Pontiac Solstice Forum banner

Cooling Fan Failure - GXP (Variable Speed?)

71K views 105 replies 36 participants last post by  alpaca 
#1 ·
Hi all.

My cooling fan no longer comes on. I've browsed some posts that says the 2.0L GXP has a variable speed controller, though I can't really see much of the unit right now. I'm guessing either the fan motor died, or the controller died.

I was hoping somebody would have some troubleshooting guide or suggestions, or leads as to where I could get a replacement controller.

I should mention the car had the blown fuse issue (following a jump start) and overheated a couple months ago because I had stupidly walked away for a few minutes. The fuse has of course been replaced. After the fuse replacement, it seemed to be OK for a month or two, then I noticed the A/C was dying while stopped.

I was going to start disassembling Sunday (Feb 2nd) and see if I could get the motor to spin up, thereby conclusively determining whether it was the motor or the controller.
 
#2 · (Edited)
One thing you can check is to see if you have power at the post on the outside of the under hood fuse block. The post on the left looking down at the block, the black cover will have to be removed to access this post. That is the power source for the fan on a turbo car. The fans are different on the 2.4 NA they don't have this "extra" post on the fuse block. There is a large plugin on the fan housing on the passenger side (shown in first image). You should be able to access that with a little effort. The "controller" as you call it is mounted on the side of the fan housing and that is what the plug goes into.

Here are some images to help you. The fan does come as a complete assembly but you can remove the controller from the housing by removing one phillips screw and disconnecting the plug from the back side. This will allow you to test and see if you have power to the back side of the controller. I believe since the hot wire is constant that the fan is controlled by the ground. You could test the module by providing power and a ground to the unit with the source plug removed. Just make sure you distinguish the hot from the ground on the controller.
 

Attachments

#6 ·
Here are some images to help you. The fan does come as a complete assembly but you can remove the controller from the housing by removing one phillips screw and disconnecting the plug from the back side. This will allow you to test and see if you have power to the back side of the controller. I believe since the hot wire is constant that the fan is controlled by the ground. You could test the module by providing power and a ground to the unit with the source plug removed. Just make sure you distinguish the hot from the ground on the controller.
Not quite. The Red/Black wire is constant 12V, the Black wire is constant ground. The fan is controlled by a PCM logic signal to the controller (on my schematic it shows as a white wire, but it looks like it may be yellow in your photo - difficult to see).

Anyway, you can unplug the fan connector from the controller and supply power and ground to that (bypassing the controller) to verify if the fan works alone.

If the controller on the fan is bad I doubt you will be able to do anything to it as it is a sealed unit.
Is it just a sealed box (that you can break apart and re-glue)? Or are the innards epoxy encapsulated? Do you know?

If the latter, then I agree, there's nothing you can do.
 
#3 ·
Doesn't really matter if it's the fan or the controller since you cannot get them separately. You can only buy the whole module with both together.

If, however, the controller's the problem and you can get it apart and debug, maybe you could fix it?
 
#5 ·
My fan stopped last year and it turned out to be the controller. I found a post here where someone had replaced their fan due to failure. I got in touch with him, even though his fan had failed the controller was still good, and I was able to buy it from him for $60 and install it without removing my fan. I see in the classifieds where "rolla rage" has posted a radiator for sale. Reading his post, he also is selling the fan assembly. The post was from August or so of last year, but he still may have the fan assembly. Just thought I'd let you know.
 
#8 ·
Thanks everyone for commenting.

I will tear down tomorrow, and check it out. I'm a comp engineer by profession, so I'm guessing the PWM signal is probably used as the driver for a FET. If the PCB inside the controller box is not epoxy covered and the damage is apparent, I might try to repair it (who knows, it might just be a blown PCB trace or fuse.

I will post back on what I discover.
 
#10 ·
Well, not much conclusively yet... I was able to split the case using a razor blade on one side, and prying up with a flat head screwdriver around all the remaining edges. I thought I would post the images of the controller internals I discovered. The controller itself is solid state, as I had assumed. I was unable to find any obvious signs of damage, though only the bottom of the circuit board is visible because of the way the controller is built.

I was able to determine the fan assembly itself is not open circuit, therefore I can assume the fan itself is likely still OK. That is a good thing, since I was able to remove the controller without removing any other components.

Now I am wondering about the signal from the ECU. I have a USB oscilloscope I was going to try to get hooked up and take a peek at the white signal wire, but it might take an hour or two to get the oscilloscope drivers working with my laptop.

Does anyone have any details as to what the signal on the white PWM wire is supposed to be? 5V, 12V, Hz?





 
#12 · (Edited)
Fan Controller Signal Tracing

Well... Once I got the cover off, I was able to run a signal line from the PCB trace to my oscilloscope.

The first trace below shows a negative pulse from the ECU, which bottomed out at ~ -5V. This was done with the controller connected to the engine harness, but the fan disconnected from the controller This test immediately resulted in an ECU trouble code, P0480 (if I remember correctly). I then cleared the code through my phone. Here is the trace which resulted in the ECU error @ 126F:



Next, I shut down the engine, reconnected the fan, and turned the AC on (so I would be sure the fan was supposed to come on). By this time, the coolant was up to 151F, and there was no ECU trouble code present. Here is that trace:



Next, I shut down the AC, and grabbed another data point:



I observed the positive pulse width widen to approximately 80%, and then all of a sudden, nothing. It looks like I burned out my oscilloscope in the process... Not really sure why (supposed to be good for +/- 20V), but I was looking for a reason to get the newer extended range (120V) version anyway.

So, it looks like the pulse from the ECU varies from -5V to +12V, at about 100Hz. It looks like the ECU is looking for a positive pulse, which is likely tied high by the fan coils themselves. I am thinking my fan is OK.

With all that being said, I guess I will have to pony up and call my local GM dealer to replace the stupid controller. I was unable to find the controller online, and AutoZone was no help either. They looked at me like I had 4 heads and their computer only came up with only the aftermarket cheapy temperature controlled relays.

Anyway, I hope this helps those tech-heads out there in the future.
 
#14 ·
Hi TomatoSoup. Unfortunately, my local GM dealer is closed on Sundays. I'm not sure how much it will cost to replace it, but I'm sure an OEM replacement will be at least $300.

It looked like the guts of the controller was just ABS plastic. I used a Dremel to cut out a small portion to get to the high side drivers. They are IR F1404S FETs, datasheet is available at Invalid Request. I ordered replacements for $15 bucks.

In an effort to repair over replace, I soldered a small lead to the FET gate. I still have one good channel on the oscilloscope, so I'm going to perform the test again using a VDR circuit to protect the scope. If the controller logic is still providing an output to the FET gate, I should be able to replace the FETs and move on. If not, I will have to order an OEM assembly.

Unfortunately, it is now raining. (It never rains in SoCal?) Will post a followup if the rain lets up.

 
#15 · (Edited)
http://www.solsticeforum.com/forum/f30/solstice-gxp-radiator-74545/

Above is the link for the guy that was selling a fan and/or radiator from last August on this forum. It might be worth checking. I was unable to find the controller by itself new when I needed one. Link below is the entire assy new at GMPartsDirect, $302. (An email they sent me with the link also stated "19130227 Engine cooling fan shroud; includes motor, blade, shroud and module".)

19130227 SHROUD
 
#17 · (Edited)
Thanks.

The OEM replacement is $370 shipped - that is a bit pricey... It is frustrating that you can't buy the controller solo.

Adding a touch of reverse engineering... The controller essentially has two sides to it. On one side, the high power circuitry (the FETs and back EMF protection Diode). On the other, we have the low power logic control. It is my guess that the severe overheat condition, originally caused by the "bad jump start", may have fried the logic control.

Additional testing revealed the FET gate is either too fast for my oscilloscope to measure (unlikely, but possible), or it is perpetually sitting at 0V due to a logic control failure. If it is just stuck at 0V, the FETs will never turn on and close the circuit to the cooling fan. I have ordered a $40 function generator from Ebay to generate a new gate signal, and an upgraded 120V capable oscilloscope.

It is now my intention to use the function generator to test the FETs themselves, manually - under the assumption that the low power portion of the board has died. This device was designed to never be taken apart, including ultra-high temp solder which my iron won't melt. So, removing the low power portion will fundamentally ruin the device, but who cares at this point?... However, with the knowledge I have obtained thus far, it might be possible to salvage the high power side of the board and design a new analog (or better yet microcontroller) based circuit board and place it in the same case.

If the existing FETs are proved to still be viable, I would likely choose the analog path initially just to get the vehicle back in full operating condition. Using an opto-isolator 4N25, I can isolate the ECU signal in case of failure (and fool the ECU into thinking the OEM controller is in use). Then, it should be a simple matter of providing a valid gate signal to the FETs. If I choose the microcontroller option, it is feasible to build in a case thermistor to sense compartment temperature and keep the engine bay a little better ventilated. (It gets hot and dry in SoCal). Ordering the OEM part will be a last resort.

It will be at least a week or so before I get my parts in, I will post back with results.

If anybody happens to know the name of the 2 pin fan connector, and/or the 3/4 pin harness connector, please let me know. I'd like to order plugs to make extension cables for the fan and harness.
 
#21 ·
Just a quick update. A forum member contacted me indicating he had a spare controller, and sold it to me. (Much appreciated!)

A 5 minute install confirmed the replacement controller worked and that the fan was indeed OK.

I won't have time to work on a new controller design (at least for now), so I will be much more careful the next time I jump start the vehicle!

Thanks to all who followed this thread, and I hope the traces help someone else down the road.
 
This is an older thread, you may not receive a response, and could be reviving an old thread. Please consider creating a new thread.
Top