Exhaust gas temp/ Cyl head temp

[Wentzel]

Do any of you use k-type thermocouples to measure Exhaust gas temp and/or cylinder head temps? I built a "gauge" to display the temp on a LCD display some time ago. I am using it but still sorting out minor things. Would any of you that have a k-type thermocouple available be willing to try it and give me some feedback? My plan is to build another circuit and let someone 'test' it and let me know if it works reasonably good and maybe give suggestions. I would just like to get it to work properly and need some input.

Thanks

P.S. Wasn't sure where to post this

[Dawie]

You could wire up a "simulated thermocouple" to generate the tiny voltages similar to a thermocouple to test your circuit. By using series resistors and a variable resistor/potentiometer.
Type K thermocouple generates around 40 microvolts per degree celcius. If the hot junction is at 200 deg C and the cold junction at 20 deg, you will have 7.34 millivolts. A difference of 1200 deg C between junctions will see around 48.83 millivolts. (The tables are available on the net, as it is not exactly linear).

You probably already know this:
1)We need to compensate for temperature at the cold junction, as a thermocouple measures relative temperature (between the cold/ hot junctions), unlike a thermistor which measures absolute temperature.
2)We need to be very carefull because of these small voltage differences and the high currents and voltages floating around in automotive situations. For instance, say you have a good earth strap between engine/gearbox and chassis of only 0.005 ohms. Alternator charging at 35 amps, and engine will be 175 millivolts below chassis. Which is several times more than our thermocouple output. Normally the thermocouple makes contact with the engine for good thermal conduction. So our circuit has to be a prescision balanced input amplifier with low offset voltage drift and cold junction compensation. And it should be able to handle inputs that are slightly below ground. Obviously the amplifying circuit must be earthed at the engine and not at the chassis.
3) The nice thing is that you can buy thermocouple wire and make your own thermocouples.

[DR WATSON]

Exactly what they said

[Wentzel]

Thanks Dawie yes I did some homework and the circuit has cold junction compensation built in. I also tested it on a bench and it seemed pretty accurate but conditions in the car may change a lot from the ideal test bench scenario.As far as I know the probe style I have is ungrounded so that shouldn't be a problem. I compared my outputs to that of a Fluke reading the TC and compared that to the tables from the internet and they are all pretty close.

Thanks for the input.

[Dawie]

As you are using ungrounded thermocouples it will be a lot less demanding on your circuit. If you screen the cables and have a good interference filter in your circuit input it should not be a problem.

Do you plan to mount the egt probes in the individual exhaust runners a few cm from the heads, or a single probe before your turbo?

[Wentzel]

I have one installed in the collector before the turbo to keep it simple and a bit less expensive. The temperature is probably a bit cooler here but should still be a good indication.

[Wentzel]

Did some work on the circuit and reinstalled it in the car. So far I'm happy it seems much more accurate and sensitive but I will see how it goes over time.