Thermistor issue - Page 2

Equipment doesn't work? Troubleshooting? If you're handy, members can help.
Supporter ♡

#11: Post by Pressino »

I just found this comment by Eric from a few years ago, which discusses the ECM/Profitec sensors:

#3: Post by erics » October 7th, 2017, 1:48 am

The Profitec 700 utilizes a PT100 RTD which is quite a bit different than a thermistor. The Profitec RTD has 1/4"-19 BSPP threading with an OD of 13.16 mm.

The PT100 has a resistance of 100 ohms at 0 C and ~109 ohms at 23 C. I BELIEVE the RTD is sealed to the boiler with a teflon sealing washer although it is not shown that way in the parts book. So, if you tried to use a sensor with 1/4" NPT threading, that would have been a mistake.

Just as a FYI, most(?) other PID machines utilize a 50K ohm NTC thermistor.

So it does look like you were right about a PT RTD being used, even if the one on that gicar site isn't it.
★ Helpful

ToM4 (original poster)

#12: Post by ToM4 (original poster) »

So it may be a K type?
I made an experiment using my oven Ikea tempertaure sensor. it has a value of about 4400 ohms at 25°c but the PID reads -15°c :D
but I don't know neither what type it is.


#13: Post by jpender »

RTDs are nearly linear in resistance with respect to temperature. 100 ohms at 0°C for a PT100 and 1000 ohms for a PT1000. At -15°C you would see 94 (or 940) ohms and at 85°C you'd get 133 (or 1330) ohms. That's not what you're seeing so your circuit isn't expecting an RTD.

A thermocouple is just two wires of different metals welded together. The resistance is very low, less than 1 ohm. Because the metals are different a very small voltage is created that varies with temperature. If you have a circuit that expects a thermocouple and you plug a resistor (or thermistor) into it there would be no voltage so it would simply report the reference temperature, room temperature probably. But it might depend on the particular circuit.

An NTC thermistor has a non-linear resistance to temperature curve, higher at lower temperatures. E.G. 10K at 25°C and 1K at 95°C. There are different curves too. You can't just choose a thermistor by a single temperature/resistance (e.g. 3.3K at 25°C). You have to know the curve too. If your circuit expects a thermistor you could insert several resistors of known value and see what the readout temperature is for each. From that you could determine the characteristic curve and know what part to buy. You IKEA oven thermometer is probably a thermistor. They are cheap and require simple circuits and some can handle 300°C.

But you are saying the PID circuit reads -15°C with a 4.4K resistor and 85°C with a 3.3K resistor. That's not typical for a thermistor. The fun of DIY.

Did you keep the original part? What did it look like? And how do you know that it was broken versus something being wrong with your circuit?

ToM4 (original poster)

#14: Post by ToM4 (original poster) »

Hello Nick

That's the post I was looking for! thanks!

John. I'm going to check the resistance with the kitchen sensor. I may have made a mistake reading the multimeter.
Just some more info. My PID has no control box. It's plugged to the 220v. The output to the relay is 6 VDC.
The schematic is the same as on the Puristila or almost. There is a bi metalic thermostat on the fase + on neutral for safety puprose in case of failure of some ohter safety parts.
As for the original sensor, it was dead. There was the A2 message on the PID. I thought it was an interesting experiment to change only the component. As it was inside a kind of resin, I had no other choice but to drill it so I turned it into dust... :o
I find that weird that these components seem to be kept as classified.
I'm at work, I will post the diagram.

thank you all.

EDIT : the kitchen sensor is about 48 000 ohms at 23°c not 4400 which make more sens


#15: Post by jpender »

ToM4 wrote:EDIT : the kitchen sensor is about 48 000 ohms at 23°c not 4400 which make more sens
That's in line with what you'd expect for an NTC thermistor.

Take a look here.

If you have 48000Ω at 23°C and 3300Ω at 85°C that's a ratio of 0.069. If that first number were for 25°C you could simply look at the Rt/R25 value for 85°C and find a match. Curve 12 is pretty close.

You should probably measure several resistors with your DMM and then see what they read with your PID. Individual thermistor families will have data sheets that specify the resistance for each temperature so you should be able to find a match that way. Or at least a pretty good one. It doesn't have to be perfect. Thermistors have some variance. And they drift with time.

ToM4 (original poster)

#16: Post by ToM4 (original poster) »

Thanks a lot John. I'm going to test a PT100. If ever I can't find the good one, I'll order a complete sensor. 40 euros for a component + a sleeve...
That's because I built my own machine so I want to keep on having a hard time :D
850 ml Single brass boiler, no steam, rotaty pump, PID, E61 grouphead, Small form factor I'll soon show it :) everything work except this bloody sensor


#17: Post by jpender »

ToM4 wrote:Thanks a lot John. I'm going to test a PT100. If ever I can't find the good one, I'll order a complete sensor. 40 euros for a component + a sleeve...
If I'm right a PT100 won't work unless you are also redesigning the PID circuit.

PT100 and PT1000:


They do kind of get you with sensor probes. I always just buy the thermistors/RTDs/thermocouples and either use them naked, if appropriate, or enclose them myself. Precision thermistors are just a few dollars.

ToM4 (original poster)

#18: Post by ToM4 (original poster) »


I eventually found a reliable source. I'm sure that on the profitec 700 it's a PT100. As it has the same PID 3K3 writtig on the enclosure of the PID. the Espresso planet site says the P6037 fits both profitec and ECM... I'm going to give it a try.


#19: Post by jpender »

If it's really a PT100 then I think you either made some error measuring the resistances or else the controller is malfunctioning. I'm going to put my money on it being a 50K thermistor. But hopefully I'm wrong and it works out. I'm wrong all the time. :-)

User avatar
Team HB

#20: Post by Jeff »

"Back in the day" there were decade resistance boxes. You could dial in any resistance in its range. For testing like this, one of those or a handful of resistors covering the potential range of interest would be helpful. That way you could determine the resistance that is somewhere close to 95°C indicated as well as how steep the curve is by finding something near 90°.