Automatic Transmission... Lurch?

Post by **precopster** » 20 Sep 2011, 08:34

Really keen to see where this thread is heading and if you have success with the solder re-flowing. I have a '95 850 wagon that has similar hard shifts on a hard throttle, though shifting fine on light throttle.

Haven't seen any discussions about the AW 50-42 in such detail until jablackburn did a heroic solenoid swap a couple of months ago when his auto crapped out. It became such a lively discussion the bookies were taking bets...

turbozutek · Post by **turbozutek** » 26 Sep 2011, 17:01

Today I spent a little time looking at this problem some more. My plan for today was to get the solenoids exercised but of course wonderful Scottish weather (rain) screwed that plan up.

Instead I decided to have a further dig into the ECU with some diagnostic equipment. I floated the pins (unsoldered one side of) and tested all the power resistors on the board and found them within tolerance. Next I decided to test the capacitors for value drift, even though they look fine; they are 16 years old and located in a harsh environment.

On the board is an NEC Super-capacitor which is used to hold the adaptation table in the event of a power failure to the ECU. This tested out OK (well, my cap meter didn't go up as high as the value, but I charged it and it drove an LED for a long time before dying off) I then turned my attention to the other caps on the board.

There are three 47uf caps there, one small 10v one, one 35v one and one 50v. I took the 50v off first and it reads 26uf on the meter. The 35v one faired even worse, now only allowing 15uf of charge and the 10v one was close to good for its age at 38uf. With value drift like that, I'd seriously doubt the caps ability to handle ripple (smooth off the power) will be anything like when new. This situation is very bad news for digital electronics and CMOS semiconductors in general.

: Old and busted.; Old Capacitor.jpg (105.93 KiB) Viewed 4206 times

In short, all of these caps are knackered.

So on with some replacements! These are Panasonic low ESR caps, all rated at 50v and 47uf. We've come a long way in component miniaturisation, if you compare the sizes of the ones that I just removed.

: New Hotness!; New Capacitors.jpg (130.77 KiB) Viewed 4206 times

As you can see, I've soldered them a CM or so proud of the PCB, this is to allow for vibration and it also makes it easier to get a good solder joint on both sides of the board (important for strength in an application like this). At this stage I haven't applied any conformal coating to the board, as I might get it out for further experiments later.

That done, I put the ECU back together and re-installed in the car. With baited breath and fingers crossed I turned the key and stuck it in 'D' - everything worked and I went for a short test drive. Gear changes were good and everything worked perfectly on the drive.

The change is certainly a tiny bit smoother and the car seems to be 'thinking quicker'! On occasion previously I'd give the car a sharp throttle change and it would seem to 'think' for a while, scratching its head and wondering what to make of it all, then changing gear. Now that feeling is gone and the gear change comes far quicker. They are still, however, slightly jerky on occasion so this hasn't been a 100% instant fix.

If there was any electrical issue it feels like it's gone though and there may still be some mechanical problem like a sticky solenoid.

That said, all of this is still fairly meaningless as the auto box ECU takes a few miles of varied driving to correctly adapt and as I removed the super-capacitor, it certainly reset all its values to '0'. If the solenoids are generally being driven better now, improvement may not be evident for a while yet, so I'll post back updates to this thread in a few hundred miles.

I'll also post if I take any further remedial steps, like exercising solenoids.

Chris...

Post by **JDS60R** » 26 Sep 2011, 20:46

Test the values of the transmission fluid temp sensor (2 wire ,brass ,in front of trans)as compared to the trans fluid temperature.

Also - if a non turbo motor - check the air box thermostat. It controls the intake air from being taken from outside or from over the exhaust manifold. It fails in the hot air position. The computer backs down the power and makes the trans shift bad. A hot airbox after a 10 minute drive is an instant reason to check the airbox thermostat.

Post by **jblackburn** » 26 Sep 2011, 21:32

Interesting fix! I never would have thought that the capacitors on the board had something to do with it!

Nice detective work

turbozutek · Post by **turbozutek** » 27 Sep 2011, 04:17

JDS60R wrote:Test the values of the transmission fluid temp sensor (2 wire ,brass ,in front of trans)as compared to the trans fluid temperature.

Hmmm, yeah, that's something to check! I do think that ATF temperature is a pretty big variable in the whole line pressure equation. I'll maybe try and give that a check tonight!

I also wonder how the TPS inputs are; I think I'll also check that as well to be sure it's a nice smooth arc from idle to WOT.

JDS60R wrote: Also - if a non turbo motor - check the air box thermostat. It controls the intake air from being taken from outside or from over the exhaust manifold. It fails in the hot air position. The computer backs down the power and makes the trans shift bad. A hot airbox after a 10 minute drive is an instant reason to check the airbox thermostat.

Ordinarily the changes can be snatchy from cold start-up onwards. That said, I really dislike the whole air thermostat arrangement and don't really see why it's needed on an EFI car. Given that it generally gets no colder than -5 here in the winter, I'm thinking of totally disabling this system and locking it on 'cold'. If you see a problem with that, let me know!

jblackburn wrote:Interesting fix! I never would have thought that the capacitors on the board had something to do with it!

Nice detective work

Makes you wonder what other caps in critical systems might have drifted a bit, hey!? I guess it all helps define an older car's 'personality'!

Chris...

Post by **JDS60R** » 27 Sep 2011, 10:13

Generally there is no issue with locking to cold only unless,
1.Ambient temps will approach 20 degrees F of the oils pour point. (minimal running time on a paralyzed lubricant is always a good idea)
2. Minimal environmental impact is a concern ( heated air help combustion chamber heat up quicker and reduces misfire rate)
3. Extreme low temp environments -40F and below some heated air is needed to keep the car running.

The turbos don't usually need one as the turbo helps mix the intake air and fuel to promote better efficiency while warming up.

FCP has the thermostats ranging from $10 to $30. I am not aware of the parts suppliers in your area so this info may not be useful to you.
Perhaps you could share your findings to help the post be more useful "worldwide".

turbozutek · Post by **turbozutek** » 27 Sep 2011, 13:26

Thanks for that JD... Based on what you are telling me it seems like the impact of this mod in my part of the world should be fairly minimal. I'll have a look at doing it through the week if I get the time.

I did take the waxstat out and test it per the instructions and it seemed to respond to temperature OK, but I noticed this summer that even when it's a nice day outside (+25'c for instance) the flap is still allowing heated air into the plenum.

turbozutek · Post by **turbozutek** » 27 Sep 2011, 13:45

OK I did a wee bit more work on this today (had a couple of hours spare in the afternoon)

First off I read the resistance of the thermistor that tells the ECU the temperature of the ATF. At ambient temperature (14'c) with the car off and cold the resistance was 1076 Ohm and as 0'c should be about 2000 ohm, this sounds about right. I then took the car for a 10 minute drive around the block (engine to 3 o'clock operating temperature for 5 minutes). When I got back I checked the resistance again and discovered it had dropped to 442 ohms, which should place the ATF at around 40 'c - I think that should be about right.

For due diligence, I will also conduct this test after a much longer run, just didn't have time today... But it seems that everything might be in order temperature sensor wise.

That done and with everything handily in bits, I decided to exercise the solenoids a wee bit. I hooked up the LPR solenoid first, connecting one terminal to - on the battery and then pulsed the other to + about 50 times (just by hand). At first I didn't notice much but after about 10 actuations the solenoid was making a lovely sharp 'CLICK!' when activated.

I then repeated the procedure for the S1 and S2 solenoids, which gave a nice click from the get go and seemed to work perfectly for about 25 - 35 cycles each.

After that I did the LCK solenoid, but didn't observe any click whatsoever. Something was going on, as the cable gave a little spark exactly like the others did, but there was no sound of the solenoid activating at all. The solenoid is obviously working OK as the trans hasn't thrown any codes, but could it be jammed up / lazy?? Or is this solenoid just in a position (dipped in ATF say) that would make it silent?

Without another car to test it on, I can't say how it should sound one way or another? (Hint to anyone who can try this for me on their car)

Anyway, on the test drive after things are a decent amount smoother. Still not perfect yet by any means, but I really think I've dislodged a huge load of crud on one of these solenoids somewhere. Perhaps the shock of un-dampened +12v straight to the solenoid is the key, or maybe they have been pushed through a range that the ECU doesn't normally send them to... But we've obviously done something right because it is shifting lots better now!

Again, I'll post further observations as mileage increases!

Chris...

Post by **JDS60R** » 27 Sep 2011, 14:02

Place a wedge in front of the airbox thermostat and see if the shifting changes. It made a nice difference on my 95 NA. I used a bolt to take up the space. Its under the filter so there is no issue of it getting to the motor. The bolt will come out as soon as the new thermostat arrives.

mxqwjskspak · Post by **mxqwjskspak** » 27 Sep 2011, 15:35

very interested in seeing the outcome of the solenoid trick, as i have a similar problem

Automatic Transmission... Lurch?

Re: Automatic Transmission... Lurch?