Speed controlled on/off relais

I am an oldtimer freak and replaced the mechanical crankshaft-driven fan by a dc-ventilator. This is temperature controlled (90 ° C in, 80 ° C out). Since de engine oil is aircooled, the fan starts after 10 min drive and never stops till the car has stopped and the oil is cooled down to under 80 °C. I have tested the cooling without fan and notice that the drive wind is sufficient to keep the oil cool enough.
So the fan has to turn 1) in extreme hot conditions (I can do that with a simple switch because I have a temperature dial and 2) when the speed of the car is too low to provide sufficient air. The engine is encapsulated by plating, has a big open mouth in the front an runs the air over the engine heads (its a 2 cylinder boxer). So at sufficient speed the air is more or less forced to run over the engine ends.
Now my question : I want an electronic circuit switching a 12 V relais on and off at a set speed. E.g. at 40 km/h the relais should be switched off, below 40 it should switch on. I should be able to preset the switching point at higher speeds too. I think of applying a hall-sensor on the shaft going to the frontwheel. The pulses from this sensor should be used to count and trigger a relais via a transistor. The circuit in between shall handle the pulses from the hall-sensor and should include a regulator so that I can preset the relais switching at a defined speed of the car.
Many thanks of the support in this topic.

Welcome to ETO!

Edce said:

I have tested the cooling without fan and notice that the drive wind is sufficient to keep the oil cool enough.

Click to expand...

Then 'cool enough' must be something above 80C, otherwise the temperture controller would have switched the fan off. Can't you simply adjust the controller to switch off at the 'cool enough' temperature?

The "controller" is just a temperature sensor (variable resistor dependent on temperature. So in fact not a component which can be adjusted.

The circuit is doable; I have the basis of one on the drawing board.
Is the temperature cutout integral with the fan motor or entirely separate?
Can you post a sketch of the present fan/controller/switch wiring?
How many pulses per wheel revolution do you intend to generate using the Hall sensor?
What is the tyre tread diameter?
Will the relay be used to switch the +12V connection of the fan motor or the ground connection of the motor?

High Alec_T: many thanks for your interest in my project. You put the right questions. I have all info you need available. The temp cutout is sitting around the tube of the oilcooler at the hottest spot (where the oil comes out of the engine and runs to the oil cooler). The fan is driven by a simple 12 V DC 80 W motor.
I can add 2 or 6 metal strips to the outgoing shaft driving the wheel. If I would use a Hall sensor then I need to stick magnets to bare metal which I think is not so reliable. If I place 2 or 6 metal strips an inductive sensor is the solution. If you are of the opinion that a Halls sensor is better I will find a reliable solution for the magnet fixing.
The wheel circumference is 184 cm. I prefer to switch the + 12V connection. If other questions pop up, please feel free to ask.
Many thanks for your cooperation.

Ok. Let's go with the inductive sensor and a 6-tooth 'tone wheel'.
What is the approximate road speed at which you want the motor to cut off?
Does the motor already have a fuse in the supply?

Typically, the windflow when driving at a speed of 40 km/h starts to cool sufficiently the engine. Of course there would be circumstances of warm and or hot weather where the cooling by drivewind only starts at 60 maybe even 70 km/h. So if I could operate a potentiometer I could tune the switch of when I see on my temperature dial that the engine starts to run too hot.
The fan motor is fused by 30Amp fuse.

I'll plug that info into my circuit design.

Here's a suggested circuit, with a pot to adjust the cutout speed. Almost any inductive sensor (automotive type) should work with this.
U1 is a quad Schmitt trigger NAND gate IC. Each sensor pulse triggers a monostable circuit comprising U1a,U1b. The monostable output pulse width is set by C3 and the pot. The output pulses are integrated by R5,C4 and the result, if high enough, trips U1c and U1d to turn on Q2 and energise the relay. D5,C5,R6 are to suppress motor switch-off voltage spikes and may need to be tweaked.

Q1 could be the open-collector output of a Hall sensor if that were your sensor choice (R3 and D3 would then be omitted).
If you want a manual over-ride for this circuit, a SP3T On-Off-On switch could be connected to Q2 to provide On, Auto and Off functions.
I've attached the LTspice simulation file.

Hi Alec_T : Many many thanks for this rapid delivery of a functional circuit for my project !!! I will go trough it tomorrow morning.
Question 1 : how can I open an .asc file ? Shall I download a program related to electronic design ?
Question 2 : what is the meaning of Rser=10m
One million thnks, I will buy the components soonest possible. Next Monday I will start to fabricate 6 metal triggerbrackets which I have to mount on the shaft fixing to the gearbox.

Q1: LTspiceXV11 simulator (a free download from Analog Devices, formerly Linear Technology).
Q2: It's the simulated internal resistance of the 12V supply battery.
The circuit would be quite compact if built on stripboard, with heavier gauge conductors for the motor switching contacts of the relay. The relay will need to have contacts rated somewhat higher than the DC current the fan motor draws.
D5 is a Schottky diode which needs to have a current rating approaching the normal motor current. The suggested one is 10A rated, but can take the occasional peak of 30A or more.
Be aware that the IC is static-sensitive.

Good Morning Alec_T : I have downloaded the software and entered the file you sent me. I noticed about 3 or 4 components were not shown (IC, potmeter, relais...) so I could not perform the simulation. Can you check the file you have sent me ?
I see btw that 40 km/h means 36 Hz pulses. How high can the potmeter drive the pulses ? Thnks
The solution you gave me seems fabulous.

Edce said:

I noticed about 3 or 4 components were not shown (IC, potmeter, relais...)

Click to expand...

Those models are third party ones which don't come with the standard LTspice download but are available (along with many others) free from the Yahoo LTspice User Group. It's worth registering with the Group if you're interested in simulating circuits. Apart from the models there is Help available there. If you can't find the models there then give us a shout and I'll zip some up for you.

Edce said:

How high can the potmeter drive the pulses ?

Click to expand...

It should allow for pulse rates at least double that 36Hz and somewhat below that. If you need a wider range then C3 and R5 could be changed.
BTW, if your sensor output is not high enough then R3 could be reduced in value to, say, 10k (it's not critical).

To save you time, here's a zip of the files I think you need.
Unzip them and put the .lib and .sub files in your .....Documents/LTspiceXV11/lib/sub folder. Put the .asy files in your ....Documents/.....lib/sym folder (or suitable sub-folder). Alternatively you can use LTspice's Control Panel to add paths to tell LTspice where to find the files.

I checked the Yahoo group but will take me much time to find the right models.
I was thinking about a kind of delay for trafic jam driving or city traffic where the cars drive between 40 and 60 km/h. Because of the quick changes in speed, the fan would start and stop frequently. A delay of 60 seconds before shut off would overcome the quick switching. Is that feasable ? Many thanks again.

Edce said:

Is that feasable ?

Click to expand...

Not without a major change to the circuit. But the manual over-ride would be simple: just an added switch.

I must be missing something in this? Any of the electric fan controllers I'm aware of work on the temperature of the radiator. Doing this by speed is going to be a fail looking to happen, in my opinion.

If the speed is enough to keep the fluid cool the fan shouldn't come on, that is the purpose of the controller. What happens if you forget to over ride the speed controlled version if stuck in traffic jam?

Here is one company that makes them - https://www.flex-a-lite.com/accessories/electric-fan-controllers.html

Here is one circuit to do it yourself - https://www.maximintegrated.com/en/app-notes/index.mvp/id/1125

My aircooled engine has no fluid controlled by a fan. A modern car uses water (with antifreeze) to coil down the engine oil. Water can be cooled down rapidly by a fan. Oil to the contrary has a higher temperature content then water. To cool down an aircooled engine, air has to be brought directly to the parts to be cooled down, that are the cylinder heads. So its not the oil directly which has to be cooled.
Cooling the heads in my case is done by the volume of drivewind and when not sufficient by the fan. So low speed driving is more of a problem, which I will solve with you well designed circuit.
A fan speed controller still would need to go with a speed dependent circuit.
OK about using the potmeter to change the setting in relation to citytraffic, or a switch for the same purpose. So no need to do a drastic change to the circuitry.

On running the sim again I found the hysteresis inherent in the CD4093 was excessive (the motor would switch on but not off unless the road speed was very high), so I've revised the circuit. The IC is now a CD4011B and U1c,d have been reconfigured as a Schmitt trigger with adjustable hysteresis. The SetSpeed pot should now allow the motor to turn on over a range of 30-67 kph. The SetHysteresis pot gives a 12 - 24 kph hysteresis. This means, near the minimum settings of both pots for example, the motor should turn on at 30kph and turn off at 42 kph.
Given the hysteresis, I don't see why any 60s timer would be needed.

Note: C2 has been increased to 100n. Q2 buffers the voltage on C4.

Oh yes, so much better !!! Also the addition of the override switch is a great help . Thanks. The introducing of the missing symbols failed. Its the first time I see a program like this, so I need some learning time to grab it. But I believe the outcome of your simulation, you'r the experienced one. The hysteresis-setting is indeed a better solution then a time delay. I now even can set to fan to start at 30 km/h and by another setting to stop at 90 km/h, super !