2seater

Oh man, I just returned to this thread again and I see my post from the this morning is missing. I will try to summarize as I am just too tired to hunt and peck type it again. The problem with the lights staying on is back, and will do it almost anytime the lights are on for more than a minute of two, at least after everything is warm. It is actually somewhat predictable. Leave them on with engine running, wait a couple of minutes and turn off. Won't go out or close, turn the switch back on, switch to hi beam and then turn off, everything works. Tried something I mentioned before, headlights stuck on, turn engine off. A sort of light knocking sound is heard from under the hood, very regular. It lasts maybe ten seconds and the lights go out but stay up. It doesn't last long enough to really pinpoint the sound but seems to from the area of the transaxle or where the air filter would normally be. Really hard to tell. The relays are in that area, sort of. They are mounted on the outside of the relay center, but it definitely isn't the click of a relay. A much heavier sound although not loud. I drove my car to work for the first time in a month last night and this triggered a memory of some other odd behavior. I will try to give a quick overview of the two possible clues to previous issues with the headlights. Several months ago, shortly after I installed the replacemnt H/L switch, when I arrived at work, I noticed the headlights turned off but did not retract. I chalked this up to shutting off the ignition at the same time as turning off the lights. I didn't know much about how this system works at the time so I thought it a coincidence. Turning the lights back on and then off again cured that issue. Never happened again. In early Aug. my replacement switch was getting crunchy feeling when operated. I was a little concerned but it seemed to work fine. One night on the way to work, the headlights suggested warning came on and I noticed the pilot light was out in the switch, although the headlights were on at the time. The lights did turn off normally enough. About two hours later, I was informed my lights were on in the parking lot. I went out to turn them off and it took several tries for them to go off and close. A couple of hours later and the same thing happened. I don't remember if I closed the doors manually or with the switch, but I determined it would be a good idea to at least pull the plugs at the bulbs to save the battery. As I removed the second plug, the lights popped right up, narrowly avoiding a nasty finger pinch. I ran them back down manually and went inside. It was about two hours later when leaving work, there it was again, with the lights up. It did it a couple more times at home in the driveway and I finally pulled the headlight switch, which was now stuck so the button cannot be depressed. I used the car a couple of days for daytime errands and the self-activation did not return. The rest has been detailed in this lengthy thread.

I did change the pump, but not that I suspected there was anything wrong with the original. The switch was so tightly installed, it tore the threads off of the initial 2/3 of the switch on removal. The aluminum is imbedded in the pump threads badly enough to prevent installing a new switch. I may be able to save it eventually but not while on the car. When I looked up the pressures I obtained in the FSM, they are spot on. I guess there are three possibilities; first it is normal, second, the new accumulator is defective, or third, the system takes more fluid than normal?

David T., I always knew in the back of my mind that the hi wattage headlights would be an issue, although I didn't suspect the headlight switch as the weak link at the time. As I said way earlier, I knew better, but out of sight, out of mind. I don't have time to go into detail at the moment, but you had asked if there was a previous issue with the lights and the answer is; maybe? I have been thinking about that and being at work sparked a memory. In any case, as Yogi would say, it's Dej'a Vue all over again. The headlights locked open at both ends of my 10-12 minute commute to work and back last night. Switch to high beam and everything turns off. I'm thinking headlight module. The BCM only makes a ground to trip the hi-lo switch. Can't see where stray power would come from.

I used to have a day job, and the headlights weren't used much, as I really dislike the look, so I suspect the switch didn't get the same abuse it does now. I have had 80/100 Hella lights for a long time, better than ten years, so I was able to prolong the original switch life by simply not using them. A night owl job and I lost two switches in six months. Coincidence, maybe? I have tried the hi-lo combo in a single bulb, Ranger pickup, and it definitely gives a nice view. The Silverstars have survived this abuse several times, but I'm sure the life has been reduced. On the Ranger, pulling the dimmer stalk will light both filaments just before going through the detent and can be held there. On very dark winter nights, in rural areas with extremely high deer populations, I felt the potential cost worthwhile. It has the side benefit of keeping the fog lights engaged as well, which were converted to Cibie driving lights. An old Saturn could be operated the same way but had four bulbs. The computer operated lights on the Reatta preclude doing anything like that.

Yes, I made the mistake that I could pump the pedal, but I wasn't thinking about the fact that I deadheaded the pump with no connection to the master cylinder. Pumping the pedal, which was rock hard, had no effect. The 8" long premade brake line places the gauge just above the strut brace so it can be read through the windshield below the rear of the open hood. The tube does need a couple of gentle bends to clear everything. I believe a relief line back to the reservoir is all that needs be added. The accumulator provides the cushion to slowly let the pressure down to pump cut-in pressure, which is really a separate integrated system from the master cylinder, pump, motor, accumulator and pressure switch potentially could be tested on the bench.

I wish I had seen the sketch earlier but what I made will work with a tweak. I have not looked at where the other end of the steel line goes, but I would think easy access would be out of the question. Perhaps your intention is to connect to the open end of the steel line? I did pick up a high pressure gauge, 8" long x 3/16" steel brake line and a couple of fittings to connect it. You must use the long nut on the brake line to connect to the pump. In order for it to bottom the line and seal properly, you must remove approx. 1/8" of the first threads on the tube nut. If you look at the stock steel line, you will see what I mean. It will not tighten all the way unless removed. The flaw in the ointment is there must be a way to relieve the pressure on the line to cycle the pump. Since there was no connection to the master cylinder, the pressure comes up to 2600 psi and the pump stops, but there is nowhere for the pressure to go, so it is trapped there ( I finally had to crack the tube fitting at the gauge end and let it drool into a drain pan. The setup will work just fine, but a needle valve needs to be installed in a tee below the gauge and a short drain line over to the reservoir so it can dump back into the opening. It can then be cycled by simply opening and closing the valve. All testing can be done under the hood, no pedal pumping needed and it can be attached with everything else in place except for removing the steel line. For reference, I saw 2600 psi for pump off and 2000 psi for pump on.

David T., I am using discrete high and low beam relays as described by the original post. One relay for the low beam and one for the high beam. I did replace the three relays this a.m., the two headlight and the original hi-lo relay in the console. I wanted all identical relays, but I needed one with a tab to hang the pair under the hood. The two without the tab show only a switch symbol for 85-86 terminals. The one with the tab shows the switch but also a small rectangular box connected across the switch symbol (resistor?). I could not find an exact match which included the tab. This is the exact same symbolism depicted on the O.E. Hella relays. I measured the coil resistance and the two without tab are 85ohm and 87ohm, with tab is 75ohm. I measured the three removed Hella relays and all are right around the 75ohm reading. I also checked across the #30 and 87a terminals and all used Hella relays measured less than 1.5 ohm. For reference, with the alligator clips in place on the VOM probes, I see almost 1ohm resistance in he probes themselves. I didn't check the new ones. The one anomaly, and possibly the source of the erratic operation with the engine running, is the O.E. hi-lo relay from the console shows apparent heat damage on both the 87 and 87a terminals. I did not notice it before when in the garage and under the dash. I had previously had the rear cover on the relay center in the console removed and actually pulled a couple of the terminals out and no heat damage was apparent. The heat looks to be from the inside of the relay as is darkest right where it meets the case. I'll open it up and see what it looks like inside. Not preferred, but the relay with 75ohm coil is on the low beam and the 85ohm on the high beam. Engine running, I cycled the headlights at least a dozen times with various "on" times and they function perfectly. Only time will tell I guess. Just one other note, or perhaps rant. I went to four different FLAPS; Autozone, Advance, NAPA and finally Carquest (actually sent there by the NAPA guy). Two places have the exact same junk, NAPA couldn't seem to find anything without a reference number or car model, but the Carquest guy took the time to look through the box of relays with me to find what I ended up with. If I had the luxury of time, I could have ordered exactly what I wanted online. Thanks for your patience and assistance.

I went out to the garage to look at the pump I removed and the steel line is indeed the common 3/16" standard tube nut thread, however it will only work with the long style nut on the tubing. It will probably seal okay even though the pump does not have the internal reverse flare like a normal female brake fitting. The original steel line seals on the flat bottom of the hole. I also looked under the capped hole on top just to the rear of the accumulator, but that passage contains what appears to be a check valve. I will investigate further.

Well, the Hellas that are operating the lights are used, they are the O.E. ones for horn and a/c. In their original location, they do not use the 87a terminal, so who knows the actual condition? I did look at the signal to and from the BCM at the original hi-lo relay in the console and it all looked correct. The wire feeding #86 on the hi low terminal is always hot and the BCM switches the ground on #85 to operate. Voltmeter showed voltage on both 85 & 86 in low beam and 85 goes to almost zero when switched to hi as the BCM makes ground. I did not wait more than 30 seconds or so when the lights locked on. What I was trying to confirm when I found everything working properly again: I had previously turned the car off when they first locked on, and the lights started to make a sort of ticking sound that didn't seem to come from the relays, more from the headlight motors. After maybe 15 seconds, the lights shut off, but I think they stayed open. I only did that the one time and wasn't recording what was going because I was so surprised. I will go to NAPA or a better parts store to see what I can find for quality relays. The only things I found close to my house are generic junk (IMHO) relays at the FLAPS. That has been getting more disappointing as I may as well order from JC Whitney relative to the quality. I will install three new relays as soon as I can find something, hopefully monday.

I think the gauge connected to the port for the steel line might be doable. The accumulator opening has that long spike that sticks up above the threaded opening, which is an unknown o-ringed configuration, making a simple fitting screwed on unlikely unless tall enough to enclose the spike. I think I will investigate further on that idea. By the way, my fluid drop is right at .45" and the brakes work very well. Since the thing was apparently air bound, I would think it will take a while to work all the bubbles out of the accumulator? The pump is so quiet, I cannot tell if it turns on while driving or not.

I posted too soon. Since it is now dark, I went for a short drive, just to see how things look with newfound brightness. There is no doubt there is a large improvement. Since they are E-code lights, the road signs light up way far ahead. They did so somewhat before, but they can be read at extended distance, not just glowing. I digress from the actual message. After only a ten minute drive, I tried to turn the lights off and the same failure of function occured. On low beam, the lights are locked in the on position, turning to hi beam, they will turn off and close. I repeated the previous test, with ignition on and engine off, and they work fine. Start the car, turn the headlights on, and they are locked there. Immediate followup to what was just said, while this post was still open, I went back out to test another combination, headlights locked on and then kill the ignition. I did that earlier when they first stuck, and it started ticking, so I wanted to confirm, but now they work perfectly, six times in a row, engine running. I am completely lost.

I know this thread had some detours due to not immediately understanding the suggested configuration and a bad selection of relays, but to ask why you would do this escapes me? In any case, I need to relate an additional happening that was not done during testing, primarily because the brake pump was apart and I couldn't fire up the ignition: When I substituted the OE Hella relays from the relay center, the headlights operated as predicted, as previously reported. I then returned the Hellas to their normal duty and installed one Bosch and one PEI relay until I could get a good matched set with mounting tab. The headlights worked exactly as with the Hellas, all normal, or so I thought. I completed the brake work and started the car to go for a drive. For whatever reason (it was mid-afternoon), I tried the headlights before leaving the garage. They turned on and opened just fine, but would not close or go out. The marker and tailights do turn off, so the H/L switch is functioning. I switched the lights to hi beam and tried turning them off and they did, as well as retracting as normal. I turned off the engine and just turned the ignition to run, and then tried the headlights, which again worked perfectly. I repeated the engine running test, and it did the same thing, if in low beam (default) position, the headlights lock open and on. I went back to the Hella relays from the relay center and they worked perfectly no matter the engine on or off. For some reason, one or both of the two other brand relays is either defective (brand new) or something doesn't like the higher voltage of the alternator charging the system. I did have a battery charger on the battery during some of the original testing, so I don't understand what is going on? I tried looking up the various relays online, to see if I could see differences in the coil assembly. I did not find much but the closest part numeber I could find for the Hellas indicated 85 ohm coil and no mention of diodes. The PEI stated a 90ohm coil, but not much else. The Bosch was no soap at all. I will list the part numbers below, but apparently this setup is sensitive to a specific style or type of relay coil assembly. Perhaps the electrical brain trust can divine something from this information so a more specific, commonly available relay can be included in the finished product. Hella 960388-02 (O.E. relays) PEI AR-5 Bosch 0332204125

I was thinking the same as you describe a few months ago. Fluid drop was in the 3/4" area, so I did install a new accumulator, labeled made in Germany. The fluid drop was reduced to approx. 1/2" with only limited change in the number of pedal applications before pump engagement. My follow on reasoning was the pressure switch on and off pressures were defective, either it turned off to soon or started too early, so the spread was smaller than it should be. The switch that I removed was a light green color as was the one I just installed, so I believe they were both replacements. I know there is a special tool to install below the accumulator to allow checking the pressures but no other easy way to do so. I am going to drive it a while to see how it settles out. To be clear, I have never had warning lights nor poor operation of the brakes, it just doesn't seem to fall in the range mentioned in the brake tests, which I believe may contain some errors.

Thanks Mc. I have tried loose steel line and accumulator but maybe I need to be more patient? Now I have good headlights and no brakes, sheesh. Edit: No success with just leaving the line and accumulator loose for five minutes. I finally resorted to applying 15 psi to the reservoir through the modified cap I have for bleeding the brakes. That took about 30 seconds and fluid was finally moving. Now everything looks good, although there is only a slight improvement in the time to pump start with brake application. Only time with tell if it is completely cured.

This started out to be a simple brake pressure switch replacement. I never had brake warning lights but could hear the pump run for a few seconds after a single brake application on many occasions. My conjecture was the pressure switch had the turn on and turn off points too close together. Luckily, Dave convinced me to take the entire pump, motor, switch and accumulator assembly from a fresh addition to the boneyard. The switch looks new. When I went to remove the pressure switch from the pump on my car, it wouldn't budge. I finally resorted to an impact wrench just to break it loose, which it did. I then used a ratchet to try to finish the removal. After less than one turn, it started to seize. I tried rocking it in and out but no soap. I finally had to resort to the impact to force the removal, and it almost rounded the outside as well. Of course the soft aluminum threads are almost gone and much is inside the iron threads of the pump. I thought the best route was to replace the pump with the takeoff that donated the switch, so far so good. I cannot get the new pump to prime? I did swap my pump motor to the replacement pump because the paint was clean and black. The pump turns freely and seems to pull vacuum on the inlet elbow when I ran it on the bench. I checked the inlet hose for unrestricted flow by draining the reservoir completely and replaced the 1qt of fluid with new fluid and the hose clamped off. I checked the hose flowed again after the new fluid addition by loosening the clamp a little to be sure the hose was full. The pump starts and runs very quietly but there is no change in reservoir fluid level. The pump did start to "gurgle" a little after a couple of minutes (cavitation?) but no apparent load on the motor or drop in fluid level. The FSM doesn't mention any sort of bleed procedure but I did loosen the accumulator and the steel pressure line but it didn't help. Does this sound familiar to anyone?

> I do not know, but after I swapped the filament wires around so the proper beam is activated, it acts exactly the same and it doesn't matter if using the Bosch or Hella relays. My original conjecture was the O.E. hi-lo relay in the console switches from high to low beam when the headlights are switch off, thereby eliminating any voltage on the hi beam signal wire. Maybe the difference in resistance between the high and low filaments? The lights go dark even before the doors start to close.

Thanks to you both for not yelling at me, but I would deserve some. I did reinvestigate my wiring harness and I did indeed make a mistake although not the cause of the weird functioning. I did reverse the connections to the headlights themselves at the relay sockets. The relay activated by the low beam (tan) wire was outputted to the high beam filaments, so apparently I traced over and over making the same error. I guess a nights sleep and fresh look uncovered that issue. Before I discovered that issue, I went to the interior relay for hi-lo thinking perhaps the BCM had somehow gone south and was pulling in the relay upon initial activation. Probing the back of the hi-lo relay connections in the relay center proved that all was working properly, default was indeed the lo beam (tan) wire. I had previously tested the relay itself for proper operation. I next pulled two of the factory Hella brand relays from the horn and a/c positions and installed them. The headlights now function as expected but there is a slight difference depending on the beam position @ shutoff. If in the default low beam position, the lights retract but do not shutoff, just as described by David T. After 5-7 seconds, there is a soft click from the relay and they turn off. If in the hi beam postion, the lights go out immediately and retract normally. My basic thinking on that would be the power to the lights drops out when the O.E. relay in the console defaults back to idle. I then tried two matching Siemens relays and the lights again acted as previously described or worse. Again, depending on beam, the lights may go out but the lights stayed open. Clearly, this setup does not like whatever is inside the Siemens relays. The symbol for the 85-86 side is marked + and -, which is not on any of other three brands of relays I have. The little diagram shows a straight line with a diode symbol between the two terminals with a switch symbol (rect. box w/diagonal line?) connected to across the ends of the diode. The diode symbol is very small and does not have much contrast to the case (black on black), so I did not notice it. I next switched them out for a Bosch and the PEI relays and the lights work just as with the Hellas. In the default lo beam position, they immediately retract and lights out after 5-7 seconds. Perfect operation as far as I am concerned. I feel somewhat better that it was an incompatible parts issue rather that an error in wiring (aside from the reversed hi-lo). I even switched relays around but there was always one of the Siemens in the mix, with changes in the headlight operation, but apparently never hit on the proper combination. Upon reflection, that should have been a clue. This headlight system does function just as originally described and I truly appreciate you sticking with me. Now I just need to switch the connectors in the relay sockets, which should be a piece of cake after all this. Whoopee!

MEA CULPA: I was going to delete the above post and repost it but if anyone has already read it I need to restate what I thought I knew and see about adding a big red faced emoticom. I do not know what this means but, after operating the dimmer switch back and forth, it appears the initial default on my system is to hi beam. In any case, if I switch to low beam, and then turn the lights off, the lights go off and the doors close. If I reengage the headlight switch, the lights open and turn on, but are on hi beam again. What makes this even more confusing, is they operate differently if I reverse the relay positions. I don't think I will detail the differences, but I suspect it is because of two different relays. One is a PEI brand and the only reason I used it is because it has a mounting tab that I can hang both relays from. The relay sockets are designed to be ganged together so both are supported from a single fastener. The other relay is a Siemens brand, which I have many of. The only difference I can see on the diagram on the case is the Siemens has a + and - on the 85 and 86 connection, and it will not work if reversed, so I imagine it contains a diode? I did not try reversing them on the PEI. I truly apologize for running everyone in circles. Please don't waste your time until I reverify my wiring. Maybe it is the mixed relays, something goofy about the default from the original Hi-Lo relay in the console or something else. Three things are now consistant, one is the headlights always turn on and open, are on hi beam when they do, and will only turn off and close if in lo-beam postion.

Yes, thanks to Dave I have completely operational headlight switch, so at least I do not have to jumper B+ to the headlight activation wire. It operates exactly the same as previously described. I use relay sockets equipped with approx. 12" pigtails for connections. The relays are wired exactly per Mc's diagram. 86 and 87a are tied together and connect to the original headlight wire for activation. 85 is ground with no splice to the left fender support using an original bolt hole, paint removed and a star washer under the ring terminal. Also is ground for the left headlight. All connections are bare crimp rings or butt connectors, soldered after crimping and heat shrink applied. 87 routed to 25A breaker connected to the stud inside the red plastic box on the right fender near battery. 30 is power out to the headlight. I have verified every connection I have with an VOM and tried all combinations to eliminate the possibilty of accidental cross connections. I applied power to the #30 terminal in both relay sockets to be certain the lights operate properly for hi and low beam. Now that I have a headlight switch, I should verify hi-low operation from the dimmer switch and try shutting them off on hi, although should default back to lo beam with power removed? I will also try this with fog lights applied as done by the OP. Not intending to be a wise guy, but the factory didn't seem to have an issue with splices within the wiring harnesses. Not that I want to do so, but my simple logic tells me that the factory connected the wire to the HDM after the on-off switch to the lights, the headlight switch in early '90 production and after the enable relay that was added in later production.

I meant to get back to this yesterday, but I will post my results now. I activated the headlights as stated before and left them activated for a full five minutes with no power to the headlight switch wire. The approx. 5 volt signal did not decay perceptably during that time, so everything stayed energized. When this issue was discussed initially, way back in the beginning, there was mention of relocating the yellow signal wire, between the headlight switch and the HDM module, so the module could see the bulb filament. I know there are two inputs to the HDM and the surmise was ths was for quad bulb setups. Originally my plan was to operate the relays as an enable relay and hi-lo selector. The yellow wire would be cut before the splice where it splits to the module, the feed portion of the wire direct from the switch would activate the enable relay. The end toward the HDM would be connected to the power input to the hi-low relay, which mimics the factory wiring. In this case, with discrete hi and lo relays, the two feeds to the HDM would be individually connected to the output of the relays, the feed wire from the switch would be terminated. I do not know if this will cause the headlight doors to activate momentarily when switching beams? I of course do not know if there would be any crosstalk inside the HDM to cause power out to the unlit filament? As it stands, it would be fairly easy for me to try it since I know where the splice is and my relays are about 6" away, but, I would have to actually cut the wire to prevent backfeed as we are getting now. I am certainly no expert, and could certainly use input as to the feasibility before I do so. Edit: I admit much of the calculation was beyond my immediate comprehension, but I agree the system voltage could be somewhat variable, probably ranging from a somewhat hi alternator voltage output to running home with alternator failure and only battery power to get there. I don't know at what level the ignition system will fail to function, but I would like the keep the option of headlights in a pinch, as remote as that chance may be.

Thanks for the detailed investigation. Since I do not have a headlight switch installed, I cannot see any means for the BCM to cause any power feed to the activation circuit as that appears to be isolated to a connection through the switch. I did leave the lights in the "on" position for a couple of minutes with no power on the headlight wire from the switch, so, if there is a delay, it is longer than I am comfortable with. I do not know if there are wiring differences between the years mentioned, but I think some had a separate headlight up-down switch on the console which the '90 lacks. Not sure if that is germaine? Five volts seems to be a common sensor voltage, although maybe just a coincidence or the product of the available voltage and resistance? It doesn't really matter I suppose, just how to deal with it. I would not mind losing any sort of headlight out warning, and it looks like it may have been used as an explanation of the headlight door operation? It would not take a lot of rewiring to do the headlights the other possible way, with an enable relay and a hi-lo selector, so that option is still available.

I just connected the voltmeter to the activation wire for low beam relay and it acts as before. Apply 12v to the output wire for the headlight switch B7, the meter shows full battery voltage and everything activates. Disconnect power from B7 and the meter voltage drops to about 5.5v, apparently enough to keep the relay pulled in. If I pull the hi-lo relay in the console, the lights go out and meter voltage drops to zero, but the lights stay open. If I replace the hi-lo relay, the headlights close. I repeated the test with new relays in both positions and the results are the same. It just dawned on me to check for voltage on the yellow wire, like a backfeed from the module? EDIT: Yep, there is a little over 5v present on the yellow wire at the headlight switch, which I imagine is the drain voltage from the headlight module. Maybe there is a reason the factory added the single on-off relay, aside from simplicity?

Yes, I will see what's what. I have an old analog 15 volt meter on the wall that hangs above the workbench just for quick checks. It may be actually easier to connect that so I can see in real time. Dave, I call if I can arrange something, maybe Fri. morning?

Sorry for the lack of clarity. I have no switch in the car. The switch went bad, not certain of the reason, but that prompted the much delayed headlight power upgrade. There are just two bare plugs in the dash, and I applied power to the yellow wire that runs down to the hi-lo relay in the console which also contains the splice for the wire to the door module.

I'm not certain I should post this quite yet as my good voltmeter decided to lose connection to the 9v battery making the display disappear, so I cannot confirm my conjecture, but perhaps there is some insight into the new relay system behavior. I do not have a good headlight switch to install as of yet but, I jumpered power momentarily to the yellow wire which should be the headlight switch output. The headlights immediately popped up and turned on, but, removing power from the headlight switch output had no effect, in other words, the lights stayed up and illuminated. Removing the relay that was activated caused the headlights to go out but the headlight doors remained open. Reinstalling the relay caused the doors to close, so it does see the original yellow wire connection to close the doors. My suspicion is there is power coming in on activation wire (original hi or low beam wire) that is enough to hold the relay engaged, although not enough to actually cause it to pull in? The only source I can think of is the original Hi-Lo relay in the console, but it is powered by the yellow wire from the headlight switch or something from the door module? I did go over my harness several times and also tried different relays.