Repair-It

Didn't know what else to call it. It started as some very firm shifting. Then I realized my PRNDL (actually on this car a PRND4 plus S and M modes) was not displaying my gear. It would do different things like sometimes display R and N but no D. Sometimes 4. Sometimes just P. Actually right now that's what it's doing. P and then a brief R as I move the lever out of park. It sets the T/C light and the trans feels like it is in limp home mode and doesn't shift past 3rd gear. Also the shifts feel like max line pressure. Nasty. I played with the shifter while driving and it went down to 2nd at one point. Good lord. Ended up driving home in 3rd but it still sucked. Engine didn't blow up. Take that, Northstar haters.

I got home and scanned. Actually I used my MDI and Tech2Win and scanned from inside my house which was pretty awesome. Anyway I found 10 codes. Of course some are old and some are irrelevant to the issue here. The real smoking gun code is this. I did try messing with the shifter adjustment and that is NOT the issue. I believe that only matters for the "shift to park" error that these are also known for, which is a switch in the shifter assembly that knows if it is in park, or not. There are no other switches in the shifter assembly itself. The range switch is in the transmission and is called the "transmission manual shift shaft switch."

I have attached the info on the P1825. I didn't really do any of the flowchart yet but I have it narrowed down to really one of two things:
1) Mice ate into my transmission wiring harness
2) Faulty manual shift shaft switch

It's hard to say which problem I would rather have but either of them will be a good time to fix.

Flow chart step 4 seems like a good start to see if there is an issue outside of the switch. They have you unplug the trans wiring harness and use a Tech 2 to verify that the voltage stage for each pin on the switch reads "HI" (because the switch pulls down the voltage so with it disconnected it should be reading the reference voltage from the TCM). If it does, then I probably have a switch issue. If not, then I probably have some kind of wiring issue, or theoretically a TCM issue.

Unfortunately I can't really reach the harness connector to twist it off without kind of burning my hand on stuff. So I will need to let it cool. Then I will unplug and do that test.

If it is the switch, check out the replacement procedure PDF. It looks.....kinda bad. It would be WAAAY better on a lift and will likely suck ass on the ground. You have to drop the exhaust (which is rusty where it connects to the cats so I'll have to play games with the old bolts, and then replace them - SRX is exactly the same). You have to pull disconnect the driveshaft and pull it back. Then you have to unbolt the trans cross member and slowly lower the transmission. Oh, don't forget to remove the transmission pan. Once the trans is dropped a bit, this allows the shift shaft to be able to slide out enough to get the switch off of it.

I almost hope it's wiring damage? The harness disappears above the trans but I'm sure I could jury rig new wires from the TCM to the trans plug and connect them in (I say that before knowing where the TCM even is). That will be no fun honestly under the car.

While texting Ed I figured out a conclusive test.

The FSM says to unplug and verify logic high for all output pins on the switch. So that's step 1. Any low logic would indicate a short in the wiring harness.

If this test passes, the next step would be to selectively ground each pin and verify on the Tech 2 that a logic low is present. If I can drive all pins low, then the wire integrity is good. No shorts to ground. No open circuits.

That would suggest it's not a mouse issue but just a bad switch unit inside the trans...so this will be interesting to conduct.

Here's the diagram I really need.
Pin 2 of the trans connector looks like a standard park/neutral switch and actually goes directly to the ECM, not even the TCM.

The other four outputs go to the TCM, and it shows which pins of the trans connector I would need to test for my wiring check. I could even theoretically test the pins of the transmission side to see if the switch is working but I have a feeling I won't be able to access that well enough. I haven't seen a logic table so I'll make my own for signals A, B, C and P, for my PRND4 shifter:

Park
A low
B high
C high
P low

Reverse
A low
B low
C high
P high

Neutral
A high
B low
C high
P low

Drive
A high
B low
C low
P high

4th gear (all the way back)
A low
B high
C low
P low

Now the diagram shows 3rd and 2nd which are only available if you slide the shifter over to S/M and manually shift. I don't really understand unless the switch is just compatible with a car that happens to have a PRND432 type of shifter. I will say I got manual 2nd more than once so something is definitely wrong here. Actually, 3rd was the best I could do which is interesting since you can't even select manual third mechanically with my shifter. So here's the logic for that.

3rd gear (not physically available on my shifter)
A low
B high
C low
P high

And while I'm at it since I also got manual 2nd more than once (that was fun at 50 mph)
A high
B high
C low
P low

Symptoms wise, I always get a Park indicator. But at its worst, I wasn't getting ANYTHING else on the display. Of course it always gave me actual reverse and actual neutral. And D or 4 always gave me a forward gear. But the wrong gear. I guess technically it wasn't in limp home mode because I assume that's just a single gear (like 3rd) and is a function of the TCM being dead.

This will be extremely useful for the diagnostic steps. From a wiring perspective, the concerns are mouse damage and the errors could be:
1) Open circuit on one or more wires
2) Bare conductor grounding out on the trans case or something else
3) A complex scenario of multiple exposed conductors doing a combination of touching each other and/or grounding.

These issues would play out this way:
1) A broken wire would be stuck on a "high" signal for that pin. It would not be able to be pulled low by the switch.
2) A grounded wire would cause a stuck logic "low" signal for that pin. The TCM would just see a ground and assume the trans switch is closed for that position.
3) Who the hell knows? Kind of depends on the scenario. If two wires were touching, that wouldn't do anything in and of itself. But when the trans switch tried to pull one of them low, it would also pull the other one low. I could test this by shifting to a position that should only ground ONE of them, and if both get pulled low, then theoretically that's my situation.

Kind of interesting process really. Pine Hollow Auto Diagnostics type process. Good thing I watch Ivan all the time.

Started it up. Working perfectly this morning. Is it heat related somehow?

All of these match my logic tables EXCEPT for D4! What does that mean? Anyway, here's what I saw with no complaints from the car:

I disconnected the shifter linkage and ran the trans linkage through all positions which, maybe not surprisingly, went to D3 and D2.

Here are shots of each position.
From what I can see, every position matches EXCEPT D4. I am also able to drive all 4 pins from high to low. According to my logic analysis, with B in low, a D4 shift position is still the most likely so perhaps that's just how the logic works. But I think it should be high in that position. So even though it is "working" at the moment, this could be indicative of the issue.

Next step is to drive the car, wait until it acts up, and see what all of the positions read. Getting D3 and D2 may be critical to fully diagnose which will require disconnecting the shift linkage and manually pulling the linkage into D3 and D2. But I will easily be able to check PRND4 logic at least and see which pin is stuck or whatever the situation is. Maybe that B switch is the issue.

Found it here. P/N 96042540. $62.96.

https://www.gmpartsgiant.com/parts/gm-s ... &filter=()

Also Rockauto carries it but I only know that because I tried to cross the P/N. I can't find it in the normal vehicle listing even though the application lists my car when I click on the part number.

https://www.rockauto.com/en/moreinfo.ph ... 56&jsn=483

Rockauto price is $51.79. Would need a seal and not sure about anything else.

https://www.cadillacforums.com/threads/ ... ng.244067/

I just wanted to come back and update what the problem was. The manual shift indicator (Range Switch) located inside the transmission failed and had to be replaced. It's the module that tells the car what position the Shifter is in. I replaced that and it took care of all the problems.

Also, unlike what the service manual states, the driveshaft and exhaust do not need to be removed. I unbolted the center drive shaft support, let it hang down and tied it with wire in the lowest position it would hang, and on the exhaust, I slipped it off the mounts at the muffler, and hung it with bungee cords so it could drop down as the transmission was dropped down. Saved a lot of time not having to remove both of those items.

The part number from AC Delco is: 96042540

Amazon had the best price on it when I ordered.

Just got back from my test drive where I got it to act up. I don't know how related this is but the trans temp was about 135F when it started. Of course before it acted up it drove 100% perfect. No lights, no funny business.

I was also able to wiggle the shifter and drift between valid and invalid states, though not all the time. But when it did it was consistent, if that makes sense.

I also feel like I am able to change the state of all pins. Nothing is stuck low or high which seems to rule out a wiring issue (open or short). So maybe that's it? Needs a new switch?

On closer examination it does appear that the B pin is hanging high because that seems like the sole pin that is wrong when there is an invalid range. So my theory was right. And the state of it does change in like P or R so, again, it's internal to the switch and not an external wiring issue.

With Adam's help, I completed the manual shift selector switch replacement. Only took 3 days. Well maybe 2 full days because the first was a half day and today was a half day. So 2 full days. The Quickjacks are awesome. Took a bit to get it set up but once you setup a car, it should be easy the next time. They must be parallel when operating due to the scissor motion.

I didn't take much for pictures but I'll post what I have.

Let me summarize the job. First, get a Quickjack and creepers.
- Remove exhaust safety brace, rear O2 sensor plugs, and exhaust. This is actually not hard. The whole thing just comes off. The car being up so high makes this much more pleasant.
- Remove exhaust heat shield covering driveshaft.
- Loosen driveshaft bolts at transfer case for rear driveshaft. They are T45.
- Drive a chisel between the output flange and the driveshaft and it will begin to separate. Just rotate and keep doing that. Comes right off. We were perplexed by this for a while.
- Hang driveshaft from the shift linkage.
- Disconnect shift linkage from transmission.
- Then we spent forever figuring out the side access plug. LONG story short (we spent a long time on this including research and finally rewatched a portion of my valvebody video), there is no plug on the side of a 5L50E. SIGH. Bad directions from the FSM! They correspond to the 5L40E I believe. Making matters worse, we also COMPLETELY UNNECESSARILY removed the front driveshaft, since it blocks the "plug" (that isn't there).
- Try and fail to remove the seal for the shift rod. Just ruin it but keep it in place for max leaking. Another oopsie. If it isn't leaking DON'T TOUCH IT.
- Drain transmission
- Jack up transmission and unbolt rear mount from body.
- Remove transmission pan and filter.
- Unbolt the detent spring and remove.
- Unplug the manual shift switch
- Drive out roll pin from shift rod which secures the shift switch and detent thing to the shaft.
- What needs to happen now is to pull the shaft from the side of the transmission far enough to slide off the pieces from the shaft. Except there isn't enough room. So we lowered the transmission per the directions.
- Well it only goes down until the front differential flange hits the steering rack. Which isn't enough to remove the shaft. SIGH.
- At this point we thought, if we could pull the inner roll pin and drive the shaft inward, that would work. But to really get access to that pin, we needed to remove the valvebody. That's right.
- Once that was out of the way, I tried to pull the pin out with vice grips. No dice. Really needs to be extracted with some kind of slide hammer or something. You can't drive it in, either, as it would contact a drum inside the transmission. SIGH #47.
- We gave up on the inner pin and said, how are we going to slide the rod out far enough? It's contacting the body. OH I KNOW. Drill a hole in the damn body. Shown above. If the hole was drilled bigger or in the correct location, more of the shaft could have slid in and possibly made it unnecessary to pull the valvebody but I think it was necessary for us since I didn't drill the hole perfectly.
- Once the hole was drilled (off center, thanks to drilling at an angle), the rod slides out enough to get the crap off the rod. At this point the job is technically halfway done. Throw the cheap plastic switch in the woods or suitable receptacle. Or put it back in the box so you can look at it later.

- Assembly is the reverse of removal. Valvebody (after a 20 minute+ search for the missing 8th E8 bolt which was in Adam's pants), filter, chase threads on transmission, clean bolts off, pan. Jack up trans and bolt mount back to body. ETC.
- Loctite and torque the driveshaft bolts.
- Torque other stuff as needed.

It was less miserable with assistance but also the Quickjack.

Road test reveals that it is fixed. It still has an overly firm 1-2 but I guess I need to do some more shift adapt learning. Or I could reset it again.

But yeah 45 minutes of driving and it functioned perfectly. The car actually drives pretty nice aside from the engine mount situation. No lights on dash. Great steering, great audio. I still like this car but it needs to go now. 158k or so.

I forgot to mention: I let the air out of the cylinder that still held air to match the one that doesn't hold air. Result? It works fine, like they said it would.