your inquiry: "Are you aware of a device or method of converting the resistance output from a Pt1000 sensor to the equivalent resistance output of an NTC 20 kΩ sensor? Background. The existing solar thermal system in my house has a Pt1000 sensor in a solar thermal collection on the roof. I am considering replacing the solar controller with a Viessmann model which uses an NTC 20 kΩ sensor. The Pt1000 sensor on the roof is difficult to access so I am investigating if it will be possible to reuse it with something to convert its resistance to the NTC 20 kΩ equivalent. My background is in systems software so this is well outside my comfort zone!

Vishay's answer: if those circuits are out of your comfort zone, I suggest you do not start to modify them and take contact with the suppliers of those systems and ask them for advice. Converting RTD sensors (Pt1000) signals to be read as an NTC is not advisable.

There are small variations between suppliers and control manufacturers - most of the NTC values were created from 3rd party suppliers like Produal... the differences in ohms and even so in 'C are small.

Ohmonwifi allows you to set your own table between 'C and ohm. There is an excel spreadsheet on ohmonwifi https://www.ohmigo.io/_files/ugd/0ea245_c1686f604dfc46d3b7d6c65a8b9ad19a.xlsx?dn=Ohmonwifi%20create%20user%20defined%20Temp%20vs%20Resistance.xlsx

The Pt1000 Resistance Table is straightforward. NTC Thermistors, Resistance/Temperature Conversion is more complex. I know it should measure 20 kΩ at 25°C but I'm not sure which of the 17 curves to apply?

Using the curve in the Viessmann guide you can see that the resistance at 140°C is about 0.5. So Rt/R25 is 0.025. You can then look at your 17 curves and see what the best match is. You could just estimate an approximate equation to match the Viessmann curve instead of trying to emulate an actual device. Or ask Viessmann what device it is?

Good suggestion. I am in the process of doing something similar to see which of the 15 (1-14, 17) Vishay curves is a good fit (It will also be interesting to see how the Ohmigo NTC 20 kΩ table compares). A table from Viessmann would have been more convenient for the comparison than the log/linear graph. However, solar thermal doesn't need as high precision as some other processes.

There are also some Vishay Tools that might be useful.

What is the problem with the old controller?

It works fine but the integration with the rest of the system is rather clumsy. For solar back off, a relay in the SKSC2 drives a contactor which adds a resistor that makes the the thermal store appear to the system controller to be hotter than it really is. Also the RESOL DL2 data logger attached to the SKSC2 has died. Both the SKSC2 and RESOL DL2 were installed around 15 years ago.

My current thoughts are to replace my existing Viessmann Vitodens 200-W (type WB2C) with the latest version (type B2HF). The existing uses KM Bus and the latest version uses PlusBus so the wall mounted mixer extension will need to change. It seems like a good time to also change the solar controller to achieve better integration. I'm considering a Vitotrol 300-E as well. Might also replace the solar pump (how long do they last?). The upgrade should be straightforward, apart from the solar collector sensor ...

Using the curve in the Viessmann guide you can see that the resistance at 140°C is about 0.5. So Rt/R25 is 0.025. You can then look at your 17 curves and see what the best match is. You could just estimate an approximate equation to match the Viessmann curve instead of trying to emulate an actual device. Or ask Viessmann what device it is?

What is the problem with the old controller?

Hi John.

It would appear that there are several possible solutions. You could go down the digital route using something like a digital rheostat, though I suspect that will entail having a computer running continuously to perform the required calculations.

There is the possibility of an analogue solution, using electronic components to achieve the same effect.

Do you have a preference?

Requirements are for reliability and acceptability to heating installers / maintainers. Ideally, a ready made device with safety approvals (may not exist). Unfortunately, the acceptable solution may be to change the sensor on the roof ...

As an aside, I found the Ohmigo® TTL which is a precision digital controllable resistor capable of generating a resistance ranging from 68.5 ohm up to 900,000,000 ohm with superior accuracy, overall better than 1%.  That should be good enough for simulating an NTC 20 kΩ (and a wide range of other sensors). Only does part of the job but demonstrates that part is achievable. Originally found the Ohm On WiFi version.

Hi John.

It would appear that there are several possible solutions. You could go down the digital route using something like a digital rheostat, though I suspect that will entail having a computer running continuously to perform the required calculations.

There is the possibility of an analogue solution, using electronic components to achieve the same effect.

Do you have a preference?

The software part is simple, but how do you output actual resistance?

My guess would be a Digital potentiometer. Might need a logarithmic one to cover the NTC 20 kΩ resistance range? I'm not an electrical engineer so this might be nonsense.

I assume that both the old and new controllers are passing a small current through the sensor and measuring the volt drop, and hence calculating the temperature. Can you measure this volt drop at various temperatures?

Sounds reasonable. I have a fairly limited electrical toolkit, not even a digital multimeter.

What make and model are the old and new controllers?

Old: Sonnenkraft SKSC2
New: Viessmann Solar EM-S1 extension (proposed).

Over what temperature range do the sensors operate and what control function do they perform.

There is a resistance against temperature graph on page 16 of the Viessmann Installation and service instructions manual (linked above) which has a range from 0 °C to 140 °C which seems reasonable for solar collectors. Curve A is for the NTC 20 kΩ sensor.

The test procedure is: Disconnect the plug of the relevant temperature sensor from the extension. Check the sensor resistance. I assume a multimeter would be used to check the resistance (by passing a small current through the sensor).

The software part is simple, but how do you output actual resistance?

I assume that both the old and new controllers are passing a small current through the sensor and measuring the volt drop, and hence calculating the temperature. Can you measure this volt drop at various temperatures?

What make and model are the old and new controllers?

Over what temperature range do the sensors operate and what control function do they perform.