29 replies to this topic

[OzOAP]

I'm trying to sort Innocenti temp gauge out.
Got sender unit of De Tomaso in Holland, that reads too high,(they have a different dash to mine).
Have tried, black, green and red Mini ones. Non accurate.

[mbolt998]

 

Mine seems to be a kind of diode. It's 59Ω at 100°C measured one way round, but 47Ω if I connect the multimeter the other way around. This also happens at other temperatures of warm water, and in air at room temperature. My multimeter is OK-- it measures the resistance across a light bulb to be the same in both directions.

 

It is odd though I don't doubt what you found. It is possible it may have some semi-conductor component to the one you have. What brand and type (or type of gauge) is it ?

 

Not sure. I think it's the one that's been on the car for ~25y but not the original. I remember replacing it for some reason back in the day. Have ordered a GTR101 from Minispares and will compare. It also seemed like it was reading lower and lower resistances if I left it stewing in hot water for several minutes, which might explain why my car seems to "overheat" after driving it for about 10m or so. But I will test this again.

 

Going to investigate more when the new one arrives and compare both of them, and will post what I find here!

[BaronVonchesto]

Mine seems to be a kind of diode. It's 59Ω at 100°C measured one way round, but 47Ω if I connect the multimeter the other way around. This also happens at other temperatures of warm water, and in air at room temperature. My multimeter is OK-- it measures the resistance across a light bulb to be the same in both directions.

 

NTC Thermistors are usually semiconductors. but the polarity shouldn't matter. In order to get current flowing across a diode in the wrong way you need to exceed the breakdown voltage, which any standard multimeter most certainly will not do.

 

I think it more likely the resistance simply changed between the time you took the measurements.

[mbolt998]

 

Mine seems to be a kind of diode. It's 59Ω at 100°C measured one way round, but 47Ω if I connect the multimeter the other way around. This also happens at other temperatures of warm water, and in air at room temperature. My multimeter is OK-- it measures the resistance across a light bulb to be the same in both directions.

 

NTC Thermistors are usually semiconductors. but the polarity shouldn't matter. In order to get current flowing across a diode in the wrong way you need to exceed the breakdown voltage, which any standard multimeter most certainly will not do.

 

I think it more likely the resistance simply changed between the time you took the measurements.

 

It was very repeatable though. I kept swapping it backwards and forwards and always the same result. But I intend to do the whole thing again, and also to try to understand whether the behaviour changes after prolonged stewing.

[Ethel]

You could probably rig something up with a thermos flask for greater consistency, or a central heating system would likely add plenty of "ballast" so smooth out the fluctuations.

[BaronVonchesto]

 

 

Mine seems to be a kind of diode. It's 59Ω at 100°C measured one way round, but 47Ω if I connect the multimeter the other way around. This also happens at other temperatures of warm water, and in air at room temperature. My multimeter is OK-- it measures the resistance across a light bulb to be the same in both directions.

 

NTC Thermistors are usually semiconductors. but the polarity shouldn't matter. In order to get current flowing across a diode in the wrong way you need to exceed the breakdown voltage, which any standard multimeter most certainly will not do.

 

I think it more likely the resistance simply changed between the time you took the measurements.

 

It was very repeatable though. I kept swapping it backwards and forwards and always the same result. But I intend to do the whole thing again, and also to try to understand whether the behaviour changes after prolonged stewing.

 

keep in mind that thermistors also often have a hysteresis effect. It is possible that there is more to the temperature sender than a thermistor. Perhaps to differentiate positive and negative ground cars.

[mbolt998]

OK sorry for sending that post multiple times. I didn't do it on purpose! I've now been boiling it away for over an hour, checking it both ways round every 10m. It's fairly consistently about 59Ω in one direction and 47Ω in the other, although over time it looks like the two values are converging. All these measurements are taken in boiling water (100°C).
.
58.2    47.2
59.5    46.8
59.5    46.8
59.0    47.3
59.3    47.5
59.0    47.3
58.7    47.4
58.9    47.1
58.9    47.5
58.8    47.5
58.3    48.3
57.9    48.9
57.2    49.1
56.6    49.4
56.7    49.7
56.7    50.3
56.1    50.1

 graph.png   106.52K   2 downloads

[mbolt998]

OK mystery solved, or part of it anyway. If I measure the voltage across the thermistor when it's in the hot water I get 1.6mV in one direction and -1.6mV in the other. This is likely due to the Seebeck effect (https://en.wikipedia...electric_effect).

 

Multimeters measure resistance by injecting a small constant current (according to the internet) and then measuring the voltage drop to work out the resistance (V=IR). But the Seebeck effect is adding 1.6mV in one direction and subtracting 1.6mV in the other. This would be consistent with a test current of about 400mA I reckon. Idk what my actual multimeter does (It's a Neoteck NT8233D Pro).

 

I believe this effect should not be significant in the actual car as the thermistor has about 10V across it. 1.6mV here or there will make no difference.

 

But it's something to be aware of if testing thermistors using a multimeter. Try connecting it the other way around. The true value should be the average of the two you record. The difference between the two readings will depend on the multimeter because they may not all use the same test current. The new one hasn't arrived yet but I expect it to do the same thing-- be different in different directions. Will let you all know when it gets here.

Edited by mbolt998, 03 June 2023 - 08:57 AM.

[mbolt998]

OK have now got the new thermistor from Minispares, which they say is a GTR101. I get 55.3Ω in one direction, 48.3Ω, which average to 52Ω. The readings from the old one averaged to about 53Ω. The DC voltage across the new one seems to a bit less (and correspondingly so is the spread between the two readings). I'm getting 0.6mV rather than 1.6mV. However this is a bit variable. At one time I did see 1.5mV.

 

So conclusion: it looks like it's supposed to be about 53Ω average at about 100°C. A new GTR101 from Minispares is giving me the same result to within 1Ω as whatever I bought from some car spares shop 25y ago, although it's still a much lower resistance than the table posted earlier in this thread indicated.

 

Still not entirely sure why my car thinks it's running hot, but there are plenty of other things to look at. I might try the sender in hot water but connected to the car itself to see where it's putting what on the actual gauge next.
 

 

 

 

 

 

 

 

 

 

 

[mad4classics]

At the end of Appendix A in Part II – Electrical senders (transmitters) of the document I posted earlier on in this thread it gives the Smiths TT3803/00A (part number GTR101) as between 43.2 and 57.6 Ohms at 100 Deg C. The reading your getting is within that band.

[sonscar]

The guage is not a calibrated instrument and its environment does not promote its accuracy.I work on the theory that if it is in the position it usually is then all is right with the world(probably),Steve..

[mad4classics]

The guage is not a calibrated instrument and its environment does not promote its accuracy.I work on the theory that if it is in the position it usually is then all is right with the world(probably),Steve..

You're correct re the gauge. However, overall sort of assumes that things are correct in the first place and half the fun is investigating.

Edited by mad4classics, 03 June 2023 - 08:58 PM.

[sonscar]

I have had much fun experimenting with air and water temp transducers to calibrate my ancient Megasquirt one equipped vehicles,Steve..

[BaronVonchesto]

 

The guage is not a calibrated instrument and its environment does not promote its accuracy.I work on the theory that if it is in the position it usually is then all is right with the world(probably),Steve..

You're correct re the gauge. However, overall sort of assumes that things are correct in the first place and half the fun is investigating.

 

 

yea because it works on resistance and current flowing through a solenoid to deflect the needle, even the corrosion on the terminals over time affects the reading of the temperature sender. Periodic recalibration is definitely necessary.

[mbolt998]

At the end of Appendix A in Part II – Electrical senders (transmitters) of the document I posted earlier on in this thread it gives the Smiths TT3803/00A (part number GTR101) as between 43.2 and 57.6 Ohms at 100 Deg C. The reading your getting is within that band.

Thanks! I missed that. The table in this post has rather higher resistances. According to the pdfs, the 6811/01 should also be 43.2-57.6 at 100 (and that range is enormous lol, much bigger than the differences I was getting in opposite directions).

 

https://www.theminif...p-sender-specs/