How to drive an electric fan?

[Fred]

About the FET: KISS. Don't worry about a diode at first, the FET has a built in clamp to ground anyway ;-) Just put the part that I recommended on a decent heat sink and see how you go without anything else. Driver box inline with ground cable, relay in line with power cable. Drive the gate with between 100ohm and 1k, the data sheet may have a recommendation for min/max. I'd probably put a line driver chip and use 200 ohms. You could step it up and use a FET driver chip and 12V too, likely a good idea with that current and switching frequency, actually. IE, turns it on faster, off faster = less heat.

You should keep all the parts close to reduce noise, short wires from fan to power source and ground and routing that takes them far from any critical engine sensors.

Page 9 shows a test circuit with a voltage meter across it, not for actual use :-)

Regards "one sensor to rule them all": Pretty much every car from the beginning of EFI till recently has (at least) 2 sensors for temperature. The gauge one is/was typically 100 ohms low accuracy coarse high internal self heating; not suitable for EFI. The EFI one is higher precision higher resistance low self heating to allow it to be accurate for the computer. You can't use the EFI style to drive a gauge directly and you wouldn't want to try to use a gauge style one to control the engine. If you want to drive the gauge directly, that's cool, but make sure the precision to the uC is good, not from a gauge sensor.

Finally, why design a special board/circuit/code for this? This is the sort of stuff that FreeEMS can/will do with its eyes closed while having a rest with one hand tied behind its back with your engine at 7k ;-) If you're going to feed an analogue signal back to FreeEMS from some other box make sure it's high precision.

Fred.

[BenFenner]

Peter, while I'm not great with EE, I do know the basic electrical components and how they work. I'm certainly knowledgeable on relay versus transistor. It would take paragraphs to explain what I was getting at, and that would be maybe pointless. Fred has answered my concern so I am happy.

[Fred]

You have to admit, Ben, you did make yourself look far more stupid than you really are :-) Trying to fool us all, no doubt. You can't blame Peter for having a crack at you really, if I didn't know you, I might have done the same :-)

Fred.

[Peter]

I do know the basic electrical components and how they work.

lol I have no idea how they work, they're pretty much magical little black boxes that run on funny smelling smoke as far as I'm concerned. I just try and understand what they do.

[oden]

relay in line with power cable

Relay? though the idea was not to use a relay.

I'd probably put a line driver chip and use 200 ohm

Really dont understand what you mean here

The gauge one is/was typically 100 ohms low accuracy coarse high internal self heating; not suitable for EFI

Aware of this, thanks for bring it up.

why design a special board/circuit/code for this?

My questions on fan control would still be relavant in the PWM signal was coming from FreeEMS, but I know thats not what you mean.

Last count there was 21 relays and my wiring loom is a mess. There are 5 relays to just for central locking, looking at it brings a tear to my eyes. To many after market add ons that are not designed to operate in conjunction (even within the same manufacturer). Building (in stages) a more elegant solution to replace some of the mess.

Dont think this sort off stuff should be jammed into FreeEMS.

[Fred]

The reason to have a relay on the power side (and a fuse!!) is in case something goes wrong with it, or the ground wire chafes before the FET box. You shouldn't have the fans live when the key is off (unless with a turbo timer) though you could, and it would likely be OK. I don't like the idea of large current carrying conductors live in random places with the key off, though.

It's exactly the same with the injectors and coils. You power the positive side through a relay (one for coils, one for injectors) and the negative side grounds through fets/ignitors so they're not powered up when the key isn't on.

With respect to a hundred relays in a big tangled mess, the idea of relays is to put them where they need to be, inline with the device between grounds and power, and run low current lighter gauge wiring for the control signals. So if all your relays are in one place, perhaps you're "doing it wrong" ? A bit off topic, but no matter :-)

PWM based on temperature is totally a good thing to have in FreeEMS, at first open loop, and later closed loop. Powering a cooling fan that way is an ideal use :-) It just takes a fairly serious driver box, that's all.

Fred.

[oden]

The reason to have a relay on the power side (and a fuse!!) is in case something goes wrong with it, or the ground wire chafes before the FET box. You shouldn't have the fans live when the key is off (unless with a turbo timer) though you could, and it would likely be OK. I don't like the idea of large current carrying conductors live in random places with the key off, though.

Now that makes alot of sense, thanks for explaining.

With respect to a hundred relays in a big tangled mess, the idea of relays is to put them where they need to be, inline with the device between grounds and power, and run low current lighter gauge wiring for the control signals. So if all your relays are in one place, perhaps you're "doing it wrong" ? A bit off topic, but no matter

Most of the relays are performing logic function not power regulations. For example if the alarm locks the door at the same time as you press the unlock button there is a direct short so I have a relay 'pack' to prevent this. Failed by not wiring up all the modules on a bench, hence there has been alot of surprises. Yes getting off topic..

PWM based on temperature is totally a good thing to have in FreeEMS

I will drive it now from my uC as its only a few lines of code and once the FreeEMS hardware is in place it can take over the task of generating the PWM signal. Curious to build something

Thanks alot for your help

[Fred]

You're most welcome!

I expect pics of your 35A omnifet and heatsink in this thread and if we're lucky a youtube video of you controlling the fan speed with a pot connected to your micro :-) Would be nice to prove out the driver. It should work, but I could be wrong. If it works well for you, I'll be copying for my stuff one day when I see my stuff again.

Fred.

[oden]

Will do, about to go on holidays so it will take a while

[oden]

Back from holiday

I got my hands on a few VNP35n07 MOSFET's as described and did some testing. It was hooked up in the following manner.

Pin 1: + 5.5 Volt generated by a MC7805 for testing
Pin 2: Fan's ground wire
Pin 3: Ground

As suggested I tied running a 100K then 200K resister between pin 1 and the 5V supply but the MOSFET would not power up.

As is the fan runs at 9.4 Amps draw (rated as 12 A). When the MOSFET is added the current draw drops to 8 Amps, understand its normal for a loss but is this to high? The fan speed is noticeably slower.