PTC resistance output from 0-10V analog input ??

[recharg]

I would really appreciate help with the following application:

I want to use a PID controller for controlling a process fluid few degrees below ambient dew point. Process fluid temperature is measured by a standard 1097 ohm PTC temperature sensor A and Dew point is measured with a sensor B which provides a 0-5V or 0-10V voltage output over 40-90 deg F range. The Honeywell T775U controller fits the bill perfectly except for a very frustrating road block - controlled temperature input accepts PTC thermistor or 0-10V BUT "reset temperature" only accepts PTC sensor input 

So now I need to convert the 0-10V (or 0-5V if easier) output of the dew point sensor to a PTC resistance input:

degF  Volt   ohm
40    0   1155.4
45    1   1174.7
50    2   1194.0
55    3   1213.2
60    4   1232.4
65    5   1251.6
70    6   1270.8
75    7   1289.9
80    8   1309.0
85    9   1328.0
90   10    1347.1

Searching for the infamous "voltage to resistor" holy grail led me to this thread and made it clear there is no simple plug-and-play solution.

What would you suggest as the best way to solve this signal conversion challenge?

[Zero999]

What voltage does it place across the PTC resistor input?

If it's below 5V, then how about a digital potentiometer? You'd need an ADC to read the voltage source, convert it to digital and set the digital potentiometer to the appropriate value. The difference between the highest and lowest values is 191.7Ohms, so it would be ideal if you could use that value digital potentiometer, in series with a 1k resistor, but unfortunately, the lowest value available is 1k. For maximum resolution, the potentiometer could be in parallel with a fixed 240R resistor, in series with 1k1, to get a 1k1 to 1k394 range.

To speed calculations up, a look-up table could be used for converting the voltage to the appropriate digital resistance value.

[recharg]

Thanks, I don't have the T775U yet but it is very likely using more than 5V across the PTC resistor input - the controller can be powered by 24V or 120V AC and can handle long temp sensor leads so it would use a higher voltage across resistor inputs. If on the other hand it is somehow below 5V, my knowledge of digital circuits is even lower than analog circuits. So would need more help for sure to get to the full solution you are suggesting.

[Ian.M]

If you use a quad 1K 8 bit digipot with all four resistance elements in parallel, you can get down to 250R with enhanced resolution.  Use with a 1K1 series resistor as Zero999 described.

However if you can determine how the T775U senses the PTC resistance it may be possible to fool it with a fixed resistor and either a controlled shunt current source or series voltage source.

[Zero999]

Thinking about this again, the current through the resistor and therefore the voltage should be kept as low as possible, to prevent the power dissipated in the resistor heating it up and interfering with the temperature measurement. In any case, once the digital potentiometer is in series with 1k1, it won't see the full voltage, so it won't be an issue.

If it measures the resistance using a constant current source, then the voltage will be proportional to the resistance, so simply connecting it to a voltage source of the correct value should be able to fool it.

If it works by using a voltage source, then it can be fooled with a current source.

If it works using a potential divider, then it's more tricky. Yes four 1k digital potentiometers can be connect in parallel to make 250R.

I've had another idea of using an op-amp to effectively scale the value of the digital potentiometer down to a lower value. 1024 position 20k, 1% tolerance, potentiometers seem to be widely available in Digikey. Although it's more components, it might be easier, as there's only one digital potentiometer to worry about, which will also carry a fraction of the current. It will probably require a frequency compensation network to avoid oscillation. I haven't done any dynamic analysis on, whichwill depend on the op-amp and MOSFET used.

The horizontal axis is the digital pot's resistance and the vertical R_OUT, which is the output. It's measured using a constant current source and applying Oham's law. I1 represents the current source in the device, but it could be a voltage source and it will still work.

[recharg]

Thank you very much both.

I am evaluating all of your suggested options along with a ready-made analog-to-resistance transducer I found at Kele.com

[mikerj]

If it measures the resistance using a constant current source, then the voltage will be proportional to the resistance, so simply connecting it to a voltage source of the correct value should be able to fool it.

If it works by using a voltage source, then it can be fooled with a current source.

If it works using a potential divider, then it's more tricky. Yes four 1k digital potentiometers can be connect in parallel to make 250R.

If the PTC is part of potential divider then it can still be emulated with a simple voltage drive provided it's ground referenced, or a diff amp if not.

[Zero999]

If it measures the resistance using a constant current source, then the voltage will be proportional to the resistance, so simply connecting it to a voltage source of the correct value should be able to fool it.

If it works by using a voltage source, then it can be fooled with a current source.

If it works using a potential divider, then it's more tricky. Yes four 1k digital potentiometers can be connect in parallel to make 250R.

If the PTC is part of potential divider then it can still be emulated with a simple voltage drive provided it's ground referenced, or a diff amp if not.

A potential divider won't be linear which will make it much harder. The value of the other resistor in the divider and open circuit voltage will need to be known in order to decide what voltage to inject, to fool it into seeing a certain resistor value. The voltage or current curve vs resistance will not be a linear slope, as would be the case for a constant voltage or current source. It's doable, but would require more than just an op-amp.

Thank you very much both.

I am evaluating all of your suggested options along with a ready-made analog-to-resistance transducer I found at Kele.com

That module has a discrete digital potentiometer on it. A micro-controller will switch different resistors in/out of the circuit to get the desired resistance. If you just want something which works, I'd recommend it.

[recharg]

Thank you very much both.

I am evaluating all of your suggested options along with a ready-made analog-to-resistance transducer I found at Kele.com

That module has a discrete digital potentiometer on it. A micro-controller will switch different resistors in/out of the circuit to get the desired resistance. If you just want something which works, I'd recommend it.

Thanks. I do need to implement a solution quickly, the 255 step resolution of the resistor ladder would be sufficient for now.