Transmission Tuning (advanced tech)

[Robottrainer]

I initially thought the same thing when I got into engine tuning, but changed my mind once I learned how it all works. If you think about it, the gdi, tivct, advanced knock control, and 10 speeds have enabled performance levels that just were not possible with 3 gears, static cams, no knock control, and port injection or carburetors. Just take the individual cylinder knock algorithms and gdi control…these both require 0.1 crank angle deg level of sensing and control at max rpm, only enabled by modern lightning-fast processor speeds.

I can’t speak super-intelligently on the transmission side because I’ve only scratched the surface (hence, this thread). But things like torque modulation, where it can Instantaneously cut engine torque from 600 ftlb to 300 for 20 ms while the shift executes, requires some level of coordination between the engine and trans, and the complication that goes along with that. I can only assume there is good reason for the parts I don’t understand. I do have to wonder, though, how they can do so much with only 2 solenoids controlling a gearshift and only 3 shaft speed sensors as feedback.

That development of all those parameters probably cost millions of dollars to develop. Computer modeling. Simulation. And the real time dyno testing with real time adjusting with every possible scenario thrown at it.

Robotics have come a long way due to advancements in computer technology. We can do things with them now that were unimaginable in the 80s.

Yes, the electronics let us do thinks that we're never possible. Drive by wire for one. However when it goes south, it won't work at all. I liken it doing banking with a computer or a pencil. Pencil is inefficient, slow, tedious, prone to mistakes, however it always works. And any pencil will work.


We rely so much on software and computer in so much that we are hanging on a thread. Power stations, banking, water purification all rely on computers. Take them out or corrupt the software and its a world of hurt. But I digress...

Sponsored

 

[shogun32]

do have to wonder, though, how they can do so much with only 2 solenoids controlling a gearshift and only 3 shaft speed sensors as feedback.

As we've seen several times. Ford believes in faking it. They just give it a fancy name: extrapolation and curve fitting.

 

[shogun32]

water purification all rely on computers. Take them out or corrupt the software and its a world of hurt

A few days ago there was a report that a water plant switched to manual control because of computer hack. I'm just amazed they HAD manual valves. And not just Siemens industrial controls.

I swear one of these days there's going to be an epic disaster because of false assumptions.

 

[Robottrainer]

A few days ago there was a report that a water plant switched to manual control because of computer hack. I'm just amazed they HAD manual valves. And not just Siemens industrial controls.

I swear one of these days there's going to be an epic disaster because of false assumptions.

Stuxnet

A few days ago there was a report that a water plant switched to manual control because of computer hack. I'm just amazed they HAD manual valves. And not just Siemens industrial controls.

I swear one of these days there's going to be an epic disaster because of false assumptions.

 

[Qcman17]

What I find funny is with all of these adjustments available to them the trans often shifts like crap. And they know it does too. The other beauty is the learning aspect intended to adapt to improve shift quality- what?

 

[Robottrainer]

What I find funny is with all of these adjustments available to them the trans often shifts like crap. And they know it does too. The other beauty is the learning aspect intended to adapt to improve shift quality- what?

So getting back to things. Here's a scenario: ( names have been changed to protect the innocent). Alt2 shift characteristics applied to Mud mode (aka track mode) solves the flare problen 3 to 4 and 4 to 5. Applied to sport mode, 2 to 3rd is a bit too harsh. Alt 0 in sport mode fixes the 3 to 4 and 4 to 5 flare, but the shifts are a bit wimpy. If one were to apply Alt 2 to sport mode and wanted to soften the 2 to 3 shift a bit, what's the best area to modify for that shift in PCMtec.

 

[engineermike]

So getting back to things. Here's a scenario: ( names have been changed to protect the innocent). Alt2 shift characteristics applied to Mud mode (aka track mode) solves the flare problen 3 to 4 and 4 to 5. Applied to sport mode, 2 to 3rd is a bit too harsh. Alt 0 in sport mode fixes the 3 to 4 and 4 to 5 flare, but the shifts are a bit wimpy. If one were to apply Alt 2 to sport mode and wanted to soften the 2 to 3 shift a bit, what's the best area to modify for that shift in PCMtec.

The [much] easier way is to use character 2 but reduce the shift torque in table 49853 until the shift is less harsh. You can enter negative values into this table. The shift will still be very quick but the engine torque during the shift will be cut.

The much harder way to do it involves making hybrid shift characters. I've done this for my own stuff. For starters, this assumes the alt1 2-3 upshifts are acceptable. You would continue to use alt2, but copy all of the 2-3 alt1 upshift data into alt2. This would include ramp time offgoing (ramp 1), ramp time oncoming (ramp 0 and 1), inertia phase upshift multiplier, torque transfer offgoing, torque rate offgoing (ramp 0), torque rate oncoming (ramp 1), OL torque rate offgoing, OL torque rate oncoming, torque modulation (including slow ramp), and possibly others.

As a side note, you can make alt1 and alt2 shift like one another, or either one shift like alt0, but alt0 can't be made to shift like alt1 or 2. This is because the inertia phase multiplier table doesn't exist for alt0; I'm assuming because they're all 1.0's for alt0.

 

[shogun32]

he much harder way to do it involves making hybrid shift characters

so what is the likes of Wengerd doing? Is he really changing cells all over the place? Or are there more 'global' knobs he's adjusting to get the desired results? I'm just glad I'm not at the mercy of million lines of code to shift.

 

[Robottrainer]

so what is the likes of Wengerd doing? Is he really changing cells all over the place? Or are there more 'global' knobs he's adjusting to get the desired results? I'm just glad I'm not at the mercy of million lines of code to shift.

I'm sure he gets more involved

 

[engineermike]

so what is the likes of Wengerd doing? Is he really changing cells all over the place? Or are there more 'global' knobs he's adjusting to get the desired results?

Most of the commercial tuners lock theirs but some have been cracked. I have a few of these and most of them expand the use of character 2 to many other drive modes beyond drag, increase max pressure (not to be confused with actual pressure), change the torque management on shifts, and change up the shift schedules. The changes are pretty simple and "safe", in that they use existing calibrated characters and don't make hybrid characters or change any of the myriad of other parameters I was trying to figure out. Some I would not call safe because they apply max torque during shifts even in character 0 and 1, which have much slower shifts.

Edelbrock and Roush don't change much of anything in the trans beyond shift schedules, and this is probably a very smart approach.

My understanding is that early Whipple 10r80 tunes attempted to totally recalibrate the transmission and it was pretty bad, so they went back to a basically stock trans calibration. However, I was able to get my hands on a Whipple 2021+ F150 cal, which is harder than it sounds. This cal is the same one you get with the Ford Performance SC kits, so it's very different than prior Whipple cals and IMO better. Anyway, in comparing the transmission cal to stock, they appear to have increased stroke pressure for some shifts, schedule changes of course, and changes to shift torque modulation a reasonable amount. Not a whole lot done there, again probably the smartest approach.

I'm just glad I'm not at the mercy of million lines of code to shift.

My buddy says the same, but when we compare logs his "power shifts" take literally 10x longer and he has zero acceleration during the shift. The acceleration never pauses during a 10r80 drag shift.

 

[shogun32]

My buddy says the same, but when we compare logs his "power shifts" take literally 10x longer and he has zero acceleration during the shift. The acceleration never pauses during a 10r80 drag shift.

that's what the JATO unit is for mounted amidships.

 

[Robottrainer]

The [much] easier way is to use character 2 but reduce the shift torque in table 49853 until the shift is less harsh. You can enter negative values into this table. The shift will still be very quick but the engine torque during the shift will be cut.

The much harder way to do it involves making hybrid shift characters. I've done this for my own stuff. For starters, this assumes the alt1 2-3 upshifts are acceptable. You would continue to use alt2, but copy all of the 2-3 alt1 upshift data into alt2. This would include ramp time offgoing (ramp 1), ramp time oncoming (ramp 0 and 1), inertia phase upshift multiplier, torque transfer offgoing, torque rate offgoing (ramp 0), torque rate oncoming (ramp 1), OL torque rate offgoing, OL torque rate oncoming, torque modulation (including slow ramp), and possibly others.

As a side note, you can make alt1 and alt2 shift like one another, or either one shift like alt0, but alt0 can't be made to shift like alt1 or 2. This is because the inertia phase multiplier table doesn't exist for alt0; I'm assuming because they're all 1.0's for alt0.

I would Imagine this would soften the track mode 2 to 3 shift as Alt 2 is global. You'd have to strike a balance. I would assume adding torque would result in flaring.

 

[engineermike]

I would Imagine this would soften the track mode 2 to 3 shift as Alt 2 is global. You'd have to strike a balance. I would assume adding torque would result in flaring.

I don’t think adding shift torque would change the flaring but who knows. The ramp rates are supposed to account for that.

The shift character in my truck uses alt2 shifts at wot, alt0 shifts at part throttle for 3-4, 4-5, and 5-6, and alt1 for all other shifts. I’m finally satisfied with it.

 

[Robottrainer]

I don’t think adding shift torque would change the flaring but who knows. The ramp rates are supposed to account for that.

The shift character in my truck uses alt2 shifts at wot, alt0 shifts at part throttle for 3-4, 4-5, and 5-6, and alt1 for all other shifts. I’m finally satisfied with it.

Mix and match

 

[Robottrainer]

Mix and match

I suppose it's the oncoming clutch ramp rate that would be the culprit

Sponsored