Any auto gearbox experts here?

Thanks. Unfortunately this is a 2001 so no CAN. There does not appear to be an equivalent to the torque reduction request signal that you have highlighted. Data communication is over a single grey wire. It appears to use some proprietary protocol that is undocumented. As far as I can tell it is two way but I have no way of knowing exactly what data is being sent.

Check A3 terminal

PS: I do not see grey wire. Only B/BL, B/Y, G/Y and Br


     

Oops, my bad on the colour. I shouldn't rely on my rather erratic memory. Grey is the diagnostics (B9 in the diagrams). As you said, comms is green/yellow (A3 in the lower diagram or O on the top diagram).

In the manual it states that comms are one way, from the TCM to ECM. However in my testing it looks like it is bidirectional. If I disconnect A3 and look at the pins with an oscilloscope I can clearly see data coming from the ECM (pulses that swing from 0V to 12V). If I reconnect A3 I see additional data that appears to be a reply back from the TCM to ECM. That's why I am puzzled as to why I am getting a P1600. As it's one wire it's not easy to prove which direction the data is going but the fact that the data changes when the two modules are connected at the very least shows it's not a wiring issue.

As far as the other signals you have marked are concerned, idle up works. You can hear the engine revs change when you engage drive. Throttle position works because it kicks down (at least it does now I have replaced the throttle position sensor). I haven't tested engine temp but it won't engage 4th when the engine is cold so it probably works. The shift points seem sensible so engine speed probably works. Again I haven't actually probed the wire.

I am coming to the conclusion the P1600 may be a bit of a red herring and the main issue may well be simply a worn out brake or clutch in the transmission. It just about worked on the old oil but now it has fresh oil it's slipping. This car has done a fairly high mileage.

Do you have ELM327 adapter?
You can check what range (PRND2L) is selected in AT section of SZ Viewer program. May be there is some clue there. Also disconnect A3 contact and check voltage on it from TCM. Manual says it is 0V in other that P/N range (third screenshot)

Thanks for the SZ viewer tip. I just tried it but it doesn't like my 327 clone. I'll see if I can find a genuine 327 adapter.

I checked A3 and it works as it should.

While browsing through the manual I found some instructions on resetting the TCM learning memory. Working on the principal that I didn't have much to lose I tried the procedure. The car now drives a lot better with no noticeable slipping on down shifts. There is still a slight slip on 2nd->3rd up shift with the engine revs rising maybe 200rpm before falling back. I'm pretty sure I had that before as well. I only took the car for a short drive so maybe with a bit more driving the TCM will learn to compensate.

I had both modules unplugged for at least a month while I was doing the welding work so I guess the learning memory could have been corrupted during that time.

I still have the P1600 code but if the gearbox drives correctly I'm not going to lose too much sleep over it.

My genuine ELM dongle arrived today so I hooked it up. The TCM reports no errors, it shows the selected gear and the shaft input and output speeds look believable when driving. It's not reporting the status of the solenoids. I'm still getting the P1600 from the ECM but I think I'm going to quietly ignore it.

By the way, if anyone is interested the Veepeak VP11 dongle works with SZ viewer and costs less than £20. It's available from the usual online suppliers such as eBay.

I took the car for a bit of a run around some back lanes to exercise the box. The shifts are definitely improving. When I first had this problem the revs would rise so much I'd instinctively let off the throttle to stop it red lining. Now it's hardly noticeable most of the time.