energeezer

Just bought this 2004 Civic. Was fine on test drive. Now I have noticed that the coolant guage is intermittant and the park brake light sometimes comes on for no reason. The engine does not overheat and the rad fan comes on as required but the temp guage will suddenly drop to cold for a while then jump back up to normal for no apparent reason. Also the brakes work fine and the fluid is topped but for some reason the park brake light will pop on. Usually when I first start driving then it will go out. I looked in the resevoir and it does not have the sensor on the cap. The sensor wiring appears to come out lower down and aft from the body of the master cylinder.
The strange thing is that I have noticed sometimes as the coolant guage jumps up and reads correctly the brake light appears to go out at the same time however....this is not always the case. I have seen it where the guage is down and the brake light is off and vice versa.
How would I troubleshoot this? Could there be a loose ground on the instrument cluster or is it likely the temp sensor needs replacing along with the master cylinder.

energeezer

I'm convinced the two issues are related. Just did it again. No temp indication and PB light on. As soon as temp indicator jumped to normal the PB light went out. It has to be a ground or loose connector. How do I access the back of the instrument cluster panel?

ezone

Remove the bezel from in front of the cluster, then pull like 4 screws out of the corners of the cluster. Tip it out and remove the wire connectors.


Almost all of the cluster operates on data that comes from the PCM, so I would have been wanting to see what the PCM thinks the coolant temps are when the gauge swings.

You could also unplug (or jumper) the connector at the master cylinder and see what the light does.

energeezer

I think I figured it out. My cluster circuit board had some cold solder joints. As the ambient temp behind the panel changed and the board expanded or contracted these joints were becoming problematic. I removed and dissasembled the cluster then resoldered any joints I could see which didn't look good.
After reinstlling the cluster I have not seen the fault again but.....Until this goes for a week or so I wont know for sure.

energeezer

To all those people replacing thier cluster and messing up their odo readings. Take a look at the board first. I was surprised just how poor some of the soldering is. Over time this will cause all sorts of issues. It can be as simple as touching each bad joint with a tinned iron to reflow. Certainly preferable to the other options.

ezone

Post pics of the suspect areas?