DIY automotive cooling fan controller

[3v0]

The commercial controllers are simple electronic thermostat type things with a sensor stuck between the fins of the radiator.

I would like to build one that can also monitor engine and trany temp with an LCD display. The computer part is in my ballpark but I would like some advice regarding the radiator sensors. One chimerical controller uses 185°F On/170°F Off for 180°F thermostat. I am thinking a sensor good to 200°F would work.

Most of the 16 inch fans I have looked at state that they draw 20 AMPS with a few as high as 30 AMPS. I think I will try to get ones int the 12 to 20 AMP range. Any suggestion on what to use to switch the power. MOSFETS ?

 

[jpanhalt]

Based on this chart, I would consider going a little higher than 200° F (say 225 to 250). The rationale being that when it got too hot I would like to know whether it was an emergency needing me to stop immediately or something that I could drive several miles more like that.

http://www.txchange.com/heatchrt.htm

Mosfets at 20 A sound great to me. The losses will be less than IGBT given Rds of less than 0.1 ohm or so.
John

 

[stevez]

Do you want the fan to switch on and off or do you want to vary the speed of the fan as needed to hold the desired temperature? I've seen others who did the same thing using PWM and MOSFETs to vary the fan speed.

Any sensor should be capable of tolerating the temperatures that you might see in an overheat condition. It doesn't need to be accurate past your control point but it shouldn't change or fail if the engine does overheat.

 

[ClydeCrashKop]

Some engines run at 210 degrees and expect the pressure to keep them from boiling. It is for emission control and efficiency but I don’t like it because you are that much closer to a problem.

 

[stevez]

The pressure combined with coolant that boils at much higher temp than water is what allows engine to operate for short periods well above 200.

 

[3v0]

The DS18S20 and DS18B20 have a range of -55°C to +125°C (-67°F to +257°F). I use these for other apps and happen to have a few on hand. I did not see any info on using 1 wire for automotive use. It may work. I could cobble up a little system and stick a sensor in the car radiator and see if noise is a problem. Is there an easier way to find out?

I was thinking of using PWM to control the fans. It would work without but why not?

Do you have a suggestion as to part number for the MOSFET ? Where to start looking. Analog is greek to me.

 

[jpanhalt]

There are just 14000 at Digikey. Got the list down to only 5 pages with a few obvious filters.

I have experience with IRFZ44 (110W,48A), IRF1010E (200W, 84A), and IRF1405 (200W, 131A). I would suggest over spec'ing them as you may have fewer problems. Thus, I would go for 60Vds, 20Vgs(most common) and something about 50A or more. I have used the IRF 1010E most and have had good luck with it. You may not even need to heat sink it or you may get away with just a small one. The 1405 is most expensive and may be overkill.

I suspect there will be almost as many opinions as there are chips available.

One question: How old is the vehicle? Is there a chance it is a BLDC motor. That might change the choices a bit. My uses have all been with brushed DC.

John

Edit: Come to think of it, I have no idea what type of motor is on my car. It may very well be BLDC from the way it freely spins.

 

[3v0]

Thanks on the info on the MOSFETS. With 14K types I could pick what I wanted in terms of AMPS and still have a 1000 or so to choose from. I will order a few of both types.

I am rebuilding a 1978 Ford F150 truck (not 4x4).
Currently I am assembling the engine.
As of yet there are no fans.

 

[3v0]

Progress Report

As a proof of concept I have setup a PIC controler with an single line LCD to display the temperature reading from a DS1820. The controler has a 100 mA low dropout regulator LM2931Z-5.0. The datasheet says it is idea for automotive use.

The DS1820 will be powered with a wire from the processor regulator rather then using parsite power. I plan to twist the 3 wires using an electric drill.

As soon as I find a box and the time I will epoxy a DS1820 to the engine block and do some testing.

 

[jpanhalt]

Sounds interesting. A lot of hardware store epoxies have quite low softening temps of 135°F or so. JB-Weld is supposed to be better in that regard and for attaching to metal, like engine blocks. Some people even claim to have repaired crankcase cracks with it; although, I don't have that experience firsthand. It's metal filled so it might be something to consider. John

 

[jessey_]

Hello, I'm interested in your thermostat project

Hi 3v0,

I'm about to do the same project as you did and am wondering how you made out with your DS1820 ic and what compiler you used? I have already wrote my program using PicBasic Pro and have my circuit bread boarded and it seems to be working good. I was wondering what type of enclosure you used for your DS1820, I was thinking of epoxieing mine into copper tubing and soldering a cap the other end of the copper tubing and pushing one into the fins of my radiator and for the other one I'd like to place it into the engine block where the old thermostat was located. Tomorrow I'll be soldering the parts onto a circuit board, Im using a 40 pin 18F452 so later on I can add other features like a voltmeter etc...

I'd really be interested in how you made out and what if any problems you encountered...

Thanks
jessey

 

[3v0]

I used Microchips C18 compiler.

The prototype is still in the box. Which is OK because the truck engine is not finished either.

I have DB18B20+'s. They have a upper limit of o +257°F. They should be good enough.

Keep us updated,