Electric fan switch hysteresis?

What temp stat have you got in? 88/79 sounds quite low to me - stats are normally around 82.

So an 88/79 is going to turn on quite quickly once the stat is open and the temp starts to rise in traffic, and never turn off, because the stat is going to keep the temp above 82.

Good question Adrian and off hand I don't know. 'Probably' whatever was stock for a 1978 Mk2 Ford Escort 1300?

Understood, in theory, but let me tell you what I see in practice.

If you watch the temperature gauge it raises from the bottom to around the middle and generally stays there, whilst you are driving through free air.

If you are then stuck it traffic the temperature often rises to just under the red before the electric fan cuts in. However, it doesn't stay in, cutting back out very quickly (it's quite a big fan and rad) but leaving the temperature way above what I would describe as the 'normal' running temperature.

The consequences of this can be that the carb overheats (the inlet manifold is water cooled / heated) and then the car starts 'choking', faltering and stalling etc (not helped by a lack of ventilation in the engine bay but like I say, it *was* ok like that for many many years).

So, whilst I hear (and agree) with what you say in theory, it's not quite what I see in practice and I know has worked well in the past with a 'different' fan switch.

With the switch mounted in the thermostat housing (I had to modify an Escort housing to take a switch as a Fiesta one faced the top hose in the wrong direction) it probably relies on the whole housing getting to temperature before allowing the (heavy brass) switch itself to come up to temperature. I think this delays (raises) the temperature that the switch actually switches, so my nominal 87 DegC when the switch is tested in isolation, could be higher when mounted in the car?

However, ignoring the usage in my particular situation for a moment, do you have any though's on the lack of hysteresis on this switch please?

Have you tried using oil temperature as apposed to water temperature to control the fan , it may be more reliable and offer a longer running period.

Other options might be some form of timed overrun circuit which allows the fan to run on even when the temperature drops placing the sensor elseware in the water system or a sensor with a larger differential switch, you may need to replace the thermostat too.

I had similar issues in a diesel transit

I would just put in a manual switch for use when stuck in traffic.

No, I haven't, not yet anyway. ;-)

I had thought about that, or making an entirely electronic solution, it's just more time and faffing for something that should (and has) 'just worked' in the first place. ;-(

I have considered making up my own sensor solution, out of suitably sized copper fittings but again, it did work perfectly ones just as it was. ;-(

I could use something like this (in the top hose)

"The switching temperature is fixed and turns the fan on at 90 deg C and off again at 80 deg C."

Ah, now we are back to my main question (and the whole issue). If I have bought a switch specifically designed to be used for what I am doing, and is supposed to have a hysteresis but doesn't, is it simply faulty? Is it maybe part of a faulty batch as I think I have now seen two that behave identically? Or, is it my tests aren't able to make the switch work as designed (I can't see how that is the case).

I think I did that the last time I looked into this 'just because' etc. ;-)

If I can't resolve the question of 'why doesn't this switch function as expected' I may just splash out on an adjustable solution that would at least give me the option to set the temperature as I choose.

Cheers, T i m

For this electronics guy and even on a 1978 Ford Escort I feel that would be a bit too 'low tech' MrC (+ daughter is now insured on it and the wife does drive it occasionally and either might forget / not realise). ;-)

No, I'm going to take the new switch back to where I bought it and ask the question, 'why is there no hysteresis as I understand there is supposed to be please?'

I'm not sure I'll get an accurate answer in reply or just the offer of my money back but that isn't what I want. I 'just want' what I was supposed to have bought and has previously worked on my kitcar for over 20 years. ;-(

Cheers, T i m

T i m wrote: [snip]

So whatever the markings on the fan thermostat, the engine gets too hot before it operates.

Try this:

Disconnect the fan thermostat and temporarily fit a manual switch.

Run the engine until the temperature gauge shows a little above normal running temperature. Operate the manual switch so that the fan runs. Wait. Does the temperature gauge show that the engine cools below the normal running temperature?

If not, the fan isn't big enough for the task.

If the engine cools in a reasonable time, cycle the manual switch as needed to maintain the value shown on the temperature gauge at your desired "normal".

Then do your best to measure the water temperature at the point where the thermostatic switch is mounted. I fully expect this to be significantly below the 88 degrees marked on your existing thermostat. If so get a thermostat set for this lower temperature.

Alternatively, when stuck in traffic switch off the engine. Many "Eco" cars do this automatically.

On this particular thermostat, yes.

Yes. I have already played with that previously.

The fan is as big as the rad and a while back I 'upgraded' the rad for one from a 2L Cortina, ready to fit my 1600 Kent or 2L Pinto.

Ok.

I have done similar previously with the thermometer in the neck of the rad. The thermostat seems to be doing what it says on the tin but the fan switch isn't (now).

Isn't that a bit of the tail wagging the dog Graham? The fist thing

Yes, I've been doing that as well (and often do in our other cars if the wait is going to be predictably 'long' (like at railway crossings). However, if I have just been say going up a hill in the kitcar and then come to some slower traffic, the heat energy has already built up in the engine stays there, meaning it's difficult to re-start (you often have to floor the throttle to clear the fumes).

No, what I think I need to do is get back to what had worked perfectly well for over 20 years and find out why this brand new fan switch doesn't do what it is supposed (before I fit it). ;-)

Cheers, T i m

Two switches with different set points, and a relay? One switch, plus a delay timer?

Could it be that the switch needs to be powered and earthed in order for it work as expected (domestic CH thermostats have this)

I'm guessing it could but I'm not sure how it could, given that the contacts aren't polarised in any way? And in use, the switch is powered (if only feeding a relay coil) and earthed and still doesn't appear to work as expected? ;-(

I think these electric fan switches use a bi-metallic dome actuator (like the little momentary dome switches you find on many button panels these days) and so it should have one temperature where it flips over to the on position and a lower temperature where it flips back.

They could be different bit what of the two wire devices that are fitted in-line with the top hose, no additional earth available there (unless one side is earth and takes the other 'powered' side to ground to activate the fan / fan relay etc)?

Do the domestic CH stats use an electro-magnet instead of a permanent magnet to provide the hysteresis now or sumnmat?

Cheers, T i m

Yes, there are loads of ways to skin this particular rabbit, I was just hoping to be able to get back to the very simple and reliable way it was for 20 + years. ;-)

It's funny, on nearly every car I've had that had anything over and above the basics required to make things work, they have gone wrong. The old 2L Sierra Estate never had any problems with the power steering, electric windows, central locking, engine or other electronics, immobiliser, alarm or electric fan switches because it didn't have any of them. And outside the Luminition Electronic ignition I fitted on the kitcar when we built it (+27 years ago), this electric fan solution has been one of the most troublesome (two Kenlow fans seized up and now this switch issue).

The fan on the Sierra was 'thermal' and in the 23 years I drove it (from new) I never had any issues with the fan. Now I'm now wondering if I could get one to fit the kitcar. ;-)

Oh yes, the one 'modern' thing that came with the Sierra was the automatic choke and that failed after two months, I replaced it with a manual one that was fine from then on. ;-)

We have just replaced the Alternator Decoupler Clutch on daughters Transit Connect ... another fancy feature waiting to go wrong. ;-(

No, I think I need to try and be sure *exactly* how these electric fan switches work and then see why the new one I have here doesn't work as expected.

Cheers, T i m

Which is what you'd expect - the 'stat controls the lower bound, there's sufficient airflow to stop the rad getting much higher.

Again, pretty much what you'd expect - you don't want the fan temps too close to the 'stat temp, because they could start fighting each other.

Mmm.

The obvious answer, of course, is that that switch is buggered.

Do you have a cheap IR thermometer - less than a tenner off the bay, and invaluable for this sort of troubleshooting. It'll also help you see what difference in temps you're getting between the manifold and where the gauge sender is, and between the gauge sender location and the fan switch location.

I wonder if the rad might be starting to silt up a bit, restricting flow slightly around the fan switch?

Is the water mechanically pumped around the vehicle or a separate electrical pump , have you checked the pump, pressure , vanes etc, if you up rated the radiator did you upgrade the pump , hose sizes etc .

A bigger rads fine but useless if the pipe work or pumping system can't feed it , it may also be possible that your overcooling the water, having a large temperature differential across the engine and your sensors picking this up. try screening off part of your rad . You get something similar happen in some central heating systems

Best of Luck Tim this is one of those issues that will take for ever to fix ad when you find the solution it will be something so simple you will be kicking your own arse

T i m wrote: [snip]

I think I may have confused you - wherever I have used the word "thermostat" I should have written "thermostatic fan switch".

As to hysteresis: ....

There should be absolutely no need for any hysteresis. In fact it is often an unwanted limitation of many devices - it's present even when you don't want it.

The only reason for requiring it is to avoid too frequent operation.

The classic example of a control system is the traditional electric bell; at rest current flows through the switch and the magnetic coil and pulls the clapper towards the bell. As the clapper moves it opens the switch so the magnetic pull stops and the clapper is released, returning to its rest position by virtue of the return spring. The frequency at which this happens is controlled by the rate of decay of the magnetic field and the mechanical intertia of the clapper and spring assembly. But it oscillates quite fast - which is what you want in a bell.

Suppose the control system is intended to maintain the level of water in a tank, when there is a variable rate of loss as water is used. You fit a level switch. When the switch indicates "low" a pump operates, and fills the tank. As the switch indicates "high" the pump stops. There may be very good reasons that the pump should not operate too frequently

- for example to avoid excessive starting currents. But the level switch might be very sensitive so would operate the pump far too frequently. So you apply hysteresis.

The analogy in your car is that you don't want the fan to operate too frequently. This would be unlikely since the rate at which the cooling water changes temperature as it passes the "thermostatic fan switch" would likely ensure that the fan would remain operating for a few seconds at least. So hysteresis might be a useful feature; but more likely it is an unavoidable by-product of manufacure, so has been designed for a specific figure rather than be whatever random value the manufacturing process achieves.

If you leave the engine running presumably the fan does cycle on and off; and the engine temperature stabilises at a value higher than you would like. This simply means that the "thermostatic fan switch" does not operate at the required temperature. So get one that operates at a lower temperature. But forget about hysteresis - there is enough in the rest of the system anyway.

Except that - there may be an unexpected feedback path. Suppose the "thermostatic fan switch" is placed in the airflow from the cooling fan. When the fan operates the "thermostatic fan switch" cools very rapidly, and the fan stops. But the rest of the engine continues to get hotter. In time this heat spreads to the "thermostatic fan switch" and it runs the fan. But the engine temperature would stabilise much higher than the specified temperature.

In a domestic CH system the reason for the power and earth is quite different.

When the house is cold, the switch closes and heat is supplied.

Whne the house reaches the set temperature the switch opens and the heating stops. However a typical house uses hot water pumped through radiators. The hot water in the radiators will continue to lose heat to the house, so the tempeature in the house will continue to rise. If the temperature then falls as heat is lost from the house, the switch closes again. So the switch operates infrequently and you notice that the house is alternately too hot and too cold. The hysteresis is excessive, but you can't avoid it.

In control system speak, it overshoots the set point.

The effect can be minimised by applying positive feedback. In my house the thermostatic switch has information printed in English and French; the English describes a component called compensation resistor; in French is is "résistance d'anticipation". The French is a much better description - it helps the thermostat anticipate the temperature rise which will be caused by operating the heating system. It raises the temperature around the thermostat by a few degrees so that it opens before house as a whole reaches the set temperature. So the heating stops. But as explained above the temperature continues to rise - hopefully by just enough.

Of course you can't guess in advance what local heating should be added, so the manufacturers rely on experience.

In essence, the control system no longer overshoots, but it might take longer to reach the set point. So by judicious adjustment of positive feedback the control system can perform quickly but without overshooting.

I think simple control theory should be taught in schools, at the same stage as wiring a mains plug - but I don't suppose they even do that now!

I'm just waiting for non-moulded on plugs to be banned.....

Std mechanical.

Only that it is there and seems to work. ;-)

No. ;-(

No, because I believe they are all the same size if it was the 1600cc engine.

Agreed, but wouldn't that also show itself when the driving normally on the open road? Then engine temperature quickly reaches what I would call 'normal' (on the gauge) and then stays there or slightly above under normal circumstances. Stop or slow considerably and the natural airflow ceases and now you only have the thermal capacity of the cooling system (better with more water etc) and any natural convection and radiation. As soon as the fan cuts in the temperature drops pretty quickly.

Whilst I can see how as things start up that could happen (stat opens and there is a large quantity of cold water going into the engine) but wouldn't the system eventually stabilise? The stat partly opens (or opens a bit more as it is a linear device) and that allows a bit more water though the system? I would consider overcooling as a possibility if it never came up to temperature or the temperature fluctuated wildly (and this is all me just thinking out loud and reporting what I see btw). ;-)

And big lorries and diesels in the winter. I've driven my old Rover

Thank you very much (for that and all your thoughts etc). However, I think my first port of call is to go back to the place that supplied the fan switch, see what they say and see if they have any alternative switches they could lend me to experiment on. [1]

You may well be right bit I have a hunch this is all around a temperature switch that is *supposed* to have some hysteresis and for some reason (that I may well have missed) isn't happening because of how I'm testing it (as was suggested elsewhere).

The problem is ... I don't know how many 'mechanics' / 'mechanical engineers' there are any more (present company accepted of course ). Many 'mechanics' are fitters these days and I'm not sure how many people would ever test the hysteresis of a fan switch pre fitting, or note *how* it was working in the car, other than the fan was going on and off.

It would be relatively simple for me to put a thermistor somewhere (or several in fact) and hook them up to an Arduino microcontroller and program it to respond exactly like I want (even controlling the fan with PWM) but I'm a firm believer that 'computers' and cars shouldn't mix (even though that horse has already well and truly bolted of course). ;-)

Cheers, T i m

Ah yes, gotcha. We already have a thermostat and it's not an electrical switch (in this case especially). ;-)

Ok ...

Agreed ... often ... or to ensure something works or stays between two settings (max / min etc)?

Understood.

Agreed.

Agreed, keeping the water between two levels. Max is obvious and Min is the lowest threshold past which the supply may not meet the demand.

Agreed.

Ok. Except in this case is seems there is no manufacturing byproduct (there is no hysteresis). ;-)

Correct.

Agreed.

I would like to but so far I've not found anything (non adjustable that is). The switch I bought last may well be suitable (even without the hysteresis) as it switches the fan on at a temperature well below that of boiling water (even at atmospheric pressure). However, from my practical experiences, 1) the temperature gauge seems to suggest the water temperature is 'high' (or much higher than 'normal' and higher than I feel I would like it to be and 2) it causes other issues, like the carb to 'kettle' and then driving becomes difficult.

Can you explain from where please? (And this is me just thinking out loud remember ) The thermostat is an analogue device and so can linearly 'moderate' the flow of water through the ( closed) system until the rate of cooling provided by the radiator matches the rate of heating provided by the engine ... or at least it can when there is a continuous supply of cooling air. Once that cooling air is restricted then the thermostat opens further to try to increase the flow of what should be cooler water but it's not cooler, so the stat eventually opens completely and if no further cooling is provided by the radiator (you are in traffic and the electric fan has failed), the engine

I can't see how you can overcool the engine with that setup as the thermostat will still restrict the flow of cold water back into the engine, no matter how much liquid is available.

I can see how a very large quantity of water could act as a thermal capacitor, delaying the time it takes for an engine to boil over (possibly long enough to overcome the need for any supplementary cooling, depending on now bad the traffic was). ;-)

I can also see that with a marginal cooling system and a lot of thermal inertia, not having a supplementary cooling solution cut in soon (low temperature) enough could see the temperature 'overshoot into a point where it boils over. I have had that say when running hard and slowly up a long hill and then stopping at the top. The system may well have brought the temperature down eventually but it would be too late if the engine had already boiled over. Applying the supplementary cooling *earlier* may have saved the situation (with a finite volume of coolant and a finite maximum cooling rate).

Understood. In this case (and I know you weren't suggesting it would affect my case) of course the switch is submersed in water right after the thermostat and whilst I can't say there wouldn't be *any* impact of the fan air (which would be hot air at that time as well) it would be blowing over the outside of the thermostat housing and it's fan switch, possibly cooling it slightly.

Understood.

So, because I may not be able to find a straight replacement non - hysteresis fan switch with a low enough temperature (and this new one may well be low enough etc), the fact that the hysteresis was

But as you say, a fully temperature adjustable electrical switch with no hysteresis would (probably) be a perfect solution here. ;-)

I may well have that (by chance) in the latest switch I bought .... I was just wondering *why* it didn't work as expected / advertised? ;-)

Cheers, T i m