Temp gauge read high with lights on?

[A TR7 16V]

Can anyone think of a logical reason why the temperature gauge would have read high with the lights on, but then gone straight back to normal when the engine was stopped and started again with the lights off? I'm pretty sure the engine wasn't overheating, like the thermostat was jammed shut or aught - the engine wasn't off long enough to cool much, didn't feel especially hot, and it's not doing it now.

I did think it was an earth problem I could look at when it was light. But, thinking about it, I can't see how any earth fault could make more current flow through the sensor and or gauge - which I assume is an ammeter, not a voltmeter. And when I look, I can't find any problems.

So I'm beginning to wonder if it's a faulty sender that's got an intermittent internal short, though that seems a bit bizarre, or something shorting the ammeter's series resistor. 

Graham

[RobPearce]

The temperature gauge isn't really an ammeter or a voltmeter but something in between.

Usual reason for reading high is the voltage regulator failing. If the behaviour you describe is persistent then I'd suggest a bad earth on the speedo (which in turn provides the earth for the regulator). With the lights off it could be earthing through them, and thus providing correct gauge supply. Lights on, regulator earth goes to +12 and the regulator stops regulating.

[A TR7 16V]

Well, actually, the gauge has to be an ammeter. That's because current is all an analogue moving coil or moving magnet meter can ever measure. It measures voltage by measuring the current in a fixed value series resistor, and gives voltage by Ohm's law. A DVM's different, but the process in the ADC is complicated and generally involves detecting when the external voltage is equal to a highly accurate digitally generated variable reference voltage using a fast differential amplifier and a comparator. But that's not important right now.

But a bad earth to the instrument VR would increase the voltage across the meter  and negative temperature coefficient resistor in the sender in series with it, and increase the current through both. And earthing through the speedo light would mask it with the lights off. I didn't notice if the speedo backlight was working at the time - it would be off with no true earth to the speedo head body.

Why that fault has gone away is the next question. 

Graham 

[RobPearce]

If you're going to claim that all analogue meters are ammeters, then why make the distinction in the first place? An ammeter is a moving coil meter with a parallel shunt resistor to make it have very low resistance. A voltmeter is the same moving coil meter but with a series dropper resistor to give it a very high resistance. The bare moving coil is sensitive to very small currents and has an internal resistance that will produce those currents at a very low voltage, and it's generally not very useful for either. The temperature gauge is, firstly, not a moving coil gauge but, secondly, it doesn't have a resistor to adjust its characteristics. It will measure either voltage or current but won't do either very well, which is OK as what it's really measuring is resistance (or conductance if you prefer).

Anyway, I think you're agreeing with my possible cause/effect hypothesis and I agree with your suggested further diagnostic. But as you say, if the problem's gone away it's hard to check, and the question is why has it changed.

[A TR7 16V]

Not all analogue meters are ammeters, just all moving coil and moving magnet galvanometers - its the current through the coil that creates the magnetic field that acts on the magnetic field from the (normally) permanent magnet to cause a torque that is opposed by the spring to give a deflection proportional to the current. Fleming's left Hooke?

The distinction seemed important because of the way an ammeter has to be in series with (in this case, above) the sender. But I guess that wasn't as obvious as it might seem to an electronics graduate.

And yes, I agree, that seems to be the only earth fault that would give that effect and one more than I thought of on me own. So thanks.

But if the fault's not there no more, then maybe the lights are a red herring, and the problem is just in the Smiths electro-thermal voltage stabilizer, which really isn't cool. If the points stuck shut, but were jiggled/jolted loose by the stop and re-start, that would do the same. I should have put the lights back on when I restarted to see if they were connected (as it were). But I didn't. I went off in a huff.

So I've ordered an LM317 module - I would have got a fixed 10 volt u7810 one, but the variable one was cheaper. Plus it allows me to calibrate it against the fuel gauge.

I'm split between filling the tank and adjusting the voltage till the gauge (another ammeter, i.e. in series above the sender) reads full, or running it till it stops and adjusting till it reads just empty, or doing both and splitting any difference. Personally, I think its more important its right when its empty - and I can drive to the petrol station on the reserve.

Graham

 

[RobPearce]

Quoted from A TR7 16V-

The distinction seemed important because of the way an ammeter has to be in series with (in this case, above) the sender. But I guess that wasn't as obvious as it might seem to an electronics graduate.

OK, I was thinking about the characteristics of the device, not its position in the circuit. Probably because I'm a physics graduate. 😛

 

Quoted from A TR7 16V-

I'm split between filling the tank and adjusting the voltage till the gauge (another ammeter, i.e. in series above the sender) reads full, or running it till it stops and adjusting till it reads just empty, or doing both and splitting any difference. Personally, I think its more important its right when its empty - and I can drive to the petrol station on the reserve.

I sort of agree that the empty point is more important but I prefer not to run out of fuel as it hits the mark. Also, you're going to need a lot of adjustment to make a fairly small difference at the bottom, where the sensor's resistance is high (dominant, much more than the gauge's), so you may end up being unable to achieve it. I'd almost be tempted to adjust for temperature gauge in the middle when fully warm because that's going to worry me more if it's wrong. As long as the fuel gauge gets "about" full when I fill up and goes off the bottom before I splutter to a halt, I'm not that fussed about details.

[A TR7 16V]

Still, the meter measures the force caused by the magnetic field caused by the current caused by the voltage accross the resitance of meter's coil. So the voltage can only be derived from the measured current, knowing the resistance of the coil. Thus the meter's primary characteristic (of the two) is as an ammeter, which can, secondarily, be used as a voltmeter given that additional information.

Since the fuel and temp gauges are near enough right at the moment, it should just be a case of setting the LM317 to give the same voltage as I've got now. But that's ignoring the effective internal resistance of the stabalizer/regulator. I assume the effective resitance of the Smiths electo-thermal thingy's is small in normal operation, or the fuel sender would visibly affect the temperature reading. And I'm assuming that of the LM317 is bog all (though I could probably look that up). But I've been wrong before.

Graham