Saab Tech Talk (STT)

This should be perhaps subtitled the "Oops I did it again" thread, as in oops I have taken on another possible rescue project.

I picked up another 9-5 yesterday. It's a 2003 Glacier Blue Vector 2.0t, so not an Aero. Mechanical issues this time, I can either repair this one or use some parts to donate to the Graphite Green Aero. Not much in the way of service history, but the engine bay is generally clean, has fresh coolant and oil level was almost on max. It's done 133,000, the timing cover is not clean with a long standing build up of oil,some fresh.

The engine is noisy, rattle from the timing chain area, and a knocking which seems too light for big end knock but may not be. I thought the knock was also coming from the chain area, but it seems to get louder around cylinders 2/3 in the sump area. Idle is fine, it gets more pronounced above 1500 rpm. No oil warning light issues.

I have been told that the engine still performs fine, so I am a little sceptical that there is a rod issue as I would have expected something more catastrophic to have happened already

I have spent a couple of hours reading around, my plan of action is as follows:

1. Remove the chain tensioner and measure how far it is projecting - that'll tell me how bad the wear on the chains / guides is
2. Drop the sump, check the strainer first, then remove/replace the rod caps one by one to inspect the bearings and crank. I don't have plastigauge, will I be able to get a micrometer in to measure in situ ?

I haven't seen anything saying that the rod bolts are stretch / TTY, are they reusable ?

Any other B205/B235 likely issues I should be looking at ?

Better go change my signature :)

WIS doesn't make any mention of needing to renew the rod bolts- they're actually studs with nuts (like the exhaust manifold)

Thanks for that, that was my deduction, at least that means I can inspect without incurring cost, so many other manufacturers had moved to TTY by this time :)

Any other thoughts on my diagnostic route ?

If you take the valve cover off, you might be able to see if the top chain guide is intact (I'm sure I remember someone with a worn chain plus broken guide that made and odd rattle + clunk sound ..) and check wear on the sprockets

From your description I would think that the bottom end has gone. I knew of one chap who had a 9-5 which had a similar problem and he had only got about £1000 to either repair or replace the car but could save a few hundred quid a month. He asked what would happen if he kept driving it and we said eventually it'd die but we didn't know how long. He drove it for 5 or 6 months after that with the bottom end knock getting progressively worse until he could afford a decent replacement!

Cheers both, that's helpful.

I'm surprised that the bottom end could survive that long when it was knocking. To be honest, I'm hoping that's the issue, sump off, a good clean and new bearings would seem to be an easy & fairly cheap fix. I'm a bit more intimidated by doing the chains, given the problems with limited access, especially as I haven't yet got my head around why the new chain needs to be rolled on.

Anyway, here's a photo. Somewhat strangely, I quite like the colour

(http://937carrera.co.uk/IMAG1217.jpg)

I've always liked that colour, it only lasted for 2 years.  We nearly bought a 9-3 in Glacier Blue, but got the Merlot one instead.

I dropped the sump off this afternoon. All the nuts came off without any difficulty, which made me think they had been off before.

I drained the oil and though there was not much in the way of shimmer, I found some small particles of metal in the bottom of the drain container.

Numbers 1 & 4 big ends were whipped off first. They looked a little worn, with some signs of micro material removal. Nice of Saab to number the rod in the factory :)

Then the crank was turned 180 degrees and numbers 2 & 3 were removed

Hmm, that's not right, looks like the bearing has rotated on number 3 and the locating tab has been destroyed allowing lateral movement - this is just as the bearing cap came off.

(http://937carrera.co.uk/IMAG1221.jpg)

A better close up

(http://937carrera.co.uk/IMAG1223.jpg)

Now let's remove the bearing and look at the journal, how bad is it ?

(http://937carrera.co.uk/IMAG1224.jpg)

Doesn't look great does it, but surprisingly after measuring with the micrometer it is still well within wear limit.

...... and now the final reveal

(http://937carrera.co.uk/IMAG1225.jpg)

Time to get my thinking cap on

I could clean the journal up "in situ" and provided it cleans up OK and is still in spec put a new set of bearings in (I have "hand repaired" a crank successfully in the past)
Drop the engine out, remove the crank and take it for a polish or regrind
Drop the engine and put a s/h crank in
View the engine as being scrap, take the bits off I need and break the car

Before I finalise my decision I suppose I should do the measurement of the timing chain adjuster

Any other thoughts / suggestions -  from what I have read the bottom ends of the B2x5 engine are supposed to be pretty strong, though perhaps not as good as the B2x4 engines ?? 

Quote from: carrera on 22 November 2018, 05:20:43 PM

Any other thoughts / suggestions -  from what I have read the bottom ends of the B2x5 engine are supposed to be pretty strong, though perhaps not as good as the B2x4 engines ??

Love the colour of that car too, always wanted one but never likely to own one now.

I'd say that the B2x5 engines are plenty strong, I would suggest as strong as the B2x4 just the oil sludge problems gave them a bad reputation. I would seriously take a look at the chains at this point but as you say the timing cover has oil on it then at this point the only way you're going to get the engine good is to drop the entire lump out and rebuild the bottom end, put on new chains, probably want a new head gasket and oil pump seal to get the timing cover sealed up again... once you get into that territory you're looking at overhauling the entire the engine and personally I wouldn't bother to do that. Unless you want to do this for fun and have access to a decent workshop you can leave it all in bits while you do the work it's looking like it's not worth the effort.

I have found the source of the oil coating.... the connector for the breather box was off, so that's good news I think - unless it has been a deliberate removal to reduce crankcase pressure in some way. I doubt that the alternator will ever suffer from corrosion.  ::)

I do do some car repairs for fun, and because of that am fortunate that I do have access to a decent, dry workshop with a ramp. It makes life much easier, but it gets cold at this time of year as there's no heating. brrr

Having taken the sump off for the first time I know how easy that is, and it's great news that mains and big end bearings can be changed in situ. It'll be a bit of a pain (meaning more work) taking the engine out, but I can see it will make doing the chains easier simply because I will be able to see what I am doing.

Do you have to take the head off to replace the timing chains ? It seems you don't if you are rolling the chain on, but do you if simply fitting new ones?

Since posting the update above I removed the strainer from the sump. Nice design by Saab I thought, deep reservoir avoiding the need for the usual baffles. Anyway, the strainer was not massively clogged, but there was more fine bearing material around the outer edges, maybe a 10mm or so band round the edge. It would certainly reduce flow, but I would see that as the consequence of the bearing failing rather than the cause. Likely cause simply low oil ?? I left it draining overnight so might put a picture up later.

Next step is to remove the tensioner, but I will need to check that I have a 27mm socket first. I'm pretty sure I have one,, even if it's 3/4 drive.

Thanks for all the help and advice, it's really appreciated as I'm a bit short on Saab specific engine knowledge.

........and it seems everybody likes a Glacier Blue 9-5  :)  :) 

 

Quote from: carrera on 22 November 2018, 10:59:23 PM

Do you have to take the head off to replace the timing chains ? It seems you don't if you are rolling the chain on, but do you if simply fitting new ones?

You don't have to take the head off but the corner of the head gasket can get mashed when removing the timing cover. I wouldn't bother rolling a new chain in as every 9-5 that has needed new chains (based on tensioner wear or noise) that I've seen also needed new guides and rollers and a balance shaft chain. Given what you've said about the bottom end and the chains it'd be far more straight forward to do a complete strip down and rebuild on a bench if you're so inclined, check oil pump tolerances etc. at that point I'd be going for a new head gasket too just simply due to how much work it is the extra time and cost will be negligible overall.

I extracted the chain tensioner this morning. Lots of faffing around working out how to get access and where the allen key goes to fix the tensioner in a locked position, but now I have done the job I know it's not difficult at all.

Chains need doing I think, about 18mm of extension. Just for fun I included the strainer in the photo as well. The sump has lots of copper bits in it. That seems quite a lot of wear for 133,000 miles ?

Looks like engine out now. As you suggest, better for doing the chains, and for me to check / repair the crank.

When I have done clutch changes on the drive on a 9000 I have always dropped the engine out from below. I now have an engine hoist and have seen a photo with just the engine removed from the top. I assume that's the best way from this point ?

(http://937carrera.co.uk/IMAG1226.jpg)

Drop it out from below if you can, only reason I can think of to take it out the top is if you didn't have a tall enough lift.

Right, slight change of tack, the engine isn't quite out yet, and oops I did it again, I did it again.

I'm picking up another 9-5 tomorrow.  :-[  It has a good engine (don't they all)  :) which I can use for this car, and then use other components to do the graphite green car which will leave me with a supply of spares, some stuff for ebay or sale on here, and a shell. I think it makes sense.

The car I am picking up is an auto. I have assumed the big metal bits on the engine are the same, but there may be some detail differences between the auto and manual on the peripherals side so expect to have to swap the engine looms over, maybe alternator,???

Have I been wise or foolish, any words of wisdom for the transplant  ;) :)

I think that the looms for the engine all go through the big connector on the bulkhead.  Maybe one or two from the ECU bypass that connector.

Most of the bits seem to be the same, but the "engine" mountings on the gearbox side are different.

I've seen someone struggle with the flywheel/torque convertor end.  I can't remember which way around, but the end of the crankshaft on one engine has a central spindle. The spindle is a lump of metal and can be "persuaded" out.  It's not part of the crankshaft.

The actual electrical bits and wiring are beyond my experience, but the instrument panels and SID are different.  The SID difference is usually just programming.  The instrument panel is physically different.  The auto version has the gear selection displayed on it.

Two main differences spring to mind:

-the brake vacuum servo which is mechanically driven on an auto via a pump on the end of the front camshaft. You'll need to remove that from the donor and swap over the blanking plug.

- the sumps are different. Pretty sure you'll need to retain the one from the manual car


Beware- the auto gearbox is heavy- twice the weight of the manual (90kg vs 45kg)

Thanks both.

No worries on the gearbox weight, I won't be lifting it manually, that's what the engine hoist is for.  :)

I didn't think about the servo, knew the reversing light switch would be different and assumed there may be changes on vacuum hoses. I did an engine swap on a non Saab last year, manual engine into a Tiptronic shell. There were pretty significant differences, so the whole of the ancillaries, intakes etc were swapped over, not least because the tiptronic box was water cooled. Hopefully the 9-5 will be less involved.

The car is at home on the drive now. 110,000 miles, loads of service history, completely rust free on the rear arches and in pretty clean condition, she's been looked after. I only put her on the trailer using Low, apparently the gear change from low to 3 is lumpy / notchy and there is a transmission warning light on the dash. When in D the engine goes into a limp mode as well. I will have to see if my reader can get the code  (no Tech 2 as yet) and have a little drive to see what the symptoms are.

I doubt it's an easy / cheap fix,  but it would be silly to embark on major works without diagnosing the problem better.

I expect you'll find it's a sticky solenoid in the valve body.

The manual car and autos have different radiators. The auto gearbox cooler is part of the radiator. The engine oil cooler is separate on both.

Thanks both again.

I've been reading around since my last post and one thread suggested sticky solenoids were likely and then said "not repairable". Another one said a Saab specialist won't touch the auto boxes any longer as they are always having problems, while another said, strange we haven't seen any issues with them. My local tech said they hadn't seen any auto issues when I said I have seen a few auto cars for sale with box problems (before this one came up for sale). Maybe it's a "in original dealer network" / out of network experience difference.

I have also gone through the thick wad of invoices, including the first service at 6,000 miles and there's no sign of the ATF having being changed or checked at any point. Maybe that's my first port of call after the scanner.

If I do transfer the engine the rad will be staying in the recipient car and the oil cooler will come over with the engine.

The oil cooler was part of my thought process when I bought this car. Having reckoned it would be a minimum £200 in parts for bearings and chains, I then realised that with all the bearing material in the sump I should also change the oil cooler, it's the only way to be sure no metal remained within it.

Curiosity got the better of me and i just went outside in the dark with my codereader. Best surprise was that it pulled the code  :)

P0748 Control pressure Solenoid A electrical

Oooh electrical, could be a much easier fix than messing with hydraulics.

I haven't cleared the code yet, that's for tomorrow, time for some more specific background reading

Quote from: carrera on 29 November 2018, 10:21:11 PM
Oooh electrical, could be a much easier fix than messing with hydraulics.

It's possibly wiring from the loom to the solenoid, but I am expecting it more likely to be a failed solenoid and needing to open the gearbox up to replace if you're keen on fixing it.

I didn't get as much time on it as I hoped today, anyway

Code cleared, engine tickover became smoother as well as EML and transmission light going out.

No performance at first when using D, after a few forward / backwards using low and R it seemed to get better, maybe setting off in second or first. Sport mode improved things.

The error code returned when I was holding the engine on a fast idle. It takes about 3 seconds for D or R to engage, and then with foot on the brake there's a fairly sharp kick as the drive engages.

I checked the ATF level, got the engine / box warm, did the drive / reverse engagement bit then left it idling. The level was just a couple of marks above minimum. The fluid was not a bright red, more of a brown/red, a bit like dried blood ?? It didn't smell burnt.

I didn't buy the car to fix it, but you never know. It's only done 110,000 miles too. I don't generally do automatics, most of the posts I read suggest that as the car is post 2003 it will have a 5 speed box, but it only goes up to 4 on the shifter and dash.  Speaking of which, why no turbo gauge  :o

My plan is to test the 4 resistances for the solenoids through the loom tomorrow, if they check out OK then I have read that the transistors in the  Transmission Control Module ECU's fail too.

I am getting the message that you suspect the pressure control solenoid regardless, sounds like the voice of experience, I just need to follow the right diagnostic process  ;)

Quote from: carrera on 30 November 2018, 09:10:51 PM

I am getting the message that you suspect the pressure control solenoid regardless, sounds like the voice of experience, I just need to follow the right diagnostic process  ;)

I do suspect that but you should always do the diagnostics, It's nice to be proven wrong and for it to be something simple.

OK, so I have whipped out the transmission control unit this morning, mainly so I could see the prices of replacements on ebay etc.

It's part number 51 64 256, two plug connections, made by Aisin, and also has the code 93-50 with a GM (not Saab) underline.

2002 model were 5540599, and the replacement is 55352924
2003 model  is simply listed as 55352924
2004-2010 is listed as 55 560 029

From what I have now read these seem to be very regular failure items. I found a thread at https://www.saabscene.com/forum/threads/162410-Gearbox-ECU-issue/page2 (https://www.saabscene.com/forum/threads/162410-Gearbox-ECU-issue/page2) which suggests that the TCU out of a 9-3 can be used. I'm not sure whether that means simply replaced, or replaced and reprogrammed with the 9-5 software. There are 9-3 TCU's on ebay at the moment at a very affordable price. Interestingly, while mine looks like new, the 9-3 ones (55 353 074, 2004 for B207E) all look weathered, are they in the engine compartment rather than the cabin ?

I have tested the resistance of the SLT solenoid winding - came in at 5.3 ohms with a spec of 5.1 to 5.5 ohms at 20 degrees, so subject to me doing a bit of loom wiggling and checking the connector in the engine compartment to make sure there isn't a broken wire it would appear to be another potential TCU failure.

Anyway, any further insight into the diagnosis or use of a 9-3 TCU / reprogramming would be appreciated.

My assumption would be that all the TCU's will be of the same component design, just with different software loaded ??

Some progress, of sorts.

I removed the battery and with a meter on the connector watched for changes in resistance for the SLT solenoid. All seemed fine, so I put some contact cleader in the 16 pin connector and reassembled.

Restarted and after about 10 seconds it threw the P0748. I cleared the code and left it ticking over. The error did not recur so with a warm engine / car moved backwards and forwards. the delay on the selector seemed to have gone, but while there was a fair thump while engaging R and  L, there wasn't when engaging D. D only seemed to give a high gear, 4th / 5th I suppose.

Switched off,  restarted, did some messing with throttle position looking at live data and retested. Still no CEL, but now D gives the low gears too. That's an improvement, but having checked the gearbox code FA57A01 I know it's a 5 speed, even though the dash doesn't have an indicator for it.

But, the thump when engaging R and L remains and the three second delay is back...... but no error codes. Wish I knew what a good 9-5 auto was like  ::)

So, progress of sorts, could still be:  solenoid / TCU / old ATF ???

If I do the pressure test will that help in diagnosis ?

This sounds like a failed solenoid, however any of the other problems will need to be sorted before you take it apart to replace the solenoid (assuming it is dead). I would suggest cutting losses here and get a replacement box unless you really enjoy the fixing part.

I had a chat with my local Saab specialist today.

TCU's from the 9-3 cannot be reprogrammed to the 9-5, different bus apparently. The guy there remembered one car with the same symptoms and a TCU swap cleared it.

I'm progressing on the diagnostic front, before potentially spending cash. I have tested the thermistor for the transmission fluid temperature sensor as it appears that there is an interaction between the SLT line pressure reading, the temperature sensor and the behaviour of the TCU. It makes sense to me as the pressure of the ATF will reduce as temperatures rise, other things being equal.

I have a reading from the TCU connection and 14 pin block that says 6140 ohms. WIS says that at 20 degrees that should be 900 ohms and at freezing it should be 2000 ohms. A bit of graphing and interpolation and at 10 degrees I reckon that should be 1400 ohms. So the temp sender looks dodgy, which could throw the pressure readings out.

Biggest problem now is I cannot find the temp sender  ::) WIS says it's underneath the box in one part, but I have followed the loom and the only connections are on the top, with what looks like two speed sensors. EPC only has the part listed for 1998-2001, part number 51 160 106. Is this a 4 speed / 5 speed difference I wonder if the sensor is integrated in somewhere else. There's a big multiplug that looks a candidate, but why integrate with another 10 or so connectors for the solenoids and how to replace it ?

Once this bit is finalised I'll do some ATF pressure readings as the final diagnostic

(http://937carrera.co.uk/IMAG1229.jpg)


 

Quote from: carrera on 03 December 2018, 07:03:32 PM
Biggest problem now is I cannot find the temp sender  ::) WIS says it's underneath the box in one part, but I have followed the loom and the only connections are on the top, with what looks like two speed sensors. 

You certain on that? My copy of WIS says it's "mounted on the valve housing" although now clue as to where or which component it is.

I'm certain I'm confused  :)

Transmission / Automatic Transmission / Technical description / Transmission Fluid Temperature Sensor does say valve housing, 2002 model year

Transmission / Automatic Transmission / Wiring Diagram / TCM part 2 / Component 535 says on underside of gearbox , but the graphic does look like the valve housing.

It isn't a translation issue either. I just created a Swedish language user and the description in the wiring diagram section is the same in Swedish

Another evenings reading (including some things called a Volvo) and I'm better informed

The thermistor is located within the valve body, not externally accessible, and according to page 11 of the pdf linked below, at 10 degrees C the expected value is 6445 ohms, pretty close to the 6416 ohms my meter said.

So, that's that theory out of the window, also meaning of course that the test values in WIS are incorrect, but the component referred to is only present in the earlier 4 speed boxes, so perhaps a different thermistor / response profile.

https://www.myvolvolibrary.info/Tech_files2_ficheiros/AW55-50%20SN.pdf (https://www.myvolvolibrary.info/Tech_files2_ficheiros/AW55-50%20SN.pdf)

That means tomorrow should be a chance to do the line pressures. Could be fun  :o

Some more references for other forum users

Link for the earlier 4 speed box - thermistor resistor readings agree with WIS, so the values in WIS only apply to the early box

http://www.vectra-club.ru/images4faq/ova/AKPP/50-42LE.pdf (http://www.vectra-club.ru/images4faq/ova/AKPP/50-42LE.pdf)

A pretty detailed set of test procedures for the 5 speed

https://www.sonnax.com/tech_resources/196-aw-55-50-af23-33-diagnosis-and-valve-body-information (https://www.sonnax.com/tech_resources/196-aw-55-50-af23-33-diagnosis-and-valve-body-information)

The system presures have been tested according to WIS at port B5.

When the car is started it shows about 5 bar in P, and within about 10 seconds it suddenly goes up to a solid 15 bar. When cold it throws P0748

Increasing the revs results in a slight increase in pressure, next the designated tests

Idling speed - 850rpm
In drive, 0 bar actual, 4 bar spec
In reverse, 15 bar actual, 5 bar spec

Stalling speed - 2400rpm
In drive, 0 bar actual, 13 bar spec
In reverse, 20 bar actual, 20 bar spec

The guidance in WIS doesn't fit this pattern, so more thinking to do, though the low pressure in D in both situations does point to "too low system pressure, fault in clutch C1"

I have found a TCU, but am not yet satisfied that is the fault. What is certain is that I am methodically getting way out of my depth  ;D


****

a bit more data, I decided to measure the pressure on the SLT, another reference document gave some typical values, though I have to convert from psi every time

Idle speed, actual value 5 bar, 0.3 bar spec
Stalling speed, actual value, 5 bar, 5.4 bar spec

Now my gauge isn't that precise and it's a long way from the drivers seat to where it is located, so for me at stalling speed the value is correct. I did notice as I engaged a gear the pressure dropped down to almost zero before returning to 5 bar.

Something seems to be forcing the SLT pressure to be high at idle

Problem solved  ;D ;D ;D

I picked up another transmission control unit and fitted it today.

Cleared the stored codes, switched off, on again and immediately felt the difference. I don;t have the ability to clear the stored adaptation values so just did a bit of the workshop test and a short test run. All seems fine.

This has been a bit of a diversion from my noisy chain / knocking thread. The engine was going to be a donor, but now this car is rescued and will be going back on the road. A bit of minor work to do and then she'll be ready for a new owner.

As there's a fair bit of auto transmission diagnosis / information here, and it could probably deserve being in its own thread, is it possible to pull them out into another thread where the info may be easier to find ?

Thanks for all the help and advice along the way

Now, I'm back on removing the engine from the glacier blue car. Everything is off, it's on the hoist, I just need to get the angles right and hopefully she will be out later today.

I have located another engine to inspect (from another auto) and hopefully drop in

Quote from: carrera on 08 December 2018, 01:32:42 PM
I picked up another transmission control unit and fitted it today.

Cleared the stored codes, switched off, on again and immediately felt the difference. I don;t have the ability to clear the stored adaptation values so just did a bit of the workshop test and a short test run. All seems fine.

Good progress, one thing I have heard but do not have experience of is that if the solenoids get sticky then the resistance increases and causes some component in the TCM to fail, so if it starts doing it again soon it would point to a possible solenoid problem.

The engine is finally out of the glacier blue car. Note for others, no you don't need to remove the electrical connectors from the fan housing, but if you do there's just 2 bolts and the whole fan housing can be removed giving necessary additional room to get the engine out.

I have been reading around about setting / adjusting the solenoid valves. When the box and engine are out of the breaker car I am going to have a little look inside the valve housing cover to see what's involved in refurbing / swapping them over

Bringing this topic back to life as the good news did not last on the transmission.

The car sailed through its MOT with no advisories.

In use however the car has started to misbehave and throw a P0748 after around 40 miles driving. The code will clear when cold, but comes back when hot. Further investigations:

The error occurs with the correct TCM and also a 2003 TCM that is "known good"
Transmission fluid temperature sensor - Has the correct resistance / translation to degrees C
Wiring, continuity and light bulb tests done on the wiring to all 3 linear solenoids
Resistance tests on the linear solenoids show all three reading 5.8 ohms. This is just a little higher that the 5-5.7 ohms I have seen quoted, but all three are the same

Now the important new information:
PWM confirmed at 300hz
Current draw - virtually zero , maybe 60 ma
Duty cycle on the SLT 55% at idle in P, dropping to 38% at 2000 rpm

Now reading around I have seen a theory that problems with the linear solenoids cause the TCM to go into "high current" mode. I am pretty convinced that this is a myth. Why ?, well these two graphs give the answer (courtesy of Sonnax)

Before I go any further I am trying to be certain that this is a mechanical problem rather than electrical. I have read and been told that with the development of the valve body the design of the SLT was changed with the electrical connector moving from the top of the SLT to the bottom which was done apparently to reduce contamination.

You can see that low amperage gives high line pressure (the opposite of the prior hypothesis), and also that the duty cycle is above the expected range. It makes sense to me that higher duty cycle would give higher line pressure.

I'll admit I haven't got my head around how the different control pressures work inside the box to provide a smooth shift, but do get how those pressures are tweaked by adaption to take account of wear over time. I also understand that the linear solenoids can be rebuilt and are mechanically adjustable.

Any further thought or suggestions for me. Key to my understanding is answering the chicken and egg question. Does the TCM adjust the amperage to achieve a target SLT pressure, or is it the duty cycle that is adjusted. What is really throwing me is the high duty cycle combined with low current.



https://www.sonnax.com/tech_resources/196-aw-55-50-af23-33-diagnosis-and-valve-body-information (https://www.sonnax.com/tech_resources/196-aw-55-50-af23-33-diagnosis-and-valve-body-information)

Can you guess what I'm doing ?

I suppose the colour is typical for 110,000 miles.

I'll be seeing linear solenoids before long :)

(http://937carrera.co.uk/DCP_9879.jpg)

Does it smell a bit burnt?  Mine did, but it was a lot blacker than that.

It smells, in a similar way that hypoid gearbox oil smells, it's not that pleasant at all, but I wouldn't say it smells burnt. I just had a whiff of GM manual gearbox oil and that was sweet in comparison.

After another check of WIS (to check that there wasn't something else to do to get the valve cover off) I spotted something new to assist my understanding.

QuoteThe solenoids are controlled by a 300hz PWM voltage via pin 26 B (SLS) and pin 35 B (SLT) and grounded via pin 36 B (SLS) and pin 15 B (SLT). They are controlled by a current that changes primarily in relation to accelerator position. A high pulse ratio producing a high current (approx 1A) results in a low system pressure. A low pulse ratio producing a low current results in a high system pressure. An open circuit produces maximum system pressure. The valves are then fully open.

The control unit uses bus information on the current "Engine Torque" and the gear currently engaged to control system pressure 

Now that's quite interesting. Though I do not know what the correct parameters are for an engine on idle and in Park, what I am seeing is a high duty cycle (55%), but a low current (60 ma)

Those two measures appear to be totally inconsistent, which is what I was guessing before.

Further head scratching required, but for now here's a picture of the internals of the valve body. The cover itself is beautifully clean, and I note that on the range of - A B C valve body designs mine is an A version, so the second iteration, which has the SLT connectors at the top, and the shorter wiring harness

(http://937carrera.co.uk/DCP_9879a.jpg)

I decided to be really careful in trying to remove the electrical connectors. Good job, as apparently if the plastic bridge on the linear solenoid breaks the solenoid is u/s........ unless you buy some aftermarket clips available on ebay in the US (Sonnax parts).

Having tried a couple of techniques I left it before I did any damage to allow me to do some more research, in the process of which I found this. Looks like I was doing it wrong and right at the same time. Of course, at the moment the solenoid is still up in the valve housing area, but it is possible to unhook the wiring harness to release the SLT from the housing

http://faeriesintheattic.com/wordpress2/wp-content/uploads/2016/08/aisin-solenoid-harness-removal.pdf (http://faeriesintheattic.com/wordpress2/wp-content/uploads/2016/08/aisin-solenoid-harness-removal.pdf)

Finally managed to get the electrical connection off  by using a small jewellers screwdriver under the plastic tab, while at the same time pushing down on the plastic tab that retains the connector.:) 

Now, having said that my brain cell has been whirring away on the failure mode. This is a fixed frequency PWM application - 300Hz, which means the solenoid switches on and off 300 times per second. More "off" time (it is a normally open solenoid) through variation of the width of the pulse is seen as an increased duty cycle, which should mean lower SLT pressure. So maybe running at too high a duty cycle on the PWM causes the electronic problem ???

A second consideration is that when there is an electrical failure (think of it as zero duty cycle) the SLT pressure defaults to high which is what is felt as a hard shift. In my case I think the TCM is trying to get the SLT pressure down (but how does it know??) by giving a high duty cycle, and then, because of the contamination the actual movement of the solenoid is reduced which makes the TCM try to go ever higher on the duty cycle. Duty cycle not current.

That's my current theory anyway, happy for others to thoughtfully develop it. Augmented slightly by having disassembled my SLT.  This was quite helpful, even though it is for something called a Volvo  :D https://www.youtube.com/watch?v=0nh0rp7HSrw (https://www.youtube.com/watch?v=0nh0rp7HSrw)

Next step is cleaning this solenoid. The holes look clear, but there was some gungy resistance to withdrawing the spindle

Time for a picture

I've never been that far into a gearbox, and I don't do anything oily anymore due to dermatitis, so I won't be going there in future :)

But, did you have to "uncrimp" the end of the cylinder to get the piston out?  It looks very rough.

Quote from: sgould on 08 April 2020, 09:12:54 PM
I've never been that far into a gearbox,

I've never been this far into an auto gearbox.

Spot on with your guess on the disassembly method. In the US you can buy repair kits where a replacement over cap is provided.

Any thoughts on the control methodology ?  I have just had another look at the wiring diagram and can only see the solenoids, input shaft output shaft and temperature sensors. Nothing that would provide direct feedback to the TCM on what pressure is being achieved.

So how does the adaption (aka variation from the default program map) work ??? 

I've seen your posts on UKS and you seem to be getting somewhere with a couple of guys with some experience.

One odd thought crossed my mind. How do Saab reduce the voltage when needed?  With all the PWM stuff around, I was reminded of the many destroyed TomTom satnavs in the early days.  They ran off 5v.  The in-car charger/power lead came with a standard cigarette lighter plug.  What no one was aware of at the start was that the voltage dropping was done inside the plug that went into the lighter socket. There were 3 wires in the cable to the TomTom.  A 12v feed and earth, and a sensor wire.  The voltage control was a fast on/off switching device that cut the 12v supply and turned it back on again - so effectively a square wave PWM - which once smoothed at the TomTom gave an effective 5v.  The speed of the switching was adjusted by the sensor wire reading the effective 5v at the TomTom end.

The problem arose when people didn't realise that the cable was no a standard 12v twin wire feed, and tried to hard wire the TomTom into the car's 12v system, after cutting off the plug that goes into the cigarette lighter.  It fried the TomTom.   My thought was whether Saab used a similar control which could damage the control module if the sensing wire got overloaded or the circuit failed to switch fast enough to control the voltage.  This may all be complete rubbish, but the TomTom design was almost 20 years ago, around the time that the 5 speed box was introduced.

Yeah, a bit of hive mind going on, building the understanding and gradually most of the pieces of the jigsaw drop into place.

I don't think it's a voltage regulation problem. WIS is fairly explicit that the TCM is fed by power from 30+ which I believe is direct from the battery positive. Reading around I think I saw something that said the TCM was good for 9-16v, so fine for flat battery and an overcharging circuit. Might be a bit of a problem if you are starting the car with a garage charger connected as they can probably put out over 16v when turned up high. (mine can do 40a on max, with start giving even more)

That's the TCM side of things. As I understand the PWM duty cycle side of things if your charging system is running at 14v and the duty cycle is 40%  the DMM will show 5.6 volts. If it was an analogue meter it would be flickering between 0v and 12v 300 times a second, if the needle could move fast enough. In practical terms that too would average around 5.6v

I'm still not sure how I can have a high duty cycle and low amps, when the TCM I am using for the test is (or was) a good one. Time will tell

Time for another update.

The linear solenoids have been dismantled, cleaned and reassembled. Not a lot of crud, but it seems enough to interfere with their operation.

Electrical tests showed that current draw was up to 0.97 amps in park with the engine off, and the duty cycle was 59%

Adaptation complete the car was road tested but after about half an hour, after a stop the P0748 returned with the proper TCM

I have the 2003 TCM and that still seems to be working correctly and the change was noticeably smoother. Only problem is that the gearbox warning light comes on with P0826 as the flappy paddles are not found.

I have a potential solution involving a mini power supply to provide 1.25v to pin 8 of the TCM to extinguish the light, fingers crossed

Well, the power supply arrived today

Pre tested the voltage regulator and set it to 1.248 volts with a spare battery

I picked up 12v from pin 32 (red / blue) and earth from pin 22 (black) on connector A

Bared some wire and soldered the feed wires direct onto the existing TCM wires, then used some heat shrink to protect the exposed areas. Not ideal with it being an unbroken Y joint, but serviceable. I didn't want to cut the existing wires.

Switched the ignition on and the error was still there, but after clearing the stored code there are no warnings after an ignition reset.

So this means a 2003 on TCM can be used with a 2002 car, no other changes, costs about £3

Here's what it looked like before I put the glovebox back. Test drives required