revilla Posted February 22, 2020 Share Posted February 22, 2020 Whilst working on the MEMS3 mapping, I've identified a number of tables which contain the calibrations for temperature sensors. For example here is the one that I've confirmed by experiment is the calibration for the coolant temperature sensor: If you change the temperature numbers, for example increase them by 10 degrees, then the engine's IAT readout on OBDII scanner increases by 10 degrees as expected. So the table is mapping the reading of the IAT sensor onto an actual temperature reading for the ECU.The question I have is - what exactly is on the X axis?I know the Z axis is correct for temperature as all of the temperature scales in the ECU are in tenths of a degree Kelvin and are very recognisable as numbers like 2732, 2932, 3132, corresponding to 0C, 10C, 20C etc.The X axis however doesn't seem to correspond to anything that I can think of, for example it doesn't vary linearly with either the voltage measured at the sensor or the sensor resistance. It seems to be roughly proportional to the square root of the resistance but that doesn't seem to make any sense as a choice by the designer. All of the other scales that I've identified in the ECU have been sensible and identifiable (e.g. temperatures in Kelvin, engine speed in RPM, MAP in kPa*100, fuel enrichment in %).You can tweak it without knowing what X is just by using the indicated temperature and working backwards, and to be honest I doubt anyone would really have cause to tweak it, it's mostly just for academic interest as I want to understand as much as I can about what's going on.Any ideas? Link to comment Share on other sites More sharing options...
rj Posted February 22, 2020 Share Posted February 22, 2020 I suppose the input of the ECU is a pull up resistor and the sensor is an NTC resistor. Could it be something that relates to the input voltage on the ECU pin generated from this? Link to comment Share on other sites More sharing options...
revilla Posted February 22, 2020 Author Share Posted February 22, 2020 I thought about that, but it didn't seem to match against the numbers I had. Mind you, the voltage measurements I had weren't my own, they came from somebody else I was helping out ... they may differ to what I might measure on my own engine. I guess I could play in Excel to see if some combination of series resistance and supply voltage matches those numbers reasonably well. I've got resistance measurements off my own sensor. Link to comment Share on other sites More sharing options...
revilla Posted February 22, 2020 Author Share Posted February 22, 2020 Regin, I think you're right! I think the voltage numbers I've got must be wrong. I'll need to measure them tomorrow properly.But this is what I've worked out:As it sits in my garage cold now, the sensor has a resistance of 4085 ohms and when connected to the ECU, the voltage across it is 3.482V. Assuming that the internal supply is either battery voltage or (more likely) a regulated 5V supply, that leaves the following likely solutions which would be consistent with the measurements stated above:12.6V in series with 10700 ohms.5V in series with 1780 ohms.If I model the second solution over the range of temperatures for which I have approximate resistance data, as shown in the table below, if I take the calculated voltage assuming 5V in series with 1780 ohms inside the ECU, and multiply by 200 I get something looks which very, very close to X indeed (my resistance and temperature measurements weren't that accurate). MeasuredCalculated Multiplied TempResistanceVoltage by 200X703900.898617512 180182605501.180257511 236241508001.550387597 3103154012002.013422819 4034063018802.568306011 5145111249003.667664671 734728So X looks to be voltage across the sensor in units of 1/200th of a volt, or 5mV.I'll try to profile the voltage against temperature for myself tomorrow to confirm it.Thank you! Link to comment Share on other sites More sharing options...
revilla Posted February 22, 2020 Author Share Posted February 22, 2020 The more I think about it ... a unit of 5mV or possibly 4.88mV would make a lot of sense with a regulated 5V feed ... if you digitise the range 0V to 5V to 10 bit precision you get a natural unit of 5V/1024. Link to comment Share on other sites More sharing options...
revilla Posted February 22, 2020 Author Share Posted February 22, 2020 And since the numbers on the X scale go to 1011, I think it must be 4.88mV not 5mV. If it was 1000ths of 5V I wouldn't expect to see numbers above 1000, but if it is 1024ths of 5V that looks sensible. Link to comment Share on other sites More sharing options...
revilla Posted February 23, 2020 Author Share Posted February 23, 2020 OK so I just warmed the car up and took my own set of voltage readings vs. ECU-reported temperatures.I think that pretty much confirms it. X is the output of a 10-bit analogue-to-digital converter in the ECU, reading sensor voltage, when the sensor is supplied with 5.0V through approximately 1780 ohms. Link to comment Share on other sites More sharing options...
Member Jonathan Kay Posted February 23, 2020 Member Share Posted February 23, 2020 Do you still get the sigmoid transfer function with that explanation?Jonathan Link to comment Share on other sites More sharing options...
bjw Posted February 23, 2020 Share Posted February 23, 2020 Do you still get the sigmoid transfer function with that explanation?Yes, because of the sender characteristics.Temp/resistance data for typical sender here:-https://www.lotus7.club/sites/default/files/images/users/42534/Temperature_Sensor_NTC_M12_Datasheet_51_en_2782569739pdf.pdf- Using Andrew's figure of a 1780 ohm source resistor results in the following relationship:-Cheers Link to comment Share on other sites More sharing options...
revilla Posted March 10, 2020 Author Share Posted March 10, 2020 Whilst poking around inside the ECU looking for something else on the mapping front, I couldn't resist checking this out:It's 5.045V:In series with ...1815 ohms.So what I guessed would be 5v supply was actually 1% higher leading me to get the resistance very slightly low, but it confirms what I was thinking. Link to comment Share on other sites More sharing options...
rj Posted March 10, 2020 Share Posted March 10, 2020 Sorry Andrew,I had missed updates on this one. I've never seen a Vref for sensors that was not regulated to +5V on aany ECU I've come across so it makes sense in every way.Pretty sure it'll go in "safe mode" if the input is above say 4.8V or below 0.2V and assume the wires are, or the sensor is, broken. Link to comment Share on other sites More sharing options...
Member Jonathan Kay Posted March 11, 2020 Member Share Posted March 11, 2020 Do you still get the sigmoid transfer function with that explanation?Yes, because of the sender characteristics.Thanks. What's the relationship between the resistance of the sender and the input to the ECU?Jonathan Link to comment Share on other sites More sharing options...
bjw Posted March 11, 2020 Share Posted March 11, 2020 It's a simple potential divider, so:-V = Vsource * Rsender /(Rsource + Rsender)The graph is drawn assuming the analogue to digital converter nominally outputs 1023 for an input V = Vsource (i.e. with an open circuit sender) so becomes:-x = 1023 * Rsender /(Rsource + Rsender)As for the 'S' shape - here's another graph showing three different source resistances - the curves of the sender and the divider fight each other and the source reistor value determines which one tends to dominate:-Cheers Link to comment Share on other sites More sharing options...
Member Jonathan Kay Posted March 11, 2020 Member Share Posted March 11, 2020 the curves of the sender and the divider fight each other and the source reistor value determines which one tends to dominateLovely.Thank youJonathan Link to comment Share on other sites More sharing options...
Recommended Posts
Create an account or sign in to comment
You need to be a member in order to leave a comment
Create an account
Sign up for a new account in our community. It's easy!
Register a new accountSign in
Already have an account? Sign in here.
Sign In Now