An accurate temperature gauge at last?

[Uncle Fester]

I emailed Emerald these questions. If I am understanding this correctly, we will soon be able to believe our temp gauges.



Can the Emerald be used to drive the existing temperature gauge?

Can an output from the Emerald be used to control the temperature gauge using a signal from the ECU temperature sensor?

I wondered if a unit to take the signal that would go to the digital dash and could be adjusted to the scale of the original temperature gauge is available? If it isn't there are a lot of TVR's with the same problem and I suspect that there would be a market for them.


Karl at Emerald replied

We have used the ECU's "user PWM output" to control a normal temperature gauge. The user PWM output can be scaled between two coolant temperature points which is fine over a small range but it's unlikely you'll be able to get the gauge sufficiently accurate across the whole cold-hot range.
There will be an ECU firmware update released soon which includes an improved user defined output option (as well as an additional user output). With this revised user output function you should be able to tailor the gauge to temperature to suit. With either function a PWM output frequency of 100Hz should work fine (the gauge smooths the signal so it won't see individual pulses but instead sees an average voltage).

[rev-erend]

Hooray..

Until then - drill a large hole thru the dash and bulkhead and aim a lazer temperature guage at the water system

[GreenV8S]

Isn't it simpler just to fit a thermister and a gauge? Some things are best done electronically, but for measuring a simple temperature all this A-to-D and variable duty cycles and tweaking it so the gauge displays the number you first thought of just seems like the wrong approach to me. You already have an analogue gauge and a thermister, why not just connect one to t'other and calibrate it?

[dnb]

I tend to agree with you Peter. The TVR gauge is calibrated use the same thermister as the Lucas ECU and the manifold can be modified to accept 2 of these.

[Uncle Fester]

Originaly I had the original sender out on the end of a little bit of pipe. The reading wandered up and down mso much as to be useless.

Now I have the original Rover sensor connected up via a little variable resistor that David Beer supplied. It's much better than the original in terms of stability. But it seems to be accurate at part of the range, but not others.

I agree simple is best, if only I could get simple to work.

I thought nobody had satisfactory temp gauges. Whats this about modified manifolds? Not heard that one before.

The other advantage would be that if the display shows the reading from the ECU sender then you are at least seeing the information the ECU is working with. If the ECU temp sender was faulty, it would show on the dash.

So what have others done that actually gives an accurate and stable reading across the entire gauge scale?

[dilbert]

Maybe you could use a thermocouple, and something like this;

http://www.national.com/ms/LB/LB-24.pdf

The thing is that thermistors are horribly inaccurate.

[shpub]

Uncle Fester said:

I thought nobody had satisfactory temp gauges. Whats this about modified manifolds? Not heard that one before.

The other advantage would be that if the display shows the reading from the ECU sender then you are at least seeing the information the ECU is working with. If the ECU temp sender was faulty, it would show on the dash.

So what have others done that actually gives an accurate and stable reading across the entire gauge scale?

My Griff gauge using the standard sensor has been accurate during the 6 years I have had the car. Right across the scale. Checked on many many occasions with an IR thermometer. The vast majority of them are OK. You only get to hear of the few that are not. It helps to get things into perspective: the gauge is not a lab standard accurate thermometer but something to gage the temp.

The David Beer mod will only give you an accurate reading only at its calibration point because it is a fixed resistance. There is nothing simple than can be done with this. You could program a small micro to calibrate the sender and send out the correct signals but that is not really simple.

As for using the Emerald... bear in mind that most temp sender failures are not catastrophic but calibration problems such as being 20 degrees out. It is still a valid signal but not one that is correct for the current engine environment. The ECU will not detect it as a fault and to be honest, I suspect that the temp gauge wouldn't be much help as it would be a valid reading but not right but how would you know unless you have a second gauge to compare it with. To be honest, it is now paranoia time. Who's checking who? If they are different which one is wrong? Maybe a third gauge is needed. It starts to get complicated and very far removed from the keep it simple routine.

Providing you can cope with the aesthetics, the digital Spa Design gauges are the best out there IMHO. Get a dual temperature one and monitor oil and water temps. That would give you a lot of information. You can even set up alarms to warn of problems.