Calibrating heated bed and hot end temperature

I suppose to use infrared digital temperature meter for calibration. Any better suggestions? Thank you!

An non contact thermometer is a dangerous thing to use for calibrating hot ends and heated beds. The are calibrated based on how the target emits IR radiation. Things that are black at IR emit it very well. Things that are shiny at IR emit it very poorly. The difference is big enough that you get major errors. To get good data you need to paint your target black with some "known good at IR" paint. That's what we do at work when we do thermal scans.

Your thermistor is probably good to a couple of degrees at the typical heated bed temperatures. You might add a few degrees for hot materials in the hot end. Unless your calibration reference is good to a degree or two, it's not worth the trouble.

If anything it's much easier to calibrate your thermistor in boiling water. Before you mount it on the hot end (or heated bed), wire it up to your Ramps board. Boil some water and dunk the thermistor in it. Your software should read 100C. There are some corrections if you have very hard water or if you live at altitude. You can look them up if that applies to you. About the only tricks are: keeping the leads dry (plastic bag), keeping them from shorting (teflon tube), and getting the bag to sink (rock in bag).

Thank You once again, uncle_bob!

Just some black tape can be good for thermal imaging.

Yes, different colors emit IR differently.

Quote
jeffegg2
Yes, different colors emit IR differently.

That's what I'm afraid. These infrared digital temperature meters have manually adjustable emissivity settings depending on surface reflectiveness, so the result may be not so accurate.

It's amazing how much variation there is in materials at IR. Some things that are clear at visible light are black at IR. Some things that are black at visible are transparent at IR. Picking the right "stuff" for IR mapping can a bit involved. The easy way is to take something like your heated bed and put aluminum foil on it. It's shiny at IR. Coat it with your magic "stuff". When the IR thermometer agrees with your bed temperature the stuff is good. That's not much use for calibrating your heated bed ....

I've used the crude method of backing out the 3mm filament from my extruder and jamming a thermocouple wire attached to my multimeter all the way down into the nozzle (so I'm measuring from the inside). I get a bit of melted ABS on the thermocouple but the insulation is made to withstand the temperatures I'm using for ABS (230C) and this gives me good confirmation of my thermistor settings.

Guys,

I did some tests this morning using a thermocouple taped to the top of my glass which has kaptop tape on it, which rests on a brand new 12/24V PCB heat bed.

At room temperature, Pronterface was showing 21.2 deg C and the thermocouple thermometer was reading 22.4
I set the bed temp to 60 deg, left it 5 mins and Pronterface was showing 60.1 deg C and the thermocouple thermometer was reading 55.7 deg C
I set the bed temp to 115 deg, left it 45 mins and Pronterface was showing 115.5 deg C and the thermocouple thermometer was reading 106 deg C

So I suspected that Pronterface (the thermistor) needed to be calibrated.

However I tried using my IR thermometer gun, and it read 115 deg (Yes I know these they are inaccurate).

So I put some towels on top of the bed (on top of the thermocouple), and left it another 15 mins

The thermocouple now read 116.4 deg C and Pronterface was reporting 115.5 deg C

I tried to stick the thermocouple to the bed more securely with some more kaptop, but it didn't make much difference. The tiny bead of the thermocouple was definitely in contact with the kaptop on top of the glass.

So I presume that 1 of 2 things is going on.

1) The thermocouple is not able to get a true reading as its loosing too much heat, because its only in point contact with the glass and its a roughly spherical shape, so 99% of its surface area is not in contact with the glass
2) There is a large thermal gradient at the top of the glass, so that the surface of the glass, (well the Kapton) is actually at around 10 degrees below the temperature of the core of the glass.


I'm not too sure if this matters that much, or whether the optimum 115 deg initial temperature for ABS includes a compensation for the thermal gradient etc. And of course as soon as the ABS or PLA comes into contact with the surface of the glass / Kapton, its going to insulate the glass / Kapton from heat losses.

But I guess the learning from this, is to insulate above the termocouple when taking readings of the bed temperature, otherwise they could be wildly off ;-)

BTW. I've not double checked the calibration of the thermocouple thermometer. I'll need to put it in some boiling water, but as all 3 temperature sensors appear to give virtually identical results at 115 deg, I suspect that its probably OK.

Edited 1 time(s). Last edit at 11/06/2013 05:08PM by rogerclark.

There is an air gap between the bed heater and the glass. You will loose quite a bit of heat at that junction. You can fill it with heat sink compound and reduce the loss. (don't do this - it makes a monster mess). The IR thermometer will mostly look through the glass at the solder mask on the heated bed pc board. IF the solder mask is black at IR, it will read that temperature. IF it's clear(isn) at IR you will hit the copper and who knows what you will get.

Quoted bed temperatures are all based on the same thing - what the thermistor on the back side of the heated bed reads. Everybody looses heat through the glass. If you "calibrate" to the glass surface, and then use their temperature settings, your bed will be much colder than their bed.

Once you do some calibration runs, you may find that in your setting and with your filament, at your speeds, with your degree of tolerance for errors, printing your sort of stuff - you want a bed temperature that's 10 or 20C off from what everybody else uses. The same thing could be true (probably to a lesser degree) with hot end temperatures.

None of that says you should not calibrate things like the sensors. It's good to know what's going on. It's also useful to go back to when something changes.

Hi Uncle_Bob

I had considered using heat sink compound, but as you say, it would make a huge mess, and taking the glass off would be difficult and messy.

My Glass is resting directly on the PCB, so although there will be an air gap, I'd have thought it would be relatively small, and in some places virtually non existant

There is a lot of data on electrical heat sinks and small air gaps. You can machine both surfaces dead flat and clamp them together. They still benefit from a bit of heat sink compound to fill the last tiny bit of gap. With a non-flat pc board (the never are truly flat) there's bound to be a gap. With a 50:1 ratio in thermal conductivity (glass to air) or 2000:1 (pcb copper to air) it doesn't take much air to decouple things. Your thermistor is sitting right on the copper heating element. It's well coupled....

One less nutty gap filler - cooking oil. At PLA temperatures water might work.

The thermistor for my heated bed is on the bottom and I find that there is usually a drop of a few degrees C going through the board to the top of the glass. I also find that there is a significant temperature variance (using a thermocouple kap-taped to the top of the glass) between the middle and edges of the heated bed. I have some ideas about making it more uniform but haven't had a chance to play with that yet...

The idea about using heatsink compound between the top of the PCB and glass is interesting. The stuff is a bit expensive but it's worth a try - I never remove the glass, even when I'm switching to milling (I just put something else on top of the glass). I think that reducing airflow around the edges of the heated bed PCB is helpful also.

Not moving the bed would help equalize the temperatures as well ....

The real question is - does this matter? Your heated bed (unlike your hot end) has a pretty wide range of acceptable temperatures. As long as it's hot enough that the material does not shrink to fast / to much you are fine. There is no advantage to a +/- 0.01 C heated bed if a +/- 20C does everything you need to do.