4L80E 3-4 Large Shift Time Error/TCC Cycling After Shift

[gearheadE30]

2004 Suburban 2500 with the 8.1 and the 4L85E. Fundamentally the same as the 4L80E.

Ever since I got it, every time the transmission shifts from 3rd to 4th and the converter was either previously locked in 3rd or the tables say it should immediately lock in 4th, the shift will happen, the converter will lock, unlock, and then lock again and be fine. The transmission and torque converter are mechanically fine and not slipping. No temperature issues or any other odd behavior. It does this with the stock tune. The converter appears to be a stock replacement rather than an original, and I am not sure if the stall speed is stock or not.

When this occurs, everything is good in 3rd gear, the 3-4 shift occurs (and feels nice and quick), and as soon as this happens the TCC slip shows ~-200 to -300 in the datalogs (I believe it still is using 3rd gear for the maths). TCC state then goes to "off due to slip" and the converter unlocks. As soon as it unlocks, TCC slip goes positive again and the converter re-locks. The shift error time roughly corresponds with the ECM thinking the 3-4 shift is "complete" when the converter unlocks and the slip number goes positive again. The 3-4 shift has a very large time error (sometimes as large as 5-6 seconds and always more than how long the shift actually takes).

I have seen one other person report this exact problem, but they never posted a solution and it was nearly 5 years ago now.

Any ideas? If the converter in this truck is higher stall than stock, can that cause the PCM to not "see" the shift? How does the PCM know that the shift is complete? Is there anything that EFILive has control over (desired shift time, shift stabilization ratios?) that would actually impact this?

[gearheadE30]

Did some more logging. The gear ratio PID shows ratio falling from 1:1 (3rd) to 8.75ish:1 before eventually falling to the correct 0.75:1. This happens even if I disable lockup in 3rd gear. The 3-4 shift doesn't feel terrible, but it is much softer than it should be. Shift time is right around 2 seconds. The transmission does not slip, get hot, or have any other bad habits. At this point it seems like it is probably a hardware issue that I can't tune out, but so far I've not read of anything that would impact only the 3-4 shift and not really have any other signs. Any suggestions would be appreciated.

[joecar]

It appears various "things" are going wrong.

Does your VSS indicate the true speed (i.e. are tire size and axle ratio correctly accounted for)...?

Was the PCM previously running a 4L80E...?

Post your .ctz file and a few log files.

[gearheadE30]

Needed to get some more logs and finally got that done. VSS is correct, all the other sensors are working properly in the trans (turbine speed, output speed, temperature). PCM is the original for this truck and has always been running a 4L85E. I can see that the turbine speed is what is causing the ratio reporting now; it isn't settling properly after the shift takes place.

At this point, I'm 99% sure I have a hardware problem, and given the lack of good information on the internet, I doubt I'll be able to diagnose it without taking it apart.

Trouble spot:
24129

[gearheadE30]

Here's the log that goes with it. Several shifts through the gears, generally 3-4 is at low load. Shift times are occasionally acceptable if I happen to be rolling off the throttle during the shift, which makes sense. Note the regular timing dips - this log also showed me my AC system needs attention.

24130

[joecar]

Instead of TURBINE, look at TIS.

Take note that TIS is read from the Forward Clutch housing which is driven by the Overdrive clutch, i.e. is input speed after overdrive, i.e. for gears 1, 2, 3 TIS is turbine speed, and for gear 4 it is turbine speed overdriven.

[gearheadE30]

@joecar thanks for the tip, I did switch to using TIS which shows that TIS and output speed seem to be what the gear ratio parameter is calculated off of, and TIS shows the engine speed instability following a shift.

I did end up replacing the valve body because I don't really have time to pull it apart, make and test individual changes, etc. I'm not sure what was wrong with the old one, but several things changed. What I am seeing of the adapt shifts in the log is the pressures are trending back towards 0 psi (previously it was pulling a lot of pressure/negative adapt). It also doesn't feel like I'm getting rear ended every time it shifts in tow haul mode, which makes some sense since the calibration says that pressure learning is disabled in tow-haul. The 3-4 shift is dramatically improved but not 100% resolved. Instead of a 1.88s shift fault, I'm generally getting a 0.3 to 0.8 second shift.

It appears that, even though there are valid shift times in the target time table for the 3-4 shift, and the adapt range is set to +16 psi across the board for all gears and -0 psi for the 3-4 shift, that the 3-4 shift is not adaptable. For every 3-4 shift I have logged so far, the "adaptable shift" state parameter drops to "no" as soon as the 3-4 shift takes place. It is yes for all other shifts under relatively steady throttle and once the engine and transmission are at operating temperature. I'd love for adaptation to work for 3-4, but I don't see any other parameters that would enable it.

I also have not found a way to reset the trans adaptations. V8 doesn't have any control that I can see to do the reset for the LS1B ECM, and if I go into the old V7, I can connect and activate DVT and there is an icon at the top that says it shows the transmission adaptation table for each shift. They are all populated with 0 psi, and clicking the "reset adaptation" button just gives an error that "this feature will be added to a later version" which is a bummer. My other scan tools don't have the functionality to view and reset this either, and leaving the battery disconnected overnight seemed to also make no difference based on the datalogs. So for now, I'm just driving carefully and waiting for the algorithm to catch up.

[joecar]

Post your tune .ctz file, let me take a look around in it.

Do you have the latest V7 software installed...?

Also log TFT, this may give us some insight (if fluid getting too hot then TC is having problems (or trans clutches/bands may be slipping)).

When you removed the VB did you notice anything unusual...? Was fluid good color/smell and was fluid level correct...?

[gearheadE30]

I'll post it in a bit, on vacation with no computer.

I do have the latest softwares installed for both V7 and V8, and updated the flashscan while I was troubleshooting so all set there.

TFT is great, even towing heavy in the mountains it has never been over 200F and usually sits around 150F to 175F. I still have the stock small trans cooler core.

The fluid was great, no debris in the pan, level was good before draining. The only quirks I found in the VB are a few check ball seats in the separator plate were quite worn and the shift accumulator pistons were pretty darn sticky. I didn't go too deep in tearing it down though. The manual valve bore seemed okay, but I didn't check the other shift spools. Likewise on the solenoids and EPC valve. There wasn't really any material in the filter or in that little screen in the valve body either.

[gearheadE30]

Still haven't resolved this issue. Put plenty of miles on the truck towing my toy hauler, no issues aside from having to pedal it after a 3-4 shift to get the converter to lock again. The shift time error is ALWAYS 1.88s, which matches the actual shift event completing based on turbine speed vs. output speed. I finally found a way to reproduce the issue with fewer variables, by letting the TCC lock in 3rd and then bumping the shifter to 4th. From the driver's seat, it is a nice crisp shift that you would never notice, except sometimes the converter unlocks. In the log, shift duration is long, engine speed is correct but turbine speed hangs high before popping down to the right value after 1.88 seconds have passed. The trans line pressure solenoid duty cycle actually drops after this error time passes, so it's not like the computer is bumping pressure to get the shift to complete.

It seems like some kind of software issue. I'm suspect of this particular OS anyways since it had some other issues that I was fortunately able to fix in visible tables.

My OS is 12612114. what I'd love to do is find OS12612115 for a 2005 Suburban 8.1 4L85E and do a full flash to see where that gets me. At one point in time this tune was posted in one of the repositories that is now gone....so I'm kind of stuck unless someone happens to have it. I have several reference tunes from 2001-2003 trucks, but they changed significantly between those and the 2004-2006. 2006 is flex fuel, so 2005 seems like it could be a happy medium if I could find that file to look at.

Anyway, here's the specific log event:
24468

[joecar]

What is the pid SHSOLA (I'm out of town away from my laptops)...?

The checkballs wearing into the separator plate are a concern... there are various wearproof balls/separators available (e.g. Sonnax).

[In-Tech]

Transgo too

[gearheadE30]

What is the pid SHSOLA (I'm out of town away from my laptops)...?

The checkballs wearing into the separator plate are a concern... there are various wearproof balls/separators available (e.g. Sonnax).

New valve body, solenoids, and accumulator are already in the trans, but the separator plate is an interesting thought. There was what my inexperienced eye thought was a small amount of wear, but unfortunately it looks like I didn't take any pictures. That would be a cheap thing to try.

SHSOLA is shift solenoid A. I was trying to match up the shift solenoid action vs. commanded shift vs. actual engine speed drop, and that all checked out.

I did bump the shift pressure up for 3-4 pretty significantly just to see what it would do...it actually seems to show some kind of tie up happening where now it is pulling the turbine speed down too much before it eventually ends up where it should be. It also seems to be pulling engine speed down a bit, but that could also be torque reduction. Reported shift times are definitely quicker.

I don't think there's an issue with the speed sensor because the only artifacts I can find in any of the datalogs are this 3-4 shift behavior.

Below screenshot has accel and torque pulled out for visibility, but ~250 ft lbs and 35% steady throttle. Stock tune does not have any torque reduction on 3-4 shift. I added torque reduction in an effort to speed the shift up until I figure out what's wrong with the hardware.

24469

[In-Tech]

https://transgo.com/

They will like it that you can alter the computer.

It's the plate.

[gearheadE30]

Got a plate and gaskets on the way, we'll see where that gets me.

[joecar]

TCC should be releasing during shifts.

Log the pid GM.GEAR which is decoded from the solenoids A and B, it is easier to digest.

Also log GM.TCCSTATE and GM.TFMPCS and GM.LASTSHIFT.

Drop some pids to get the fastest logging rate (10 frames per second for LS1).

[joecar]

Yes, TransGo also.

[gearheadE30]

Have the plate, got new balls as well, forgot to order the separate accumulator gasket so waiting for that to show up. Have a road trip coming up, and decided I'm not going to risk causing other issues swapping everything out until I get back from that trip.

I do have GM.GEAR in there, the solenoid A was just for that one log so I could see what that would show me, if anything.

Logging rate is a good idea, I'll see if there's something that's getting aliased out of the data.

GM.TCCSTATE shows "off due to slip" when the converter pops in and out of lockup. The slip number is actually negative, and checking engine speed vs. output speed, there isn't actually any slip occurring. It's only showing slip between turbine speed and output speed, somehow.

I have TFMDC% in there, have no tried TFMPCS to see what that shows me.

LASTSHIFT is in there, it's generally in the .25 to .4 range for 1-2 and 2-3 which is where it should be. target is .35 generally. For a "good" 3-4 shift, it's in the .7 range. 90% of the shifts with any torque behind them are 1.88 seconds. Always exactly 1.88, never more, never less.

I would like the TCC to release during shifts, but I haven't changed anything in the tune that would make that behavior different than stock. It has stayed locked through shifts based on seat of the pants feel since the day that I bought it with an untouched tune. The only thing we have control over through software is "allow TCC to lock during shift" checkbox, but all this seems to do in my testing is allow simultaneous upshift+TCC lock instead of stock, where the shift has to complete before lockup can be commanded. This is turned off.