Car temperature monitor

given that increased temperatures are the usual early indications of a problem, and that the normal car temp gauge is completely useless, I am going to make a temperature monitoring system for the car.

very early days - it's all still in my head, but i will try to keep this thread updated with progress.

the requirement

Multiple measurements
LCD display - plus warning lights
programmable warnings
programmable channels/functions
52mm gauge size
K type thermocouples

I can think of a lot of places that temps could be measured - oil and water are obvious, but once you get into serious trackdays or start pushing hard on the road, you may want to measure brake pad temps, intake temps, effectiveness of intercoolers/radiators/oil coolers. on turbo cars then EGT and oil return temps can help too.

so much to monitor, so little time.

AD595 is looking like a good choice of input chip, it's a pity I can't really multiplex the thermocouples to the chip, it would save a lot of money, but thermocouple signals being what thet are......

(if anyone has successfully multiplexed TC before the amplifier chip, please let me know)

it's going to be a rolling project. the first thermocouple will be going in when i change my turbo gaskets next week

Graham

Sounds like an interesting project. First comment I'd make is that some changes in temps are to be expected - some are not. The changes may be up or down. A drop in engine temp, as indicated by the circulating coolant temp may also indicate a loss of coolant or at least a bubble of gas/air near the sensor.

More data will require more attention to read, interpret, analyze, etc. It might be helpful to characterize engine behavior and the relationship of each data point. Temperatures, rates of change, among other things, could be calculated measured/calculated then compared to things like throttle position, changes in throttle position (possibly indicating driver demands), etc. The whole point would be to carefully sift thru the data (real time) to 1) require less attention of the driver allowing time for other critical tasks and 2) look for things that a driver may not ever see.

Again, sounds interesting.

All you need is an EGT monitor after the turbo to tell you what needs to be tweeked in the motor. Other than that, proper coolant mix, properly sized intercooler coils, and upgraded brakes shouldn't need monitoring. Simpler is better.

KISS is definately a sensible approach

i could just buy an egt sensor and guage

in fact i could probably buy all of it if i wanted to

but where's the fun in that?

I see it as one of those projects that will grow with my needs. initially one sensor, and one guage. then add alarm points. but from the start allow for the customisation aspects.

lets say I am happy for my water to be between 90 and 105'c. if it is then the display either won't show it, or will give it a low priority. over 105 it starts being displayed more prominantly and over 115 alarms ring. (those aren't actual temps, just guides)

the important thing though is to fit the tc this weekend, when i have the turbo off

Why not use a cheap junction (from a diode for example)? U can read the voltage drop at constant current using a amplifier and a uC with a decent ADC. It's ~the same precision as using a thermocouple. Are thermocouples a must?
About multiplexing the output of thermocouples? What's the problem. U'l need like a reading at 1/4 s. The ADC can't handle that switching time? The multiplexing can be done with some of uC's outputs activating one at a time a MOS-FET o BJT.

yes, diodes can be used for some measurements, but a lot of the ones i am looking at would fry a diode - the oil from the turbo , the exhaust gasses and the brake pads, for example

multiplexing the thermocouples is a problem, because the signal is so small, the switching interface can cause more noise than the the signal it is switching, so really a relay is the best option, but the dis-similar metals introduced then cause more problems

Ok ...you're right about the diodes. About the thermocouples...I mentioned some BJT or FET's. Use a small amplifier for each imput. A voltage controled amp IC or a transistor amp switched on&of trought a MOS-FET by outputs of a uC. Same thing can be achieved using a opamp and 1 or 2 JFETs (Google).

the tc signal is so small though that anything between the tc and the chip will cause problems.

I've seen some tc's interfaced trought a FET amp to a AD2020 chip. Those things where used in a aut instalation from a factory. Just use good shielding. The FET's have very low noise and very high imput impedance. Also some opamps have more then OK caracteristics. The low voltage imput (few mV) will create problems for ADC chip.

BTW...I was talking about ACTIVE circuits. Not just a switch but a amplifier switched on&of. That wil boost the voltage.

i was going to stick the input into a AD595 (probably) to get 10mv per 'c out