Engine overheating when idling.

[skiing]

Are you sure the temperature gauge is reading correctly? I would see if you can find another way of checking the coolant temperature? the voltage regulator can cause the gauges to read high - especially when the current draw if high (e.g. when fans on!) Mine dumped coolant as soon as it was anyway near that indicated temperature - out of the header tank overflow (mine had an airlock sorted by running up to temperature with rear of car on a jack and the coolant cap off while squeezing hoses etc.) - I am sure you have already done this though - 

[NG5]

I'm impressed it will idle smoothly at 750rpm !

Book value for an S3 is 900-1000rpm. On most classic cars I've worked on achieving book value is often a struggle.

Yes there is a big difference in flow for a small change in revs. Flow is proportional to revs (ignoring pumping losses which get worse at high rpm) so going from 750 to 1000 rpm will increase the flow by 25%.

[ChrisJ]

~750rpm is the correct idle speed.

I did mention to raise the revs speed what stationary / stuck in traffic earlier in the thread. Try raising the revs to ~1000 when the fans are running - rather than 2500.

I'm not sure you have proved your otter switch is/isn't working. In my car, the temp goes up to around the mark on the gauge between the 90 & 140, then the fans cut in - very audible - which will bring the temps down to just over 90 - and if you leave the car idling (or are stuck in traffic) the gauge will dance between the 90 & the upper mark. Normal running is just under 90.

I don't have A/C and I also turn the heating/fans on full whenever I think I might be stuck in traffic.

 

 

[NG5]

Quote

~750rpm is the correct idle speed.

Not according to the S3 manual I have...

[ChrisJ]

Mine's always been ~750 (since 1987)

[NG5]

Age! my maths is getting very rusty..

Quote

so going from 750 to 1000 rpm will increase the flow by 25%.

Of course it increases the flow by 33%  Duh.  Quite a lot considering.

No expert on carbed Esprits but still impressed at a 750rpm idle for a highly tuned engine.

[GreenGoddess]

1 hour ago, skiing said:

Are you sure the temperature gauge is reading correctly? I would see if you can find another way of checking the coolant temperature? the voltage regulator can cause the gauges to read high - especially when the current draw if high (e.g. when fans on!) Mine dumped coolant as soon as it was anyway near that indicated temperature - out of the header tank overflow (mine had an airlock sorted by running up to temperature with rear of car on a jack and the coolant cap off while squeezing hoses etc.) - I am sure you have already done this though - 

I checked the gauge today. When the temperature sensor is in boiling water the gauge reads exactly 100 degrees. I was actually quite impressed! The voltage regulator is fine. I’m aware of its foibles. 👍

1 hour ago, ChrisJ said:

~750rpm is the correct idle speed.

I did mention to raise the revs speed what stationary / stuck in traffic earlier in the thread. Try raising the revs to ~1000 when the fans are running - rather than 2500.

I'm not sure you have proved your otter switch is/isn't working. In my car, the temp goes up to around the mark on the gauge between the 90 & 140, then the fans cut in - very audible - which will bring the temps down to just over 90 - and if you leave the car idling (or are stuck in traffic) the gauge will dance between the 90 & the upper mark. Normal running is just under 90.

I don't have A/C and I also turn the heating/fans on full whenever I think I might be stuck in traffic.

 

 

The otter switch triggered today when the gauge was reading approx 120 degrees. But the temperature didn’t move a millimetre even after the fans had been going for about 10 minutes. 

[Chillidoggy]

I'm surprised that the coolant temperature can reach 120 - 140C and still not boil up, or dump anything. If the gauge is to be believed, the coolant hoses must be like balloons!

What the pressure cap rating? 

[Gis]

10 hours ago, GreenGoddess said:

The otter switch triggered today when the gauge was reading approx 120 degrees. But the temperature didn’t move a millimetre even after the fans had been going for about 10 minutes. 

The standard otter switch triggers at 82-92 degrees, not 120 degrees. I think the culprit is the 10v voltage regulator. These things are super crude and if mounted incorrectly (they have an arrow on the case for correct vertical alignment) deliver an incorrect signal to the temp gauge. I’d replace the old bimetallic voltage regulator with a modern 10V electronic one that you can get for a few quid on eBay and try again. That way you can eliminate a potential error source.  Clearly the car would boil at a real 120 degrees, let alone 140 degrees. 

[GreenGoddess]

48 minutes ago, Chillidoggy said:

I'm surprised that the coolant temperature can reach 120 - 140C and still not boil up, or dump anything. If the gauge is to be believed, the coolant hoses must be like balloons!

What the pressure cap rating? 

I don’t know what the pressure cap rating is. I would need to check. I believe only the water circulating around the engine and past the temperature sensor is getting over 120 degrees. The rest of the water in the chassis pipes and passing the otter switch into the radiator is much lower than that. I could understand it boiling over if the entire coolant system was at 120+ degrees but it’s not.

11 minutes ago, Gis said:

The standard otter switch triggers at 82-92 degrees, not 120 degrees. I think the culprit is the 10v voltage regulator. These things are super crude and if mounted incorrectly (they have an arrow on the case for correct vertical alignment) deliver an incorrect signal to the temp gauge. I’d replace the old bimetallic voltage regulator with a modern 10V electronic one that you can get for a few quid on eBay and try again. That way you can eliminate a potential error source.  Clearly the car would boil at a real 120 degrees, let alone 140 degrees. 

I think the voltage regulator is fine. The temperature gauge is 100% accurate, as confirmed yesterday. The voltage regulator has, on rare occasions, had a bad 30 seconds every now and then but the temperature gauge and fuel gauge always DROP simultaneously for a few seconds, they never go up. 
When my car is overheating the fuel gauge doesn’t move. 
 

My next thing to check and adjust is the idle speed. I’ll try and see if the problem comes back when it’s idling at 950rpm instead of 750rpm. If that doesn’t cure it then it’s time to test and change the thermostat I think. 

[Chillidoggy]

The coolant out of the engine will be the same temperature until it’s cooled by the the radiator, and returned. You have a laser thermometer, what’s the temperature of the coolant going into, and out of, the radiator?

You mentioned in one of your earlier posts that the coolant temperature had reached 140C. Unless there’s a blanking cap on the header tank, the system would definitely have boiled up at that temperature, and dumped coolant. Even at 120 C, boiling is a strong possibility.

Also, there will be differing temperatures in the system, but pressure will be the same throughout.

[NG5]

If it's not boiling at 140C then it sounds more like the sender or gauge than an actual coolant problem.

The senders are simple variable resistors. Circuit is power to gauge to sender to earth. 140C looks like max reading which is what you get if its direct to earth with no resistance from the sender. The inside of senders vary in the way they work but it looks like yours might be earthing and sticking when hot, then releasing when a lot cooler. You say it goes down rapidly when it does go down. How rapid is rapid? a few seconds or a couple of minutes. If it's the former then odds on it's the sender, the coolant can't cool down that quickly no matter how good the circulation. If the latter then perhaps not, not sure why it wouldn't stick at 100C though when you tested it, but possibly the angle of the sender into the coolant.

Interesting problem

[GreenGoddess]

3 hours ago, Chillidoggy said:

The coolant out of the engine will be the same temperature until it’s cooled by the the radiator, and returned. You have a laser thermometer, what’s the temperature of the coolant going into, and out of, the radiator?

You mentioned in one of your earlier posts that the coolant temperature had reached 140C. Unless there’s a blanking cap on the header tank, the system would definitely have boiled up at that temperature, and dumped coolant. Even at 120 C, boiling is a strong possibility.

Also, there will be differing temperatures in the system, but pressure will be the same throughout.

I tried a laser thermometer a few months ago. It was a Chinese special from Amazon and was absolutely hopeless. I sent it back. I would like to try one again but only if I can get hold of one that is accurate and has a very concentrated beam.

Regarding boiling, it's quite difficult to figure out. If the coolant around the engine is at 120 degrees and the rest of it (in the chassis pipes and radiator) is, say 85 degrees, then would the system boil over? The water passing the otter switch must be less than 90 degrees because the switch has never triggered the fans until yesterday after the car had sat idling for ~45 minutes.

30 minutes ago, NG5 said:

If it's not boiling at 140C then it sounds more like the sender or gauge than an actual coolant problem.

The senders are simple variable resistors. Circuit is power to gauge to sender to earth. 140C looks like max reading which is what you get if its direct to earth with no resistance from the sender. The inside of senders vary in the way they work but it looks like yours might be earthing and sticking when hot, then releasing when a lot cooler. You say it goes down rapidly when it does go down. How rapid is rapid? a few seconds or a couple of minutes. If it's the former then odds on it's the sender, the coolant can't cool down that quickly no matter how good the circulation. If the latter then perhaps not, not sure why it wouldn't stick at 100C though when you tested it, but possibly the angle of the sender into the coolant.

Interesting problem

I am absolutely certain that the sender and the gauge are OK. I replaced the sender a few months ago as a precaution and when I tested it yesterday in boiling water the gauge read exactly 100 degrees. The needle on the temperature gauge does not move quickly like it would if the voltage suddenly changed. When I increase the revs to 2000rpm from idle, the temperature gauge goes down from 120+ to 90 degrees within a couple of minutes.

I'm off to research how to adjust the idle speed. I'll be amazed (and delighted) if that cures the problem.

[Gis]

Idle speed adjust is a simple screw sitting between the carbs where the throttle cable meets the carb spindle lever. Set it to 950-1000 rpm when the car is warm. Basically a 5 mins job

[GreenGoddess]

2 minutes ago, Gis said:

Idle speed adjust is a simple screw sitting between the carbs where the throttle cable meets the carb spindle lever. Set it to 950-1000 rpm when the car is warm. Basically a 5 mins job

Perfect, thanks. I'll take the car out for a decent drive next weekend and then try to adjust the idle speed when I get back.

[jonwat]

1 hour ago, GreenGoddess said:

Perfect, thanks. I'll take the car out for a decent drive next weekend and then try to adjust the idle speed when I get back.

Number 2, don't get confused with the balance screw 😉 

[semi42]

When the engine is hot and you switch on the heater, does the air blasting from it correspond to what your temperature gauge is showing? 
if it is relatively cool, you’ve got an airlock 

[GreenGoddess]

3 hours ago, semi42 said:

When the engine is hot and you switch on the heater, does the air blasting from it correspond to what your temperature gauge is showing? 
if it is relatively cool, you’ve got an airlock 

It just feels nice and warm like a heater should. Definitely not cool but not excessively hot. 

[GreenGoddess]

Well today I took the car for a 90 mile drive and it drove perfectly. When I got it home I let it sit with the engine idling for a few minutes. As usual, the temperature started to rise so this time I held the revs at 1000rpm instead of letting it idle at 750rpm. Sadly the temperature continued to rise until it was over 120 degrees again. I just turned the engine off at that point and parked it up. 
So, unfortunately, it appears that the low idle speed is not causing this problem.

My next move is to take off the front undertray and check that the front of the radiator is not blocked up with leaves etc. I am 99.9% certain it won’t be because the radiator looks brand new (aluminium V8 radiator). 
 

Then I will borrow a decent infrared thermometer and check the temperature of the pipe that houses the otter switch. I am fully convinced that it will be around 80-85 degrees when the temperature gauge is showing 120+. Naturally I will also measure the temperature of the curved pipe across the top of the engine bay that houses the temperature sensor and I strongly suspect that will correlate with the temperature gauge reading. 
 

If those tests match my suspicions then I can see no other option other than to change the thermostat. I can’t see it being anything else that is causing this other than perhaps the water pump. 

[rgh0]

Infrared thermometres will only give a rough indication and normally read low on rubber pipes due to the outer surface being cooler than the inner as rubber is a good insulator.  If used on Aulminium or stainless they also read low due to the emissivity of the surface being lower and registering a lower temperature.  The best way is to use a contact thermistor probe. many multi meters have a temperature function on them.

 

Either the otter switch or temperature reading on the gauge is incorrect.   

 

cheers

Rohan

[GreenGoddess]

6 hours ago, rgh0 said:

Infrared thermometres will only give a rough indication and normally read low on rubber pipes due to the outer surface being cooler than the inner as rubber is a good insulator.  If used on Aulminium or stainless they also read low due to the emissivity of the surface being lower and registering a lower temperature.  The best way is to use a contact thermistor probe. many multi meters have a temperature function on them.

 

Either the otter switch or temperature reading on the gauge is incorrect.   

 

cheers

Rohan

My intention was to use the thermometer on the aluminium pipes. I have no idea what a contact thermistor probe is but I will Google it. 
I don’t think the otter switch or gauge are faulty.  Certainly not the gauge, that is 100% accurate as per my previous test. 

Edited May 14, 2023 by GreenGoddess

[rgh0]

Infrared thermometers on alumimium pipe will typically read 20 to 30 C low.  If the temperature gauge is accurate then your Otter switch is not working

[GreenGoddess]

26 minutes ago, rgh0 said:

Infrared thermometers on alumimium pipe will typically read 20 to 30 C low.  If the temperature gauge is accurate then your Otter switch is not working

Thanks. Sounds like the infrared thermometer idea is a waste of time then. I believe the offer switch is working. I just think that the coolant passing it is not reaching 90 degrees, unlike the coolant circulating around the engine which exceeds 120 degrees. 

[Bibs]

Why not just change the otter switch? Worse case, you've got a new one in there and best case, problem solved

[rgh0]

The temperature the indicator gauge sees is the same the otter switch sees as they are just connected by a straight pipe.  if the gauge sees around  90C then the Otter switch should turn the fans on. if the gauge is accurate and then fans do not turn on then replace the otter switch or fix the fans