Megasquirt Support Forum (MSEXTRA)

I'm using a Denso MAF Sensor #197-6130 also known as a Subaru #22680-AA38A and need to get the calibration values for the IAT part of the senor. Problem is its integrated with the MAF and I do not think it would be a good idea to submerge it in ice water and then boiling water for the low and high ohm readings needed to generate a calibration table in TunerStudio.

Anyone done this before or have an idea how to do this?

I hate to say this but (and yes the pun is intended) the path of least resistance might just be to use a separate temperature sensor that has known characteristics.

Or... do you have access to the factory service manual for the car? Sometimes the manual will have test procedures for temperature sensors, that tell what the resistance should be at various temperatures. Or they'll give ranges of resistance, in which case you just use the middle of the range.

Two easy options I see here.

Pull the data from the original ECU. Use rom raider and dig around in the factory ROM for the data points.

You can get three points with household appliances. *all with a calibrated thermometer! At room temp.In the freezer. In your oven set to 180F or whatever the lowest setting is.

jacky4566 wrote:Two easy options I see here.
Pull the data from the original ECU. Use rom raider and dig around in the factory ROM for the data points.

I did pull the factory ROM and found the IAT scale factors but without knowing the factory ECU bias resistor the values do not have much use. I pluged them in as a custom inc file and the ambient temp readings were way off using the 2490 Ohm bias resistor built into my MS3-Pro.

jacky4566 wrote:Two easy options I see here.
You can get three points with household appliances. *all with a calibrated thermometer! At room temp.In the freezer. In your oven set to 180F or whatever the lowest setting is.

I was going to go there as a last resort, especially the oven one. The wife really loves it when I put car parts into her oven.

I'd still use a different sensor but for what it's worth, figuring out what bias resistor the OE ECU uses should be pretty easy, if you have an OE setup to fool around with.

First thing to do is disconnect the sensor and measure the open-circuit voltage going to the sensor.

Then hook up a potentiometer in place of the sensor, with the wiper connection going to ground and one (only one!) of the other pins going to the sensor lead. You may have to fool around to find the correct value but 10K is a good guess. Put a voltmeter across the potentiometer and adjust the pot so the voltage reads exactly half of the open-circuit voltage. Then shut the power off, disconnect the pot, and measure the resistance across the same terminals you had connected in place of the sensor. The pot's resistance will equal the bias resistor in the ECU.

elaw wrote: figuring out what bias resistor the OE ECU uses should be pretty easy, if you have an OE setup to fool around with.

That's a problem. I don't have the OE ECU. I only have the ROM image from RomRaider site.

But, thanks for the idea.

The ROM data you have... does it translate voltage to temperature?

If so, and if you think you can safely assume the OE ECU provides 5 volts to the bias resistor, you can probably calculate it mathematically.

elaw wrote:Or... do you have access to the factory service manual for the car? Sometimes the manual will have test procedures for temperature sensors, that tell what the resistance should be at various temperatures. Or they'll give ranges of resistance, in which case you just use the middle of the range.

This suggestion may have panned out! I found the factory service manual for the engine and it does list three data points for testing. I will plug them into TunerStudio tomorrow and see if the sensor outputs sensible data.

Thanks!

elaw wrote:I'd still use a different sensor but for what it's worth, figuring out what bias resistor the OE ECU uses should be pretty easy, if you have an OE setup to fool around with.

First thing to do is disconnect the sensor and measure the open-circuit voltage going to the sensor.

Then hook up a potentiometer in place of the sensor, with the wiper connection going to ground and one (only one!) of the other pins going to the sensor lead. You may have to fool around to find the correct value but 10K is a good guess. Put a voltmeter across the potentiometer and adjust the pot so the voltage reads exactly half of the open-circuit voltage. Then shut the power off, disconnect the pot, and measure the resistance across the same terminals you had connected in place of the sensor. The pot's resistance will equal the bias resistor in the ECU.

Nice trick, trying to remember when I need it.