I need to simulate various sensors so I can automate the testing of automotive ECUs.
Some just provide a 0-5v analog out which I can simulate no problem.
Some sensors are a resistive type sensor mainly thermistors connected to ground like in the image:
I need to replace the sensor with an Arduino so at the moment I'm thinking a digital pot with the wiper connected to the ecu input, connect the lower side of the pot to ground and leave the top open.
I know with a physical pot this would work but would it work with a digital one?
have never used one but yes it will work as a matter of fact there are tutorials on this web site for ad5171 digital pot and arduino or use a servo motor turning a pot
not sure why you think an arduino is your answer.....Automotive ECU's look for a value within a range for each sensor input...look up sensor range spec for each sensor and you will find the acceptable range/values....then use a resistor in that circuit to simulate a value within that range. They are all 5v to ground..accept the A/C generator sensor/pickup's(crank and cam sensors)...(and O2's)..they produce a/c volts.
The reason I plan to use an Arduino is because modern ECUs react different in different conditions. For example ECUs won't go into static DPF regen if the water & exhaust temps are above a certain amount and the ECU will know if they aren't rising as they should and shut the process down. It isn't just testing I am interested in but reversing engineering among satisfying my curiosity.
My solution has to be purely circuit based i.e no servos turing a pot etc.
My only problem is I'm not sure if a digital pot will work the same as a mechanical one with no positive connected to one side of it.
I know the obvious solution would be to test it but I'm waiting for some to arrive and was hoping for a heads up before they do.
Hi
My solution has to be purely circuit based i.e no servos turing a pot etc.
Any reason why, some of those resistance values I gather are quite low, and as you say digital resistors may not like the bias the ECU circuit applies.
Tom... 
If the current needed is too high or resistence (range) expected is too high or low you can use DAC or filtered PWM as input: according to your scheme the ECU measures voltage anyway.
not sure what " DPF regen"...means....but you can adjust the output of any sensor but simply adding a value to the "digitalRead" output and increment it over time..to sim the warm up....or if your not actually reading a value....just increment from the set value for each sensor input..... and set a know value of time frame for the next output with the variation added.
I was just playing around with a thermistor setup and resister values...and the output was off...tried diff resisters to adjust the out put....but the simple solution was to just add (or subtract) a value to the output...ie: output was 65 deg but it really was 75 so...print (or use value) + 10. You can do the same with your simulated outputs by adding a time variable plus and increase or decrease +/- value before the output is used in the next equation for that sensor
If you want to simulate an engine, you have to change the simulated sensor values over time, as they would change during operation of the real engine. It may be a good idea to record the sensor values of a real engine, for replay on the simulated engine, and then modify the values as you like.
I don't see a reason against using a pot and servo for the simulation of resistive sensors, do you?
I've worked with some resistance sensors in the Rotax area of engines for light aircraft... as I remember, the datasheets generally show linear coefficients (slope) for thermistors, oil pressure senders, fuel level senders, etc.
If one can find the specific data-sheets, then any physical engine conditions should simply be a ratio applied against a stored array of either high/low values (or the linear high or linear low.)
Ray
Maybe you want some isolation between the ECU and the Arduino. A motor and a pot seems primitive but at least it won't toast an ECU.