New A6 Transmission parameters walkthrough

[Chris@HPTuners]

carneb,

you are correct, after all the idea of this whole complicated system is that *it* calculates the optimum input torque to achieve the shift in the optimum way. These tables look like a fudge factor to account for any issues that arise, but no doubt they are used by calibrators who cannot work out the complicated stuff.

You will never convince racing people that it is better to let the PCM cut torque to get a faster shift that isn't as hard on parts

The shift "feel" is primarily controlled via the shift time and as i explained in an earlier post you can either mod the shift time directly or use the intertia profiles to select a higher profile at part throttle.

I think we definately have enough params uncovered now for people to make the A6 shift well. The only remaining thing i have is to find out what makes the TUTD react slowly and if it can be changed.

[ktoonsez]

carneb,

you are correct, after all the idea of this whole complicated system is that *it* calculates the optimum input torque to achieve the shift in the optimum way. These tables look like a fudge factor to account for any issues that arise, but no doubt they are used by calibrators who cannot work out the complicated stuff.

You will never convince racing people that it is better to let the PCM cut torque to get a faster shift that isn't as hard on parts

The shift "feel" is primarily controlled via the shift time and as i explained in an earlier post you can either mod the shift time directly or use the intertia profiles to select a higher profile at part throttle.

I think we definately have enough params uncovered now for people to make the A6 shift well. The only remaining thing i have is to find out what makes the TUTD react slowly and if it can be changed.

Has anybody actually used the "shift time" on there '08 corvette? It does not have any effect on my car, I used 1-2, 2-3 and a downshift one as well 2-1 and they all had no effect. Wondering if anyone else has tried them on there '08?

[carneb]

The only remaining thing i have is to find out what makes the TUTD react slowly and if it can be changed.

Would we have to make changes to the timing of the fill phase and torque phase of the shift to speed up the TUTD shifts? It seems that the shift times we can change now are only for the inertia phase. In auto we only feel the inertia phase and so the shift seems to happen quick. When you request a shift in TUTD mode though you have to wait for the trans to fill the oncoming clutch, then transfer torque from the offgoing to oncoming clutch, and then we feel the inertia phase happen. All this makes it seem that the TUTD shift is delayed.

I guess changing fill and torque phase times runs a greater risk of blowing something up than what we have access to now though.

[Chris@HPTuners]

all shifts happen begin the same way with the TCM deciding a shift is required, all TUTD does is tell the code to shift to the next gear when you hit the lever, just as the shift speed/rpm tables would. At this stage i don't know if it is an input processing delay, a calibrated delay or some other delay.

You would *think* that given the vehicle is a corvette, GM would have the TUTD working as fast as it can and not purposely delaying the time it takes to begin the shift after you hit the paddle.

The desired shift times are the total target shift time for the entire shift, from being in one gear to being in the next gear. This is the time the pressure adapt works to achieve.

[johnh]

The Final Torque Factor is calculated like this:

FTF = {Shift Torque Factor + (Shift Torque Factor Adder * Shift Torque Factor Adder Mod)} * Torque Factor Cold

Shift Torque Factor is per shift & usually in most cals set to 1.0

Are you sure this is right?
In my G8 bins..the Shift Torque Adder Mod is all 0s which would make the Adder irrelevant.
The formula would then become: STF * CTF

But if that were true why are there different values in STF Adder if they are just zero by the adder mod?

[Chris@HPTuners]

yes it is correct. it not uncommon for base values to be left alone and disabled by setting a multiplier or other calibration. eg. just becuase a vehicle doesn't have DOD doesn't mean all the DOD cals are zero.

[thesubfloor]

Hey Chris, have you had any luck yet figuring out why the A6 is shifting based solely on MPH and not both MPH and RPM like it's supposed to?

Thanks,
Christopher

[g8_795]

Hey Chris, have you had any luck yet figuring out why the A6 is shifting based solely on MPH and not both MPH and RPM like it's supposed to?

Thanks,
Christopher

Very interested in this. Even shifting based on MPH is wrong, but seems to effect the shifts unlike the RPM's.

[D-GTO]

Hey Chris, have you had any luck yet figuring out why the A6 is shifting based solely on MPH and not both MPH and RPM like it's supposed to?

Thanks,
Christopher

In my truck I lowered the shift speed a little bit and upped the RPM a little bit to 5800. The WOT shifts are right on the money.

[thesubfloor]

In my truck I lowered the shift speed a little bit and upped the RPM a little bit to 5800. The WOT shifts are right on the money.

Unless something's different about your setup, I'd be willing to bet money that if you left the MPH values where they are but either lower or raised the RPM it won't shift any differently.

I actually hope I'm wrong as it means your car isn't affected by the glitch but I'm guessing it'll behave the same as mine.

Christopher

[carneb]

The desired shift times are the total target shift time for the entire shift, from being in one gear to being in the next gear. This is the time the pressure adapt works to achieve.

The reason I think that the desired shift time we have access to is only for the inertia phase is from looking at logs from my car and trying to work out how the desired shift time and the "Time Shift" PID relate.

I've attached a screen grab of a 2-3 shift and marked on it when I think the shift starts, when the shift ends, and when the inertia phase starts. The time between shift start and shift end is 0.86s. The time between inertia phase start and shift end is 0.3s. The "Time shift" value in the scanner usually seems to be close to the start and end of the inertia phase once the trans has had a while to learn. In this case the commanded shift time (base + inertia adder + torque adder times) was around 0.32s. It's not exactly the same as the actual inertia phase or "Time shift" time but it's much closer than to the total shift time of 0.86s.

As for adjusting the turbine acceleration, I was looking for complicated ways and missed the most obvious one of adjusting shift time!

[johnh]

Found this, makes for some light reading.

[D-GTO]

My vehicle = '08 AWD Sierra Denali 6.2L A6, TVS1900
I am NOT an expert at this and have tons left to learn so please take these observations with a grain of salt because I don't know all of the technical terms. I may be looking at the wrong points but they seem to make sense ... feel free to correct me if I am wrong!

I performed a bunch more testing today to check this out a little bit more.
With Speed set to 28MPH and RPM setpoint @ 5800 it shifted everytime within a few RPM of the setpoint.
Leaving the Speed alone and changing only the RPM to 5400 netted an overshoot of about 300RPM shifting at near 5700.
Changing the speed to 10MPH and leaving the RPM @ 5400 the shift clutch was chattering on and off @ about 15MPH with the initial engagment @ 10MPH ... I am not doing this test again.
Adjusting the Speed to 27MPH and the RPM to 5600 it shifted within a few RPM of the setpoint repeatedly.
Leaving the Speed set @ 27 and adjusting the RPM setpoint anywhere from 5600 - 5800 netted shifting within a few RPM of the setpoint.

For all of the above:
When I say shifting I am talking about when the timing is first pulled for the Shift Tq. reduction. The 1-2 shifts complete on my truck in .4 sec and 2-3 in .2 sec. I use the point of when the timing returns to normal from the Shift Tq. timing reduction as the end point of the shift.

Using the diagram from carneb above ...
If the line labeled "start of shift" were moved to the left a little bit, where the "curr gear" begins it's transition from 1-2, on my truck is exactly at the MPH setpoint I entered.

If the line labeled start of inertia phase was moved left a little bit, to the point where the timing takes a nose dive, the RPM value on my truck would read within a few RPM of the setpoint I entered. This point is at the peak RPM of the shift.

If the line labeled "end of shift" were moved to the right a little bit to the point where all of the torque reduction timing was restored, that is the point where torque is coming back up and RPMs quit dropping.

Again ... I don't know if any of this is right it is simply based on my observations.

[thesubfloor]

My vehicle = '08 AWD Sierra Denali 6.2L A6, TVS1900
I am NOT an expert at this and have tons left to learn so please take these observations with a grain of salt because I don't know all of the technical terms. I may be looking at the wrong points but they seem to make sense ... feel free to correct me if I am wrong!

I performed a bunch more testing today to check this out a little bit more.
With Speed set to 28MPH and RPM setpoint @ 5800 it shifted everytime within a few RPM of the setpoint.
Leaving the Speed alone and changing only the RPM to 5400 netted an overshoot of about 300RPM shifting at near 5700.
Changing the speed to 10MPH and leaving the RPM @ 5400 the shift clutch was chattering on and off @ about 15MPH with the initial engagment @ 10MPH ... I am not doing this test again.
Adjusting the Speed to 27MPH and the RPM to 5600 it shifted within a few RPM of the setpoint repeatedly.
Leaving the Speed set @ 27 and adjusting the RPM setpoint anywhere from 5600 - 5800 netted shifting within a few RPM of the setpoint.

For all of the above:
When I say shifting I am talking about when the timing is first pulled for the Shift Tq. reduction. The 1-2 shifts complete on my truck in .4 sec and 2-3 in .2 sec. I use the point of when the timing returns to normal from the Shift Tq. timing reduction as the end point of the shift.

Using the diagram from carneb above ...
If the line labeled "start of shift" were moved to the left a little bit, where the "curr gear" begins it's transition from 1-2, on my truck is exactly at the MPH setpoint I entered.

If the line labeled start of inertia phase was moved left a little bit, to the point where the timing takes a nose dive, the RPM value on my truck would read within a few RPM of the setpoint I entered. This point is at the peak RPM of the shift.

If the line labeled "end of shift" were moved to the right a little bit to the point where all of the torque reduction timing was restored, that is the point where torque is coming back up and RPMs quit dropping.

Again ... I don't know if any of this is right it is simply based on my observations.

Were these tests done at part-throttle or WOT? The issue I mentioned earlier about my car only shifting based on MPH was at WOT so I won't be too surprised if part-throttle works correctly as I've never had any reason to test that out.

Christopher

[thesubfloor]

Adjusting the Speed to 27MPH and the RPM to 5600 it shifted within a few RPM of the setpoint repeatedly.
Leaving the Speed set @ 27 and adjusting the RPM setpoint anywhere from 5600 - 5800 netted shifting within a few RPM of the setpoint.

Also, I'm guessing that if you leave the speed at 27MPH and bump the RPM up to 3000 or drop it down to something low like 2000 you'll still find that it shifts right around 5600 every time as if you hadn't changed the RPM at all.

[D-GTO]

All of the above testing was done at WOT. I think that the initial clutch engage which appears to be triggered by the MPH is done premptively so that the inertial factors and timing can then take over. The RPM is the desired RPM but I think that the trans will not allow a shift outside of the timing factors programmed through the inertia adders plus base time etc. likely as listed above. I also noticed the TCC MPH limits which I have not adjusted yet ... still learning.

[LSX378]

Found this, makes for some light reading.

Thanks for posting this. Have been looking for this or something similar for a while and had not turned anything up!

Cheerz.

[LSX378]

If you have a look at this document http://www.google.com/patents?id=6KQ...BAJ&dq=6364811 it seems to describe how the 6L80 works. It talks about changing output torque trajectory and turbine acceleration trajectory to calibrate shift feel, then the TCM calculates the TM and oncoming clutch pressure required to deliver that.

Thanks for posting the patent. have searched for this before and missed it.

[johnh]

Well I did some testing.

First, I disabled Discrete Shift Torque..it was set enabled for 1-2, 2-3. I saw a few other had disabled this as well. Not sure why it was only enabled for those shifts, but my car shifts better with it disabled.

Second, I slightly tweaked some pressure settings and torque profiles (I used C6 stock bin as bit of a guide). Very slight changes.

Third, shift times all reduced to .15 (only in sport mode for me).

Car shifted decent but still had a bit of the tell tale lag during the shift (I call it granny shifting becuase of the huge amount of timing pulled (in some cases going to -15*!)).

Lastly I played a little with Shift Torque Factor (in beta), changed the 1-2, 2-3, 3-4 to 0.75 from 1. Still pulls timing, but not nearly as much sometimes even staying positive on the advance. However, the SOTP feel is much, much improved. Shifts "more normal" without the lag, just a nice smooth shift and under WOT car tends to push forward instead of having a lull. I tried 0.5 but the shifts seemed a little to harsh normally/part throttle. I think the next experiment will be to use the STF adder/adder mod tables so that the factor gets reduced at upper RPMs, stays same at lower.

All TQM is still enabled.

Overall it seems to be working pretty well..shift aren't harsh, but they aren't so lazy anymore. I'd say shift time is a big factor, make sure to examine the shift time adders as well. The other thing is the Shift Torque Factor...this made a dramatic improvement.

Chris any downside to keeping the Shift Torque Factor less than 1 that you know about?

[johnh]

I think the next experiment will be to use the STF adder/adder mod tables so that the factor gets reduced at upper RPMs, stays same at lower.

I tried this out.
STF at .8 (to help the lazy shifts all the time).
Put -1 in the Adder Multiplier, and .4 in the adder mod, for only 1-2, 3-2, 3-4.
Definitely a change, car shifts smoothly now normally, but WOT it invokes stability control as it breaks the tires loose. I have since backed down the adder mod to .2. So with the mulitplier of -1, my STF = .6, might still be harsh still have to test.

The nice part is shifts are smooth and easy until you really get on it, then they become firm and quick. As mentioned previously, I have tweaked shift times, a bit of the pressure and intertia settings (based off what the Vette does).