fan controller issues

I am having fan controller problems and wonder if someone can offer some insight to a possible solution. These are very sophisticated controllers, but I'm going through them like crazy. Their engineers claim they are dying due to excessive heat and noise on the lines. Can't do much about the noise...but I have suspected all along their controllers are not up to the 257 degree underhood temp standard. So I'm moving it into the cab. Their engineers believe this should fix the problem. We'll see. But I want to do more. These are rated at 30A and I'm running a fan that draws 18A. I fuse it for 25A...and it never blows the fuse until the controller outright fails. The controller takes power and sends it to the fan, then takes the return from the fan and uses PWM on the ground side to control the fan speed and ramp up the fan on initial turn on. What if I were to supply power to the fan via a relay, and use the power output side of the controller to trigger the relay instead of powering the fan? Then run the ground back to the controller and let it use it's Pulse Width Modulation circuitry to do it's thing? Their people say I haven't done a thing since the return is still going to the controller and subjecting it to the power needs of the fan. I say that is somewhat true, but the power the controller sees will not be as severe as before as it's only seeing the down side of the fan. If this were not true, there would be no benefit to down side (negative side) switching...which ALL OEM's incorporate a ton. Any thoughts?

They are correct.

I would assume, since it simply passes the positive power through the module that they are not subjecting the positive voltage to any failure points. 

The (-) side is where all the magic is at.  I would assume they are using a fairly large FET to control the PWM and most likely that is what is failing.  However, without knowing the exact failure mode it would be impossible to deterimine what is causing the issue.

The two main possibilities I see are an output failure - the transistor is either shorting or opening from too much stress - or a logic failure - the main processor (or even voltage regulator) is failing.  If you can find out exactly how it is failing then maybe we can determine what is causing it.

-------------
Kevin Pierson

KPierson wrote:

---

I would assume, since it simply passes the positive power through the module that they are not subjecting the positive voltage to any failure points

---

To me, that's a pretty big assumption. That's why I'm thinking the relay might help. I also suspect...like you...it's a transistor that's failing. You know, Painless is coming out with a fan controller in the next 30 days or so that is really going to be something. One thing they told me in initial testing was when they put the controller in an oven and heated it up, it failed miserably. They didn't realize how robust they had to make it to survive the underhood temp standard. I have a feeling heat is killing this one too. I don't see the problems at all until we see the summer heat of Arizona. Then they fail quite rapidly. And I mount them as far away from the engine as possible.

By the way, when they fail, the controller shorts the main power lead to ground the moment it tries to turn on the fan. Same failure every single time.

Where would the failure point be?  I don't see how adding a relay to a relay output would help at all.  If anything, run a dedicated 12vdc fused line directly from the battery to the controller and leave the fan + supply on the fan relay like it is now.

-------------
Kevin Pierson

I agree with KP (in both his replies).

Below is what I replied a some hours before I lost webland.
KP's as usual is somewhat more concise and only my last 2 lines add anything substantial(?), but FTR.....


I don't think it's a big assumption at all.
The -ve side is PWM modulated; the +ve side would merely be +ve power to power the unit and (presumably) on to the fans - there would be no +ve conditioning etc (other than perhaps a fuse).

The only thing might be if they don't trust the -ve MOSFETs (or transistors - yuk!) they might have a +ve relay (or MOSFET - but I doubt that).
Otherwise you can bypass the heavy power to the fans - ie, a fuse direct from the battery to the fan +12V.
The controller itself should consume little power.

At that temp (125C), many semiconductors would be stressed and it is likely that heatsinking is inadequate - though that would be less of a problem for good MOSFETS (very low on resistance).

And chance it could be moved to inside the cabin?   

They should not fail due to noise.

Interesting comments from both of you. And I certainly appreciate each and every one.

The + heavy battery feed must be doing something in the controller. Otherwise, why would you bother with running the lead there? Why wouldn't you go directly to the fan with it? The controller has another + feed that is ignition controlled. The module retains all settings when disconnected from power, so it's not utilizing constant battery power for KAM.

Oldspark, there is quite a bit going on inside this module. And it does create quite a bit of heat that I don't believe is dissipated in an efficient manner. It's such a high quality piece that it's a shame it can't withstand the environment it's meant to. This may give some insight as to why electronics designed for use in this type of environment are so expensive. The components need to be very robust and the design such that it will survive. But when it works, it does so much so well. And it's programmable.

Thank you for all your thoughts and ideas. You have me thinking...and although that could be dangerous :), it's also a good thing.

I agree...the 'noise factor' is a bunch of hooey from an engineer that doesn't like admitting his controller won't live where it's advertised to live.

Yes...I am going to move it to the interior. Not only that, I'm going to either put a small fan on it, or drill a small duct from the a/c side of the climate control and cool it. That shouldn't be hard at all.

cobraguy wrote:

---

The + heavy battery feed must be doing something in the controller. Otherwise, why would you bother with running the lead there?

---

Maybe I should ask the opposite - what IS IT doing?


The ignition power obviously turns its circuit on, but I suspect it also provides the full power necessary for the controller. (Hence fused by smaller IGN fuses.)    

The heavy power lead may simply be for an in-line wiring layout (at the expense of "direct" efficiency perhaps), or maybe a memory function - though I doubt that these days.

Certainly if noise is a destructor, then I'd suspect immature or bad design. (And I've seen heaps of those - from the simplest to the most sophisticated - both often better replaced with cheaper simpler (and superior) methods or components).
With such designs may come unsuitably spec'd components. Ignitors are often mounted on the engine yet they survive (albeit at lower currents up to ~15A, but withstanding higher temps & IgCoil feedback - and many of those use transistors!).


But you should be able to the heavy power's function....
I'd expect a PICAXE circuit which PWMs MOSFETs. The only "heavy" current should be the grounded MOSFETS.   

But keep adding your thinking to mine/ours. I can be wrong.