Houpty GT

I have a 1991 convertible. I went out to my car after a heavy rain and the headlights were up all on their own. I started the car and could not get them to go down. The next day I was able to get them to go down. For a while they would go down on their own but still stay up sometimes. Now they rarely go down after you turn the headlights on except for maybe once a week.

I have determined the following. If the headlights are down, you can use the switch to pop them up for cleaning to open and close them all day long. If you turn on the headlights they pop up fine but will not retract on their own or with the switch for cleaning option. The only way to retract them is to disconnect the battery and they will retract properly after the battery is reconnected. The same is true for disconnecting the headlight retractor fuse or the headlight switch itself. Disconnecting the CPU makes no difference.

I guess there is something wrong with something in the switch circuit that stays charged once turned on and holds the relay closed. I am not sure once. I am currently enjoying the task of desoldering the switch but the solder seems to flow to the bottom were I cannot reach it to wick it or suck it out.

Has anyone done this repair or does someone have good knowledge with electronics and can point me in the right direction?

A new switch is going to cost $250.

Icemark

on a 91 vert, the controls for the headlight relays are in the CPU.

So if the disconnecting the CPU really does nothing, then I would be looking at replacing the switch.

Houpty GT

The diagrams I have seen show diodes and a relay built into the switch. Is this incorrect?

Most of the threads I found during my search were from people toasting the CPU by jumping or hooking their battery up backwords. One post had a car were the CPU and the switch were both bad.

I figure the relay inside the switch is worn and getting stuck or there is a bad inductor or something.

Saving $200 would be nice.

Icemark

On a coupe, or 88-89 convertible, yes that is correct... not on a 90 or 91 vert (SRS equipped).

Don't confuse the CPU with the ECU... two different things. The ECU is often damaged when the battery is hooked up backwards. CPU is almost never damaged (but does suffer from cold solder joints).

ECU= engine (or emissions) computer
CPU= body computer

ECU controls the engine, while the CPU controls headlights (on your 91 vert), horn, security, key lumination, turn signals, etc.

Houpty GT

I took the switch apart and there is a small black relay on the inside. It is a Taiko FXE-10N and it seems it is discontinued so I need to find an alternative. I inadvertantly broke it and have thus converted the car to one wiht the Daytime running lights modification which is an improvement over the previous condition since I can control it manually. I would like it to work properly though. I guess there is some relay I could use even if it requires a little custom wiring because it may not fit the board.

Icemark

The S4 switches use the Omron G6C-2114-US-DC12 as a replacement relay... which may be able to be wired into your S5 switch by jumpers.

I can check and see if I have any used FXE-10Ns laying around, but they all would be used.

Houpty GT

Let me know about the relays. I found some data sheets for the FXE relays. It claims the 10 represents the voltage but I would have thought it would have been 12 volts unless I am overlooking a resistor. There is a 12V listed on the data sheet. I think everything my be discontinued now though. The data sheet breaks down the rest of the model number which should be blank but makes not explanation of the 'N'.

http://www.globalcomponents.com/specs/web_fxe.pdf

Houpty GT

I tested the relay and it is 12 volts across the coil.

Houpty GT

I bought a 5 pin 12v PC relay from Radio Shack for $5 and wired it up using 6 inch long jumper wires. The headlights still went up and would not go down. I found out that I was disconnecting the wrong plug on the CPU and then it would work right. I removed the black/green wire. Now there are a few questions that remain. What exactly caused the circuit to short out during the heavy rain storm? Has anyone determined how to repair the CPU?