Electric trickery check required...

[Rich H]

Hey all,

Quick electical question to check I'm not being too daft...

Esprit has 2 water temp gauges, one on the dash and another for the ECU.
The ECU is accurate and reliable the dash one is crap and cheap. (sender is new, wiring is OK)

The thermostat is 88* and when up to temp the ECU one reads around 88* but the dash says about 75* which is nearly in the cold section of the gauge.
I'd like to bring the dash gague so that the needle is in the middle of the normal range as it should be when it's at 88*

Now the sender resistance decreases with temperature so zero resistance = really hot.
If I was to put a high value trimmer (1Mohm?) in parallel with the sender would that allow me to trim the gauge in the right way?
I think it would but it's been a while and my elecronics is rusty...

In my head adding a parrallel trimmer will effectively lower the resistance of the sender/trimmer combo and therefore raise the indicated temp - yes?
High value will lower the resistance a little bit so give a relatively small rise in temp. Or would high value resistor + a small trimmer be better for fine control?

The oil temp gauge needs the same treatment...

Thanks
Rich

[robin]

Of course the correct solution is a carefully design op-amp circuit - this is what they were invented for (applying y = mx + c to signals).

However, if you must take the cheezy route ....


You are right, in a way; the temperature sensor will be a NTC (Negative Temperature Coefficient) resistor.

Assuming you don't care what the thing reads at any other temperature, allow the engine to heat up, then measure the resistance of the sensor. You want a 17% increase in resistance, so you need to drop the effective resistance by 17%.

Assume resistance at 88 = R.

Resistance with your shunt S in parallel = 1/(1/R + 1/S).

You want that to be 17% less, so 0.83R = 1/(1/R + 1/S).

1/(0.83R) = 1/R + 1/S

1/S = 1/(0.83R) - 1/R = 1.205/R - 1/R = .205/R

S = R/.205 or approximately 5 x R.

Cheers,
Robin

[Rich H]

Nice one! Yes cheesy route, as it's easier!
I got 6 x r but close enough

Not bothered about response away from nominal as it'll still give hotter and colder than nominal but I'll check that it still reads sensibly elsewhere. I have a spare sensor so I can boil that up and measure it's responses

I suppose the only problems will be if the sensor fails but that should show up as a off scale cold still

Cheers!
Rich