Can the source of an N-channel FET be left open to inhibit switching?

Hi guys. I’m trying to work out a simple delay off timer that will keep the relay of an electric fan closed for 3 or 4-minutes after the ignition in my car is turned off. This is to let the fan run for a little while to move some hot air out of the engine bay. The thermo switch closes to ground when the high temp is reached, starting the fan. However, due to the build up of heat after the engine has stopped and the fairly wide differential of the thermo switch, the battery will be flat before it stops. So I have a simple FET timer circuit that just uses an R-C network on the gate to keep the FET switched for the required time when the ignition is turned off. My problem is, although we all know that the gate should never be left floating, what about the source? In a normal switch-to-ground set-up, the source is tied to ground. Can this ground connection be controlled directly by the thermo switch? With the ignition on, the FET would have a voltage at the gate all the time but it can’t switch until the source is grounded. The cable from the thermo switch to the source could be up to 3-feet long. Would there be problems with this in a vehicle environment?. I could just use the thermo switch to trigger a transistor and pull the source down but I’m trying to keep the parts count down and in any case, the question of a floating source still needs a response. Thanking you for any advice.

Why not just try it? I’m not an expert but fwiw my opinion is that the circuit will work. Without a series connection the FET certainly cannot conduct. When the connection is made by the thermo switch a reference will be established for the gate.

The automotive environment would worry me - hot, cold, wet, etc.

Is your car highly modified? Why the concern to get the temperature down with a fan?

I have had two different VW Passats - 95 and a 2001. These have auxiliary coolant pumps that circulate engine coolant after the ignition is shut off. I don’t remember what happened with the fan. This helped cool the engine a bit faster but I am not totally sure of the reason for it. not exactly making a point - your idea just reminded me of these cars.

Thank you mjkuwp94. I will try it but I am concerned about reliability. It may work but it may not be the “done thing”. I invite others to comment on this. My car is a fully restored 1966 Mustang GT. When the 289 V8s are rebored, they often run hot as there isn’t a lot of metal between the cylinders and the water jacket. My car had such an overheating problem and a 3,000 cfm electric fan has finally fixed it. However, as not everyone knows, an engine initially gets hotter when it is switched off, some actually creating an afterboil condition as coolant isn’t circulating. Again, I have this and the temperature gauge is off the scale to the right within a minute of switching off, then switching the ignition on again (to see the temp gauge reading). Letting the fan run on for a few minutes really helps to drop the temperature under the hood. It can’t be good for the fluids, hoses, electrics etc. to be exposed to such highly elevated temperatures with no air movement. A simple delay off timer is all that’s needed to keep the air blowing for a few minutes.

Any paticular reason why you want an IMP to do this? You could get away with a little attiny86 (if i remember the number correct), or it is even possible to do it all analog, like extending the time the dome light stays on.

Maybe I’m in the wrong forum. All I wanted was a comment on the wisdom of leaving the source of a FET open and controlling a load by switching it to ground. I thought the experts on IMP would find this easy. I just can’t find any reference on line to this question. Plenty about high-side and low-side switching, using pull up/down resistors on the gate, but nothing on leaving the source open, or is it not the done thing. Sure, I tried it and it works but is it likely to cause damage down the track or be an unreliable way of switching a load ?. The load by the way is a 12V automotive relay with a fly-wheel diode across it and it draws 140mA.

I’d personally put a huge (eg 1M) resistor across the thermo switch contacts. This ensures that source will be at a known voltage when floating (ground) and hence the Vgs doesn’t go negative (not that this is necessarily a problem, but “floating” is rarely good with FETs).

~12uA will flow through this resistor (and the fan) when the thermo switch is off and the FET is off.

As this is the real world, there’s likely several megs of leakage across the thermo switch anyway, especially if it’s humid there, so even without this resistor you’ll probably be fine :slight_smile:

Thank you Hugo. I took your advice and added the 1 meg resistor across the thermoswitch input. Using a 220uF low ESR electrolytic cap in parallel with a 1 meg resistor from gate to ground, I get an off delay of 4-1/2 minutes after the ignition is switched off if the thermoswitch is still closed. This is perfect for the application.