Since my last foray into European cars went so well... How about an old Saab?

First, no worries on the relay. 

Second, and this is coming from someone with a pretty poor understanding of electronics, so take it for what that may be worth, but, looking at the schematic, if the brake fluid light comes on when grounding the parking brake switch, I'd assume that the fluid light has short circuited and is finding a ground via the parking light switch. 

I don't know how the lights are triggered to go off once the car is running, but if I'm right about the short, it's moot, unless you have a fluid leak at some point, in which case you wouldn't know unless the parking brake switch is grounded. 

But the behavior of the lights when the car is running is weird. D+ on the alternator has a diode that ought to keep power from backfeeding, according to the internet. Do you see voltage at D+ if you disconnect the ignition relay, with the engine off? If you disconnect the wire from D+ does that affect the cluster lights?

In reply to procainestart :

I've tested the brake fluid switch and confirmed that it is open.  I also get the exact same behaviour with the brake pressure switch and fluid switch disconnected.  And with the pressure and level switches as well as the gauge cluster disconnected, I have infinite resistance on pin 11 on the right side gauge cluster connector (which is what the light should ground through if either the level or pressure switches detects a problem).

47M is the park brake light and 47F is the fluid light.  With pins 11 and 9 both verified as open, it has to be getting its ground from something on the other half of the cluster.  Now, I really thought that I had tried disconnecting the D+ connector and still saw the same behaviour, but I've tried so much crap at this point, I don't remember.  Just to try to get ahead of the game a tiny bit, I did go ahead and order a new alternator late last night as well as a new brake fluid accumulator.  While I wait, I think I may just pull the alternator off entirely and see what happens.

the audio system in this car should be programmed to play nothing but Led Zeppelin's Immigrant Song.

Valhalla, I am coming!

I have to get to work shortly but if you ordered an alt and not bushings, add bushings to the order - they go bad, the alt bounces around and the ground wire breaks. Skip polyurethane and get standard rubber. Getting the alt removed can be a PITA - I'll get you some more info when I have time later

In reply to procainestart :

Good tip, thank you!  I've got a couple of the Dorman rubber bushings on the way.  It kind of looks like it's a bit of a challenge.  It's almost like someone installed the whole motor backwards.....

In reply to AngryCorvair (Forum Supporter) :

I think you might be on to something there!

Thank you for the package and kind words. :-)

So, the alternator can be "fun" to remove.

Loosen the bolt on the arm that holds it in place and untension the belts. You may also want to move (not remove) the coolant reservoir out of the way along with its associated plumbing for more room. That's because the mounting bolts are buried and you need to push the alt toward the engine to gain some access.

You don't remove the alt from its mount -- you remove the mount with it. I can't find my photo of the perfect combo of extensions and wobblies, but you basically are going to figure that out and snake a hex down below the alt, making sure that the hex goes into the cap bolt fully and squarely. If you have a worn or cheap hex, you're going to hate yourself when you strip one of the bolts. Well, that's how I felt when I did the last time I R&R'd one.

Inspect the ground wire for fraying. At some point, Saab went to a burly version, but I don't recall when. If you're a pack rat, pull the voltage regulator from the old alt. If the brushes are still decently long and it's a Bosch, it's probably worth keeping.

If you're replacing the belts, fwiw, I've used Continentals forever, because that's what my indy has always used. They can be a pain to get on; calm mindfulness breathing (read: frustration-induced profanity) can help. I have no idea how to properly set the tension. I just go conservatively and if they squeal on start up, I snug them down a bit more.

Here's what the bracket looks like; it's separated from the alt so you can see it better, and it's oriented as if it were mounted to the block. The one on the bottom is what makes it "fun":

Well, I have the alternator completely disconnected - the D+ terminal is removed, the battery cable is removed and secured, and the little wire that I assume is a ground strap of some sort is removed.  Other than the charging light not coming on, the gauge lights work the exact same way.

I officially don't have the slightest clue what's wrong with this thing and I'm kind of sick and tired of berkeleying around with it.

Don't give up. As someone who hates electronic repairs, I'm watching this with great interest. I'm fascinated by what the solution will be but I know you'll find it.

I decided to do something stupid.

Oh.  No, not that.  That actually might have been a good idea, though.

I de-pinned the left gauge cluster connector except for the earth connection and the wire that goes to the right side of the gauge cluster.  Parking brake light worked fine.  Starting at pin #3 then, I re-pinned the connector.  Everything worked as expected until I got to pin #7.  Now I've got weird lights again.  That's the oil pressure sender connection.  If I disconnect that and plug in #8 for the connection from the charge light to the d+ terminal on the alternator (which I have grounded right now), weird lights.  Everything else is normal as long as both 7 and 8 are de-pinned.

I'm not sure what that's telling me, but it seems like it's probably significant.

Have you verified the left gauge cluster ground wire shows continuity? Will the car run (and no dash lights) without the oil pressure sender? What does the d+ terminal normally send to the cluster, or how does the circuit complete to light the bulb?

I verified that the #2 wire on the left side has zero Ohms of resistance to its earthing point, I've verified that there's almost no voltage drop under load to that earthing point, and I've visually inspected that earthing point to verify that it looks good and the connection is physically solid and immobile.  The car will run fine without the dash attached and, in fact, runs fine even with the funky lighting situation. 

My understanding of how the D+ terminal on the alternator works is a little bit fuzzy, though.  The theory is that when the alternator isn't charging, D+ becomes a path to ground and that will illuminate the charge light.  Somehow the little gnomes inside the alternator that turn the pump that make the electrons move need to be tazered in the butt to start working and having the 2-watt bulb of the charge light in the path somehow provides that shock to pull them out of bed and get them to work.  Once that starts, D+ will be at the same voltage as the supply from fuse #7 (which is the other end of the charge light circuit) and the light will extinguish.

They also use that path to ground to provide the lamp test for other bulbs through some sort of witchcraft.  Apparently, when the +54 ignition source goes away when you turn the key to engage the starter, it's supposed to earth some lamps, like the brake fluid lamp, through the diodes in the cluster to the charging light and from there to the alternator.  Same thing with the park brake lamp.  Now, it also says that the check engine light is supposed to be on solid before the car is started, but mine is acting differently.  If I remove the pins for the alternator and the oil pressure signals, the park brake, abs, and brake fluid lights (the "right side lights") all come on full brightness and then the park brake and brake fluid lights go out, leaving just the ABS light.  If I toggle the position of the parking brake, the light turns on and off at full brightness like it should.  Turn the key off and right back on and now the right side lights behave the same way - although they don't stay on quite as long as the first time - and the check engine light comes on and then turns back off.

Disconnected the tach - no change.

Disconnected the clock - no change.

I really wish I would have grabbed the cluster off that car in the junkyard.  I'm still not sure why I didn't since that was on my list of things to bring back.  I guess I just got distracted and forgot.  It'd be nice to be able to try a different cluster to either rule in or out the cluster itself.

I get the same behavior when either 47D (oil pressure light) or 47E (charge light) are connected.  The only difference is which light gets illuminated.  And with both hooked up, the behavior is also the same as when only one is connected.  If both are disconnected, though, the other lights start working the way they're supposed to. 

If I disconnect +15 from the ignition switch, I get no illumination on 47P (check engine), 47D, or 47E.  Also, the right side lights are dark.  Turn signal and bright beam indicators work, however.

47P, 47D, and 47E are supposed to get their power source from pin 3 on the left side connector which comes through pin 1 on 152B, fuse 7, some other connector, and then from the ignition relay.  I've already jumped the relay and verified power on both sides of the socket.  The relay coil is getting voltage with the ignition switch in drive as it should.  Interestingly, when the ignition is off, I'm pulling about .6 amps across the relay distribution terminals - there's some sort of other relay clicking under the hood somewhere that I don't think I care about.  When I turn the ignition to "drive", however, I still only have .6 amps.

Pin 3 has battery voltage at the connector and with the ignition switch on, 0 volts of drop between pin 3 and the fuse that feeds it.  Checking the traces on the gauge cluster, though, I don't have any breaks that I can fine.  Pin 3 on the connector is good to the tach, 47P, 47D, 47E, and even 47K (shift up) - which I don't have a bulb for since this is an automatic. 

The way I'm looking at it, with the +15 source disconnected, I should still be getting the charge and oil pressure lights, right?  They get power from fuse 7 and then ground through pins 7 and 8.  I've got normal bulb resistance between pins 3 and 7 and 3 and 8.  I need to dig up some little aligator clips and see if I can put power directly on it and see what I get maybe.

I've got alligator clips, my digital multimeter, and a 2 amp, 12VDC power supply:

When I connect the +12 to pin 3, I get a charge light when I ground pin 8 and an oil pressure light when I ground pin 7.  So that works like it should.   When I put my voltmeter on the alligator clips, I get 12.04V.  If I move the probes down from the clips and on to the traces, though, I get 11.15V.  That's reading the voltage right where the clips are clipped on.  So that's a little weird.

Otherwise, though, the lights seem to operate as expected individually.  I think I need a nap to think on that.

I'm kinda lost in numbers and details and don't have my Bentley atm, but what happens when you disconnect the oil pressure sender connector at the sender? It's a single spade connector, sits above the filter boss, under the intake manifold.

As an aside, major kudos to you for sticking with this. FWIW, I've learned a lot, though, alas, it's through your frustration. (And some of my favorite GRM threads are mechanical mysteries.)

EDIT: When you re-install the cluster. :-)

In reply to procainestart :

Yeah, there's a lot going there and it's mostly notes so I can look back on what I've done.

I haven't tried disconnecting the oil pressure sender wire at the sender side, but I did unpin it from the cluster connector, so same basic thing.  With that unpinned, the oil pressure light doesn't come on of course, but the other weirdness with the lights still happens.  So I'm kind of ruling the sender and its wiring out.

Hmm, I wonder if any of these Dubai call girls know anything about automotive electrical troubleshooting.

In reply to wae :

In Dubai, you can't get your call girl license without a BSEE, so they know ALL ABOUT automotive electrical troubleshooting. It's just too bad Dubai is so far away.

I've picked over dozens of these cars in the junkyard over the years, and it never occurred to me to pull a cluster as a spare. If I had one, I'd have shipped it. That way, as soon as I dropped it off at the post office, you'd figure out the problem with the one you've got.

You know, if you have a minute to spare...  What do your gauge lights do when you turn the key to drive without starting it?  I'm assuming that the oil pressure and charge lights would come on solid.  What about the check engine, brake fluid, and park brake lights?  Follow-up question, do the lights behave differently if you turn the key to drive, let it sit for a moment, and then quickly turn the key off and right back to drive?

With the gauge cluster on my dining room table, if I leave the park brake and brake fluid inputs ungrounded but ground the main ground wire, those two lights come on for a moment and the go back off.  If I unground and reground rapidly, they come on again but stay on for a much shorter interval.  Just trying to see if that's normal behavior.

Okay, on my '89 Turbo, which is essentially the same car but no ABS or SRS, with the key at Drive, engine not running, the CE, oil, alt, and p-brake lights come on. 

A hair under 2 seconds later, the CE light goes off. 

The brake fluid light isn't coming on, but it could be a bad switch or the cluster, as the clock doesn't work and there's a spotty connection to the gas gauge. (Yes, the car has some deferred maintenance...)

When I quickly turn the key off and back to Drive, there is no difference: same 2-second CE, same other lights. 

Hmm, okay.  Well that sounds like that's all normal then.

When everything is connected, the lights are weird until the alternator spins long enough that it excites itself and starts putting voltage on d+.  Once there's voltage there, everything in the cluster goes back to normal.

My most recent test was to connect the grounds on both sides, the power to both sides, and then I grounded the charge light.  As soon as the charge light is grounded, the park and fluid lights come on dimly. 

It's not the best test in the world since I don't have everything hooked up, but I'm inclined to think that it might be the cluster somehow.

I'd like to repair this cluster instead of replacing it, but I'm not sure what could possibly be the problem.  I don't really see anything that doesn't look right and the traces are all pretty simple.

The charge, p-brake, and brake fluid lights all have diodes -- I assume, then, that they all turn off the same way when the car is started: D+ sends 12V "upstream." All three are also connected: when the charge circuit is grounded, so are the brake-related lights. 

The brake-related lights are connected to the charge circuit  by an actual wire, violet color, 0.5 mm. If when the charge circuit is grounded the brake lights are only dimly lit, then something at either end of the wire or the wire itself is hogging most of the electrons trying to get to ground from the brake lights. At least, that's what it looks like to me, fwiw.

I cleaned up the wiring diagram to show exactly how I have it connected on the table:

47E = Charge light, 47M = Park Brake light, 47F = Brake Fluid light

In that configuration, I get a bright charge light and the 47F and 47M lights come on briefly and go completely dark.

If, however, I break the +12V connection to pin 3 on the left side:

Now I get very dim 47M and 47F and 47E (charge light) is on almost at full brightness.

So it seems like my problem is when #3 doesn't get +12 from pin 3.  I don't fully understand how it works, but the circuit is pulling power from the other side of the cluster across the wire that goes from left #9 to right #7.  Once the path to ground on left side #8 is removed - either by the alternator finally producing voltage on D+ or by disconnecting the wire - then there's no path to ground there and everything is okay.

So, then, why is it acting like it isn't getting +12 on pin 3 when it's hooked up in the car?  On the table, I'm touching my probe to the same ribbon cable that the connector touches.  I've been moving the probe around on that ribbon to see if there are any dead spots on it and I can't find any.  I get battery voltage on pin 3 on the connector when I turn the key to drive.  There's 0 voltage drop when I backprobe pin 3 against the fuse.  But when I watch the amps on that circuit, I get the same draw no matter if the cluster is plugged in or not.

wae said:

So, then, why is it acting like it isn't getting +12 on pin 3 when it's hooked up in the car?  On the table, I'm touching my probe to the same ribbon cable that the connector touches.  I've been moving the probe around on that ribbon to see if there are any dead spots on it and I can't find any.  I get battery voltage on pin 3 on the connector when I turn the key to drive.  There's 0 voltage drop when I backprobe pin 3 against the fuse.  But when I watch the amps on that circuit, I get the same draw no matter if the cluster is plugged in or not.

You may have done this already, but if not: with the cluster installed, have you pulled the speaker grille out and tried to probe the left-side warning lights that all share power via #3? I know that the illumination bulbs are accessible from the speaker grille; presumably, you can get to the warning lights, too, though you may need to get creative with some bent probes.

This is just weird.  I've got battery voltage at the proper pin (#3).  I've checked both sides of the connector - back probed it and also checked the actual contact surface.  +12.someodd volts.  I'm checking the other gauge cluster connector as well - pin 3 is tied in to pin 9 which goes to the other side of the cluster.  I've got battery voltage on that other connector where I should. 

But when I check voltage on the light sockets that are tied in to pins 3 and 9...  Nothing.  I've got 0 Ohms of resistance from the pin to the contact on the light sockets.  But no voltage.  WTF is up with that?