I just took the ice bin out of my fridge to put on the

I just took the ice bin out of my fridge to put on the heated bed. I am awaiting the 0C reading. It thinks the room is 75C. I need to wait for my thermometer to get back up to room temperature before I can see what temperature it is and I suspect that thermometer may be off too.

What tips do you guys have for calibrating thermistors and figuring out which one you have when you forget?

It sure seems to take long enough for the ice bin which covers almost the entire bed to drop the bed temperature 24C. It makes me wonder if the bed heated too well and the thermistor got unseated or if I am using the wrong thermistor table. I removed the glass. Ah. A single ice cube in a plastic bag pressed onto the center of the bed above where the thermistor is drops it faster. I must not have had direct contact between the ice bin and the spot above the thermistor. I finally got 0C for a moment and then it went back up to 1C.

I suppose I could borrow an oven thermometer from the neighbor and try to measure the high range.

I have only tested for the bed so far.

I use an infrared thermometer.
It also helps to check the temperature during heating. The thermistor measures the heatbed, but the top of the glass heats with a delay.
Depending on the room temperature, the heating power of the bed and thickness of the glass it can trail even more than 10 degrees.
I have lost a few prints because the top of the bed wasn’t at the right temperature yet when I started to print.

Maybe drip some water on the bed, put the thermistor tip in that, then put a baggie full of ice water on top of that. That way you get really good thermal contact.

imho, buy known-value thermistors and replace them if you’re getting weird readings like that. No need to play around with all this crap. It’s a big waste of time and can potentially be dangerous. Thermistors are like what…$4/ea? – is it really worth all that time and effort to not have to get a handful of known-good thermistors?

@ThantiK If you could just walk down to the store and get one, I might just do it. If I spend time figuring it out…well, I have not been doing much with my time anyways. It seems there are over 6 tables for 100K thermistors.

@NathanielStenzel Amazon prime, takes like 24 hrs to get a thermistor.

@NathanielStenzel as long as you have the right basic type (eg 100k NTC) and the expected thermistor reading circuit (eg voltage divider) for the MCU ADC, the exact table doesn’t matter all that much. The difference between the thermistor and build plate surface differs by an amount similar to the calibration error you get from the wrong thermistor table. It should read more or less correct at room temp, and you should then empirically figure out the right temp setting to get your desired bed surface temp for ANY printer whether you know the right table or not.

Are you sure it isn’t reading in F? :wink:

@CELSS quite sure. I do not keep my living room below freezing temperatures.

@Ryan_Carlyle if the resistance to temperature graph is curved, I would think the units of heat needed in the PID math may be off. The PID tuning may still compensate for that though. I have never seemed to print at the recommended temperature anyways. Thanks.

@NathanielStenzel PID doesn’t know what heat is :slight_smile: it’s just a feedback loop. It can be tuned to work as long as heater power makes measured temp rise within a certain amount of time. Sensing lag can make PID unstable but otherwise it’s pretty crazy robust.

PID doesn’t do well with nonlinear systems, but as long as it’s roughly linear in the setpoint area, it does a fantastic job.

@John_Bump yeah, all these printer heater PID loops are saturated (100% power) and often outright bypassing PID entirely until they’re close to the setpoint. The thermistor sensing circuits are typically set up to be reasonably linear near the setpoint range. Does create a lot of low-accuracy reads at room temp and below though…