Here's an update on my "switch project" - it may not be of much interest to most, but should at least give you an indication that I haven't abandoned this project (I do that sometime when I get too busy, frustrated, lazy, or a sprinkling of all). I got the high amp kit and after I put it together, I decided that I didn't like it, so I decided to design my own, but I didn't want to "toast" a Spal unit by attaching my DIY solid state switch to drive a Taurus fan, so I decided to to do some testing of my whole concept without the Spal included. To do this, I decided that I needed a Pulse Width Modulation unit, so I built one that's capable about 12 volts + and about 50 amps. I finished this unit last night but have only tested the output on my oscilloscope and that looks good - I'll test it out with a fan this week sometime. Today, I did some experimenting with my switch and got a working prototype - I need to fabricate some mounting and cooling parts for it so that I don't end up melting my "bread board" when I finally run high current through it. While I'm fabricating the mounting, I'll add a cheap 555 timer circuit to my prototype to create some variable square waves (DC pulses) to drive the gate on the MOSFET in the switch and make sure this works, then I can finish the switch, test the PWM unit, combine them and test the combination & the get around to including the Spal controller.

 

Cool!

When I mounted the Taurus fan into my Jeep I found that with a bit of trimming it fit quite nicely into the factory shroud and I attached the two with some self tapping screws. Almost looks stock. Might want to pull your shroud and see if something similar can be done.

 

I've installed them in a YJ, and they do fit very close. Since the mounts points for the Taurus and Jeep shrouds are so close, I found it fairly easy to make an adapter out of sheet ABS, it's easy to cut and I used ABS solvent and screws to pull it all together. I need to have a close look at the Dakota to see if I can do something similar & if I have enough 1/2" and 1/8" ABS pieces left from my last project. I didn't go with the shroud-to-shroud approach for a couple of reasons, but mainly due to the fact that the jeep shroud was broken and screwing another shroud onto it didn't look all that stable.

 

I had an opportunity today to make some additional progress on my making and testing. I added a simple, adjustable pulse train generator to my switch prototype to ensure the whole arrangement would make a higher amp duplicate of a low amp pulse train. It works. My next step with this switch will be to make it into a unit that I can use with a real fan, so I’ll need to get the final parts list and solder them onto a PCB along with a substantial heat sink for the FET.

In addition, I was able to do some preliminary testing with a real fan. I was able to scrounge an HHR unit and to make sure it worked, I hooked it up to an adjustable power supply. It appears to work well and draws approximately 18.4 amps continuous when running, so for those that have been interested in this number, there it is. My next step was to attach the Pulse Width Modulator that I had built & tested earlier, and see how the fan reacted to that – this PWM unit has manual (a couple of *****) frequency and duty-cycle adjustments and is set up to create a pulse train anywhere from about 100 Hz to 1KHz – it doesn’t seem like much, but is plenty to move these motors. I started my tests with the PWM set at 430 Hz (it was an arbitrary selection) and 50% duty cycle and then stepped up the duty cycle and noted the amps used. I didn’t have anything to test the rpm, so all I can say is “the lower the duty cycle, the slower the rpm, up to full rpm at 100%” (I also didn’t check the airflow – I have no way of doing this other than with a number of equations & I don’t feel like beating my head against a wall). Next, I increased the frequency about 50 Hz at a time and for any 1 duty cycle setting, the current requirement was less than it was for the same setting at a lower frequency (I’ll show some examples). I also found that at close to 600 Hz that the motor started to make a high pitched hum, so I backed it off to 550 Hz where it worked well. I finished off this bit of testing at 100 Hz. At this frequency and 50% duty cycle, you could hear the motor laboring to start, and it also used substantially more current than at any other setting (this wasn’t surprising). Here are the details for anyone interested.

For 430 Hz:
50% = 5.6 amps
60% = 6.9
70% = 8.3
80% = 10.9
90% = 14.2
100% = 18.4 (obviously)

For the 550 Hz test, there was an average of about a half amp saving (except for full on) at any one setting (greater at lower duty cycle settings). As an example, at 50% it used 4.9 amps and at 80% it was 10.6. I suspect that the rpm response would have been best here as well with a more constant series of “bumps” to the motor.

The 100 Hz test didn’t provide any real surprises. That frequency is just too low to keep any large amount of the magnetic flux established and additional energy was needed at each pulse to re-establish the field instead of rotating the motor. At this frequency, the 50% duty cycle used 9.2 amps and 14.2 at 80%.

I ordered some different parts for my final switch since I found that the variety of MOSFET I was planning to use was a real pain in the back side to get it to turn off, so to keep things simple (I really like simplicity), I decided to go with a different FET type – I got a close equivalent today just to confirm my thoughts and got the switch to work properly with very few parts, but this FET won’t handle the Taurus fan. Next week I should have the new parts and they’ll give me the capability of over a 100 amps continuous. In preparation for their arrival, I’ll be making the switch housing that will be able to keep the FET cool without a fan, and to make it easy to connect some 10 gauge wires to a “transistor”. If I have a break this weekend, I might be able to lash up the Spal unit with my current switch and attach that to the HHR fan. I’ll post the results and probably a few pictures which should make reading this stuff a bit more interesting.

 

Well, my efforts towards extending the Spal use might turn out to be a wasted effort. I was reading the Spal manual this morning and noticed something I had missed earlier - the unit might not operate with a load of less than 4 amps. My switch will give almost no load, so if the Spal doesn't like that, it won't turn on. To overcome this, I'd need to add a dummy 4 or 5 amp load and this is starting to add up in terms of cost for the Spal specific high amperage switch and wouldn't be worth it. I'll continue my project to completion to see for sure - the reason I started this task is that I know someone that has a Spal unit that wants to run a Taurus fan, but he may also need to change to a relay based system or go with the Painless controller & sell his current unit to someone that's happy with fans that use 30 amps or less (a lot of them do, and provide adequate cooling to our trucks).

Anyway, I'll post my results once I get assembled with it's final parts - I'm sure I'll find another use for it if it's not needed here, and the PWM unit was "cool" to build and play around with - the switch will certainly work with my unit & I may just expand the project and add a micro-controller to it at some point to "smarten it up" and make it useful as an automatic fan controller.