm the thermistors are Kapton taped under the bed in most cases with a couple soldered direct to the extended traces provided and Kapton covered.

CL1

Seems like if the thermistor is only kapton taped to the board with only a layer of kapton as insulation, that a good share of the 10 to 20 degree error could be because the thermistor is actually cooler than the board.

I thought so to but I've had mine between the bed and insulation - it should be pretty close the the right temperature in there.

[Dick Brewster]

Seems like if the thermistor is only kapton taped to the board with only a layer of kapton as insulation, that a good share of the 10 to 20 degree error could be because the thermistor is actually cooler than the board.

I thought so to but I've had mine between the bed and insulation - it should be pretty close the the right temperature in there.

Thanks, that's useful information, not the answer I wanted , but useful none the less.

[scantrontb]

Would anybody with a temp gun like to measure their bed temps at several different temps 10 to 20 degrees apart, and post measured and Pronterface values? Is the error consistent or way different for different bots? How's your thermistor attached and insulated?

Not sure how you'd check the hot end temp, or if an IR thermometer would take an accurate reading at the nozzle. Does somebody know?

I'm Botless at the moment myself, but once i get mine and get it built, i can do the IR Camera angle, as i have one (it's calibrated too). I'll do the bed and the nozzle as well. currently I'm not sure if I'll go glass or aluminum plate for the bed, but either way, I'll also do a set with and without it for baseline temp studies...

[Mooselake]

I'm Botless at the moment myself, but once i get mine and get it built, i can do the IR Camera angle, as i have one (it's calibrated too). I'll do the bed and the nozzle as well. currently I'm not sure if I'll go glass or aluminum plate for the bed, but either way, I'll also do a set with and without it for baseline temp studies...

Thanks! Calibrated would be great, since I hear the cheap IR guns aren't all that great (probably worth checking mine against ice and boiling water temp non-reflective objects).

I'd like to see if there's any consistency in actual vs measured temps. Probably not, besides differences in thermistors there's different bed attachments and insulators, plus glass (needs something. like blue tape, to measure against), etc. Be nice if somebody came up with a datasheet, too.

Kirk

[CL1]

I would be good if someone can find out the manuafacturer & part number of the thermistors (if the bed & hot-end are different).

Epcos NTC G560 100K 1% from my notes. The only difference between the Hotend and HeatBed are the package types. Hotend is radial glass bead (like a grain of wheat lamp or LED), Heatbed is axial glass bead (like a resistor). If buying new, I'd use the Hotend type in both places.

Here's an Epcos general Datasheet for handling and mounting consids: http://www.epcos.com/inf/50/db/ntc_09/G ... _G1560.pdf
Here's a useful thread at Reprap: http://forums.reprap.org/read.php?146,1 ... 14,quote=1

Thus we can get a datasheet.

Should be easy now, eh?

CL1

[CL1]

I was planning to try a cardboard insulator, always around and looked good on the insulation value tables posted a while back. It could be a problem with how flat the bed is.

Corrugated cardboard was Alex's favorite. Alex started out with Brook in Brook's garage, and he built and ran the first HQ Botfarm. He told me he used cardboard at home and *tried to get Brook to use* it at HQ, while he was still there.

@CL1 - thanks, as always. Were the botfarm beds close to the same, or do the thermistors vary a lot? No feel for how consistent cheap thermistors will be.

You're welcome. Can't say for sure because while I was there too many variables in this area were changing daily, even hourly. New PCB's, firmware updates and changes, temp profile updates, Heatbed mounting and insulators, filament source, etc. Much was done by familiarity with certain machines, only to have to update that *familiar* into something new regularly. The part itself is 1%, so that should answer partly?

CL1

[Mooselake]

Epcos NTC G560 100K 1% from my notes. The only difference between the Hotend and HeatBed are the package types. Hotend is radial glass bead (like a grain of wheat lamp or LED), Heatbed is axial glass bead (like a resistor). If buying new, I'd use the Hotend type in both places.

Here's an Epcos general Datasheet for handling and mounting consids: http://www.epcos.com/inf/50/db/ntc_09/G ... _G1560.pdf
Here's a useful thread at Reprap: http://forums.reprap.org/read.php?146,1 ... 14,quote=1

Thanks! The <tm>Real Answer</tm>

Thus we can get a datasheet.

Should be easy now, eh?

CL1

Sure, just feed it into the python table generator

The second time will be easy. Won't have to figure out git, eventually do a fork, arduino IDE this box, compile, link, upload, swear at it a few times. Plus get and build a bot. Bout time I do all that again, anyway.