[canderson]

Will assume for the moment that you are correct and that it really is the radiator (cooling) fan and not the condenser fan (which to avoid fogging, can also run at this time of year without A/C on if you've got the front defrost going). Fan is controlled by a relay which is driven by the PCM. There's some chance that the relay is sticking. It's in the box, under the hood, on the driver's side. It's unlikely, but so is your situation. One other possibility - that the wire that the PCM uses to generate the ground to fire the relay is somehow exposed and shorted to the frame (or any ground) somewhere. That would be somewhere between the PCM (which is back near the relay box under the hood) and the relay box. Also not very likely - it's a short run and there shouldn't be any mechanical issues to damage the wiring. It's in a fair bundle anyway.

Edit: For what it's worth - if it WAS a condenser fan and not the cooling fan, all of the above still applies identically apart from which relay is involved.

 

[canderson]

Sounds like the CAA guy has a decent head on his shoulders. For the price, and since this is an intermittent problem (total bugger to diagnose when it's not misbehaving), I'd be inclined to ditch the relay for a new one as a preventive measure -- as an attempt to reduce the odds of getting stuck with a dead battery again. While being able to ring up CAA is nice, you never know where she'll get stranded, or how long it will be until someone shows up.

 

[canderson]

Hard to say - wasn't there to hear it!

It's an interesting lash-up. There are (as you note) three different relays operating the two fans up there. It's even possible to run the two fans in series, reducing the speed of both of them. Knowing which had been stuck requires a good ear or observation of which of the fans is turning. That's why I wish you were sure about that.

One relay controls only the +12 to the radiator fan.
One of the condenser relays takes the ground from the radiator fan and either sends it straight to ground, operating the radiator fan at high speed, or through the condenser fan (causing both to operate a low speed).
The other condenser relay can operate the condenser fan at full speed by sending +12 directly to it. So it's all a bit of a mess sorting which is the culprit until you know what the fans were doing when this problem occurred. They run about $11 to $18 each (ouch). You'd be out around $30 to replace the two condenser fan relays.

 

[canderson]

If you pulled the 50A CON/FAN2 fuse and the fan shut down immediately, understand that this fuse supplies power through the corresponding CON/FAN2 high speed condenser fan relay. Either the relay is sticking or you've got a ground on the PCM side of the relay coil that is enabling the relay. Here's hoping it's the former. It's also in the box under the hood.

 

[canderson]

Pulling the relay is the same as pulling the fuse. No surprise there, as you say. Fan shuts off either way.
However, if he pulls it and gives it a good rap and re-installs it, and the fan remains off, then 99% odds it's a relay issue. Easy enough to replace in any case.

 

[canderson]

I fully understand the difference.

Since this is such an intermittent issue, using sufficient force to pull the relay may in itself be enough to knock the contacts loose (these relays can be pretty tightly held), so measuring them after removal may not provide good information about the condition of the contacts.

However, if tested while the problem is happening, and giving the relay a sharp rap after removal (to give the contacts a chance to unstick IF they're still stuck), reinstalling it, and discovering that the fan remains off after reinstall is a good indicator that the contacts were the issue. I suggest this first because it's an easy test, and doesn't even require the OP to locate a meter. Of course one could measure them as well, but see first paragraph. A measurement that indicates open contacts after removal isn't a guarantee that they were open before removal. If a good rap doesn't do it, then before replacing the relay...

... we're down to either a short to ground of the wire headed back to the PCM, or the PCM itself. Finding +12V between the contact pins in the coil socket when the fan should be off would occur in either of those cases. The former will likely be a pain in the butt to locate, and the latter is expensive.

If there's no 12V between the coil pins, then the relay was the problem, even if giving it a whack didn't loosen it up.

For the OP's sake, I'm hoping it's just the relay.