Heat bed reaching temperatures 20°C lower than target
Hi all,
I have a i3 MK3S I have assembled myself just yesterday, and that's printing pristinely with the stock PLA. I've tried treefrog and Benchy with really good looking results. In the end I want to use the heated bed to reach the maximum temperature it can go, say 110°C. So I tried going at these temperatures with my printer.
Getting the bed to temperatures of 100°C was rather fast, I could even achieve 110°C easily and then I pushed things to 118°C, still no problem, and finally at 120°C I got a bed preheat error after some time, pretty normal.
Then, in order to check whether I was really achieving these temps (it was too good to be true) I used a thermometer with a Type K thermocouple and taped it with Kapton tape to the print bed, and heated the bed at various temperatures. I found that my readings were off compared to what the printer was sensing, by a large amount: from 10 to 25°C.
Here's a photo of the set-up while I was printing PLA at a heatbed of 60°C, supposedly. The thermometer read 50°C instead. On other corners, I had 56, 50, and 48. The hotter was the front right corner. That's more or less 10°C difference in average, but it didn't seem to affect the print as the treefrog came out really good.
Then I monitored the value in the location that I think is the closest to where the thermistor sensor is located, i.e. like 
And read the following:
- At bed saying 70°C, A is 58°C; B is 51°C
- At bed saying 100°C, A is 81°C; B is 77°C
- At 108°C, A is 88°C
- At 118°C, A is 96°C.
- At 120°C, I get BED PREHEAT ERROR.
So as you can see, from 100°C the temperature on top of the sensor is approximately 20°C lower than what it should be. The evenness is also not good, with consistently more than 5°C difference between places. I've tried to use a thermal IR camera but the readings are not very accurate since I don't know the emissivity of the steel/PEI sheet. If anybody has made tests for it, I'm also interested.
For me, a 20°C difference isn't acceptable as this way I can barely reach the temperature needed for ABS. Looks like a sensor problem, right? Has anybody encountered such an issue before?
Thanks for your help!
RE: Heat bed reaching temperatures 20°C lower than target
The bed temp feedback thermistor is on the bottom of the heat bed. Thermal loss through the bed and steel sheet is normal and expected. As is some variation in surface temperature across the sheet, especially near the edges where the area of concern isn't fully enveloped by heat coils.
My bed is about 5c to 7c lower than set temperature. Even more around the edges. I don't expect much better. Yes, Prusa could add a compensation adjustment - but is it really needed? You know your bed is a bit cooler than the setting. so increase the bed temp accordingly.
Then there's the question of the meter you are using. When was the last time it was calibrated to a standard? I don't see a cal sticker. What are it's specifications? Is the cold junction offset properly set? Etc. And the top of the hot junction isn't properly shielded, and it almost looks like the weld isn't even touching the bed surface, so you have several errors right there. Isn't metrology fun?
ps: Normally a TC is embedded in a thermal compound when attaching and measuring flat surfaces. Since yours isn't, you should expect major temp errors in your measurement.
RE: Heat bed reaching temperatures 20°C lower than target
My actual temperature reads between 76 and 81 using a fluke infra red thermometer at various points all over the sheet after a 15 minute stabilising time, yet the setting is 90 on the printer and this is what is shown, so yes, the bed temperature is inaccurate.
For most users (PLA, PETG) this isn't really an issue, but when you really need that temperature, it will be.
My suggestion is to try and place a piece of kapton tape between the thermistor and underside of the hot plate so that it insulates slightly (This should raise the temperature of the actual hotbed), and stick some insulation underneath the plate.
You will have to check the actual temperature vs reported temperature, and maybe leave a punched hole in the insulation where the actual thermistor bead is to loose some heat from there if the temperature is still to low.
Yes it does unfortunately mean having to deal with those spacers again 🙂 a wad cutter will be able to make nice clean holes in the insulation for where the spacers are.
Normal people believe that if it ain’t broke, don’t fix it. Engineers believe that if it ain’t broke, it doesn’t have enough features yet.
RE: Heat bed reaching temperatures 20°C lower than target
I'd say that it's irrelevant for almost all users, not just those who print PLA/PETG.
All printers use a similar setup, with the thermistor under the main heatbed. It's expected that there will be a difference between the thermistor and the build plate. So long as this is fairly consistent across printers, then it should be of little concern.
If you use a temperature that accurately represents the top of the heatbed, then you'll need to compensate for the filaments you use. eg: if an ABS says 110ºC, you may need to use 100ºC or so as that 110ºC is chosen based on results with the common heatbed arrangement.
RE: Heat bed reaching temperatures 20°C lower than target
Thank you all for your replies!
@tim-m30 the thing is, I would love to accept that it's lower than expected and push it to higher temperatures until my actual surface reaches this, but since I want to print at 110°C and the bed doesn't accept getting past 120°C, I can't push it beyond. I have never managed to get past 100°C by setting 118°C (after that I get the BED PREHEAT ERROR).
Your comments on the set up are absolutely valid, so I double-checked the 2 different thermocouples I had used with a calibrated one. The error stays within 2°C even at an output of 120°C. I agree for the thermal paste, I don't have any though, so I made sure that the weld was touching the surface by applying pressure on it. The tape itself doesn't help insulating from the colder environment, so I'll order some thermal paste. Since I was also not so convinced with my temperature reading setup, I tried using a straightforward physics approach and added a drop of distilled water in the center of the print bed while it was supposedly at 110°C. Water don't lie, it didn't show any slight sign of evaporation. 😣
Then I used a Fluke IR thermometer too. Thing is, I don't have the emissivity of that bed surface, so I asked Prusa Research for it directly. Hope they will reply soon. That's a very critical value to know for all those who want to read with IR. I'll add it here whenever I get it. Using the good old 0.95 I got more or less the same as @chocki, so at least I know that's not a faulty bed or anything.
Interestingly enough, I also installed a thermocouple on the bottom of the heatbed, close to where the thermistor is taped. I used aluminium foil and a Kapton tape again trying to imitate the way the thermistor was installed. The readings I had were of course higher than on the surface, but for example at 118°C I could read 101° C on the bottom TC while the one installed on the top surface above was at 95°C. So indeed, there's a lot of loss by just crossing the metal sheet.
The reason why I'm being so picky about all of this is that it's a research project where the bed temperature will play a strong influence on how the polymer develops, so as you can imagine, we have to have a reproducible and critical approach to what we're doing, and not just settle with 'whatever works'.
Interesting @martin-a22, I'll keep that in mind. To be honest I'm a bit surprised that everyone is ok with never knowing at what temperature they are printing with. As you said since all beds are more or less the same, through trial and error, at room temperature with a good temperature surface measurement set up, it should be fairly easy to come up with a table of correcting factors at different bed temperatures. If I do this, I'll post it here once my surface measurement set up is improved.
Anyway, so now comes the question of 'how to remove that bed temp error so that my surface temperature actually reaches 100°C?'. Before changing the PSU to overvolt the PCB or insulating the underside which require some more ordering/hardware change, I tried changing the firmware for it. I found this link: https://github.com/prusa3d/Prusa-Firmware/issues/1435 that gives a solution to remove the PREHEAT BED ERROR and allows you to set a higher BED_MAXTEMP also. I've just done that, it worked out well. I fixed the max temperature to 135°C instead of 125°C. So I could reach values of 130°C according to the printer. What I read with a TC at the same location underneath the bed was 112°C, and on the surface was more along the lines of 106°C but that's good enough for me so far especially considering there is no enclosure yet.