Need advice on Thermistors

Hi everybody! My extruder got plugged up, and in the process of taking it apart I destroyed the thermistor. In the Merlin Configuration.h file it was a Type-5 ATC Semitec 104GT-2. At the Semitec web site it has the parameters for their thermistors specified

[www.atcsemitec.co.uk]


The 104GT-2 has its parameters listed as:

R25 100.0k Ohm±3%
B Value 4267K±2%
Operating temp (°C) -50~300

I can't find that for sale in small quantities anywhere (let alone at a reasonable price). So I ordered this part on eBay:

[www.ebay.com]

It has these parameters specified in the auction listing:

Specifications:
Item:NTC Thermistor
Model:B2-100-3950-1
R @ 25: K 100K-1%
B25/50:K 3950-1%
Element:Thermistor,type MJB2

Looking at the thermistortables.h in the Marlin code base, there is an entry for a Type-71 thermistor of Honeywell 135-104LAF-J01. It's
parameters are:

R0 = 100K
Beta = 3974

Question #1: Do I dare just use the newly ordered thermistor from eBay and call it a Type-71 thermistor in Marlin's Config.h file?

Question #2: Is there an easy and inexpensive place to get a legitimate Semitec 104GT-2 Thermistor?

Question #3: Is there any other good advice you can give me?


THANKS!

#1: You can use it, but the B=3950 vs B=3974 would read about 2C low around 200C.

#2: Your new thermistor is probably a good replacement, since the 1% tolerance is better than the 3% tolerance

#3: Use/Make a table that matches your actual thermistor. [www.geeetech.com] has an old sample table and a datasheet. Or you could read a couple of those values off of the datasheet and plug them into createTemperatureLookupMarlin.py Or, if you can measure temperatures well, you could measure three actual resistances and temperatures and plug them into the script and get a well-calibrated table.

Hey, the given sample table appears to match the Marlin table 60.

Quote
DaveX
#1: You can use it, but the B=3950 vs B=3974 would read about 2C low around 200C.

Thanks for the concise answer!!! At worst case, I can live with being off 2 degrees C at 200 C. I was worried about melting down my extruder.

Quote
DaveX
#2: Your new thermistor is probably a good replacement, since the 1% tolerance is better than the 3% tolerance

OK! It is good to hear your opinion on that! That 1% versus 3% tolerance is on a per part basis, right? What ever the part is, it will consistantly read the same, right? That 1% or 3% deviation is between similar parts at identical conditions, right? Or stated differently... an individual part won't read randomly +/- 1% or 3% ???

Quote
DaveX
#3: Use/Make a table that matches your actual thermistor. [www.geeetech.com] has an old sample table and a datasheet. Or you could read a couple of those values off of the datasheet and plug them into createTemperatureLookupMarlin.py Or, if you can measure temperatures well, you could measure three actual resistances and temperatures and plug them into the script and get a well-calibrated table.

Wow! Thanks for the suggestion! I didn't realize createTemperatureLookupMarlin.py was there and I didn't realize that was its purpose. I guess as a fall back position I will learn about that. But instead... I'm thinking I will do what you pointed out next:

Quote
DaveX
Hey, the given sample table appears to match the Marlin table 60.

I just went through this table and see what you are pointing out! I think I have a good handle on this now thanks to your help! I guess I will declare my new thermistor to be of type 60 and I will use a digital infrared thermometer to see that it is reading close to what is expected.

THANKS SO MUCH FOR THE HELP!

You can get the proper thermistor here:

[www.rapidonline.com]

I bought half a dozen as spares. Quick delivery.

Yeah, the 1% or 3% are fairly systematic. I think it could affect the beta or the R_25C value, but both should be constant for the same individual part. You could test this yourself by putting the thermistor or hotend in a icewater bath or boiling water bath and measure its resistance at temperatures to see, but I imagine you'd have more variance in keeping your baths consistently stirred versus the repeatability of the thermistor.

An infrared thermometer might not be able to read accurately enough. I've got a handheld one that that measures average temp over a 10 degree angle, and it would just pick up portions of the outer surface, not the internal working temperature.

[hydraraptor.blogspot.com] is the main rep-rap reference. There Nophead says he uses a thermocouple slipped down into the barrel to get the internal temperature in the melt chamber. I calibrated mine with one of these [www.amazon.com] hooked up to a $25 PID controller.

With the script, you can calibrate for your measured pull-up resistor and measured temperatures and resistances to get temperatures as accurate as your measurements. With calibration, you could use a cheaper 3% or 5% thermistor and get higher accuracy.

Edited 1 time(s). Last edit at 03/27/2014 12:32PM by DaveX.

Quote
dave3d
You can get the proper thermistor here:

[www.rapidonline.com]

I bought half a dozen as spares. Quick delivery.

I need to do the same thing and order a few spares. Unfortunately, this place charges a lot for shipping to the US (where I am).