CarTalk.com Blogs Car Info Our Show Deals Mechanics Files Vehicle Donation

Subaru A/C fan won't turn on

Hitting a diagnostic dead end here - car wants to overheat/drink coolant on hot days when inevitably the A/C is on.
A/C fan never comes on. Car was in a front-ender accident about 18 months ago, could be related, unclear. The A/C condensor was damaged then, but I replaced it, recharged the system, and the A/C now blows nice and cold.
I’ve swapped the fans and the same fan works fine when plugged into the other side.
I’ve swapped the fuses.
I’ve taken a working relay (as the other fan stops when I pull the relay out) and rotated that relay thru the other 3 sockets with the engine and A/C on. Still no movement.
So I’m a bit stumped puzzled as to what to consider next, and then a part of me started to wonder if the fan is the red herring, and might I have a radiator issue which is the real culprit. I don’t see any leaks. I had the thermostat fail about 1-2 months after the accident but a shop replaced it and this overheating issue really only flared this summer.
This is a 99 Subaru Legacy SUS Sedan, the 30th Anniversary edition, with 189K+ miles, so minimal expenditure is the game plan.
Thanks for any ideas.

You need to get an ohm meter and check all the wiring for continuity. The fan needs to run.

If the fan is not running, check to see if voltage is going to the fan, if you have voltage, then bench test the fan.

You can find a schematic to your car here:
http://www.powerlibrary.net/Remote/Remote.asp?ID=Monroe
To get in, enter barcode 02378001078755

It is a given that you’ve checked the 20A fuse and the fan is good.

Your AC fan has two relays controlling it, MainRelay 1 and 2. One turns it on, the other turns on high speed. I’m not sure if relay1 needs to be active for the high speed to turn on - I suspect so. It may be a good idea to pull both relays and short both contacts to see what happens. If it doesn’t turn on, the problem is past the relays, else before the relays.
If it doesn’t turn on, maybe what they have in common - the black ground wire on the connector - isn’t making good contact to the chassis.
There are more thing to try but that^ is good start.

Oh, and definitely don’t drive it with AC or defrost on. Overheating will kill that engine quick.

It sounds like you have done a lot of parts replacement but you don’t mention doing any voltage checks to the fan circuit. Using the diagram that RemcoW provided and a test light probe you should be able to find where the wiring trouble is fairly quickly. You should be able to test the fans by plugging the green diagnostic connections together under the dash. The fans should pulse on and off while they are plugged together. Other tests happen also when the connectors are together.

Thanks for all the feedback and sorry for the slow follow-up from my side.
I’ve never done anything electrical aside from bulbs, so I’m at kindergarden level on this front.
I have a Harbor Freight multimeter which I guess if I’m reading it right gives me the Ohm reading on the relays. The Haynes book says poles 1 & 3 should be 100 Ohm and 2 & 4 should be infinite. I seem to get closer to having 2 & 4 be 100 Ohm and 1 & 3 be infinite (I think I’m reading it correctly as far as the Top being the engine/inside side.
My tester says:
Relay 1: 2&4: around 100, 1&3: infinite
Relay 2: 2&4 around 100, 1&3: some low number, close to zero
Relay 3: 2&4 around 100, 1&3 infinite
Relay 4: 2&4 infinite, 1&3 around 20.

I plugged in the 2 green connections under the dash. I think I saw the regular fan give a slight turn (once I started the engine) but the A/C fan again didn’t budge.

I disconnected the green connections and unplugged the A/C fan and spun the multimeter around to the various voltage options, but couldn’t get any readings from the side that leads back to the relay box. Wasn’t sure if I needed to have the engine running and A/C on to expect to see any voltage there, but figured I’d stop here as my odd relay readings might be the more insightful data.

I’m not sure how to short the relays as RemcoW suggests.
I did manage to get that fan work previously by plugging into the plug for the other fan, so I believe the fan itself (motor, wiring to plug) shouldn’t be the problem.

Thanks again.

The connection that are infinite are the contacts. The corresponding connections in the socket for the main fan are ones that, when the socket’s contacts are shorted, will cause the fan to turn on. That’s basically what the relay does but under control of another voltage.
Since by your admission you’re at kindergarten level as far as electrical stuff is concerned, I don’t want you to hurt things or get hurt so do this instead:

Turn the multimeter on read resistance or continuity. It is the mode that, when the two test leads are shorted, the display goes from 1999 to near zero. The multimeter may even have a mode that buzzes or beeps when you hold those leads together. Experiment with that meter so you understand how to measure continuity before you crawl under the car.
Use that mode and measure between one of the connections of the connector on the body of the car to the fan that isn’t working and the negative terminal of the battery or any bare metal part on the engine. You don’t have to turn the car on. Just unplug the fan and measure on the connector it plugs into.

One of the connections to the fan is supposed to read near zero continuity and/or beep. In other words, one of the connections to the fan is supposed to be attached to ground.
You can probably figure out which one is supposed to be grounded by looking at the connector to the fan that is working and measure between one of its three connections and ground (ie negative battery terminal or bare piece of metal on the engine).

If you don’t see a connection to ground on one of the terminals to the inoperable fan, trace it back and see where it is broken. That connection could be at issue.

When you look at the body of the relay, you should see a schematic on it. It will have the symbol for a coil between two of the posts. If the drawing on the relay body shows the coil between posts 2 and 4, then that is where you should be seeing the 100 ohms. If the Haynes manual is showing the coil between posts 1 & 3 but the relay is showing the coil between 2 & 4, then either you have the wrong relay or the Haynes manual is wrong.

The relay will show the switch between the other two posts and the switch is normally open, therefore it will read infinite with the ohm meter. Once you determine exactly which posts are the switch terminals, then you take a piece of wire and short out those two terminals on the relay socket. This is what RemcoW meant. Make absolutely sure that you have the right terminals. If the fan starts right up, then either the relay was bad or it is not getting the 12 vdc to the coil.

To check for the latter, set your VOM to the DC volts position, put the black lead on ground, then test each of the coil terminals in the relay socket. One of them should have 12 vdc on it, the other zero. If you get the 12 vdc on one terminal in the socket, then put the meter back into ohms and test for continuity between the other terminal (the one that read zero) and ground.

If the fan did not start running, then check the switch terminals on the relay socket for 12 vdc. One of them should have the 12 vdc, if neither has, then there is a wiring problem from the source to the socket. If one has the 12 vdc, then check the other terminal to ground. If you get infinite resistance, then there is a problem between the relay socket and the motor, or the motor is bad.

My experience with this issue is that the brushes hang up in their tracks inside the motor. They will work intermittently sometimes and that gets people to think the problem is elsewhere. A new motor may be in your future.

Btw, those relays are all the same part number?
If so, the should measure roughly the same…

re: 446pm post. I couldn’t manage to disconnect the electrical connectors, try as I may for a silly amount of time, using a screwdriver, pliers, vice grips, etc. When I used the multimeter to try to reach the connection on the relay side from the other side of the connector (on the inside). I seemed to find continuity in 2 of the lines and a zero/ground on the 3d black wire, but whether I was really doing that right is hard for me to say, to be honest. Seemed a challenge to replicate.

re: Keith’s comments - just so I’m not missing something, the relays are the cartridge type things that plug into the box, right (you’re using the word socket, which leads me to think I’m using the right words and following you…)? I took them (relays) out and looked around everywhere. I didn’t see any icons/logos in the relays or the plug in areas (sockets). They are all 4 the same, factory original I think, DENSO 056700-8480 12V Made in Japan. Inside the relays it’s noticeable that 2 of the prongs are more copper colored (vertically) and the other 2 more yellow-ish.

I rechecked all my readings with my admittedly crappy (it’s the $8 model) HF multimeter. I got the same #s I recorded before, following the Haynes logic that 2&4 are left side.

The Haynes Manual is admittedly 90 thru 98 Legacy, and this is the 99 special edition. It’s been my experience with 3 different Subarus that things don’t change dramatically, so I’ve tended to use this one, although it’s true that the relays/layout are different from my 98 Impreza.

I guess by looking at the schematic I pulled down from the web link I could try to verify which side is 1-3 and which is 2-4 if I could match the colors mentioned in the schematic? When I started to take the box out earlier it seemed like the wires were hard to access/see from the bottom.

I should be able to check for 12vdc w/o figuring that out though, and see what I get. I’ll try that tomorrow. When you reference motor you mean the starter motor? I haven’t tried to scrutinize how the fuse/relay box is connected into power, not having had problems in this area before.
Thanks.

In following up on Keith’s suggestions, I could find the 12 reading one on of the 4 poles in sockets #1 and #3, but got nothing but zeros from all 4 poles in sockets #2 and #4.

In addition, as I retested for continuity, in sockets #2 & #3 I’m now getting an infinite reading on poles 2&4 where I was previously getting 100. I’m not sure what has changed, unless I somehow damaged some of the wiring in my efforts to disconnect the fan connection at the base of the radiator/fan under the car.

?
I don’t have a sense as to what one would try to replace even in steering towards potential problems here. Would the approach be to replace some of the wires from underneath the socket in the relay box and re-run them to the fan, whether that’s by way of taking off the fender or rigging it in a way that the wires are out of the way of other moving parts?

Maybe you just made a simple mistake with the relay connections.
Do you have continuity on that one inoperable connector to ground?

When I set to Ohms (my mm doesn’t have the light/beep, just a digital readout) I can find a zero post on sockets 2 & 4 with the black cable touching the car’s frame in the area where the battery is grounded.
Specifically:
Socket 1: close to zero, 100, and 2 infinites
2: 30, infinite, zero, zero
3: 30 infinite, infinite, 100
4: 30, infinite, 30, zero

You do not use the Ohm scale on the socket terminals if you don’t have a schematic. Two of those terminals have 12 vdc on them, though normally only when the ignition switch is on and the ac switch is on. 12 vdc applied to your meter in the ohms scale can damage the meter. In some vehicles, 12 vdc is on one of the contact terminals all the time, but with the contacts open, there is no current flow.

To close this out, I think I’m good now with a diagnosis of a bad relay in socket #3. Couldn’t have figured it out without the help here though, so many thanks. After parsing this very carefully, I figured out enough to jump 2 of the poles in that socket and got the fan to turn. My local parts store didn’t have the relays so I had to order them online (ebay, 2 for $13, used). What confused me greatly was I had rotated the other relays around into the AC socket, figuring if the relay was bad, the A/C wouldn’t turn on. But with that test, I couldn’t identify a bad relay. But I followed Keith’s advice and measured 12 vdc and then zero in the right places, so I figured I’d try a new/different relay and it seemingly has worked.
My diagnosis was a bit further confused in that the connection between the fan motor and the wiring down under the car got a bit strained as I was heavily struggling and ultimately failing to disconnect the connectors last time. I realized that even with the new relay in socket 3 I still wasn’t working although when I started the car/AC it would turn for 5 seconds or so then stop. I also saw that jumping the other relevant socket (#1) with a piece of wire I couldn’t get the fan to turn. So with the engine & A/C running I started wiggling wires, and in wiggling the ones under the radiator for the aforementioned connectors, I realized I had a loose connection there. I pulled the wire out of the connector and tried to reinsert it as straight and deep as possible, so hopefully that holds. Thanks again everyone for the help and advice. Long live Car Talk!

There are a lot of good books you can get to study up on basic DC circuit testing for automobiles. You would be wise to get one and learn the basics. It can be fun to learn and it will be a big help in doing that kind of thing in the future. There is a lot free stuff on the net also you can get.

Glad you got to the bottom of it. Thanks for the report.