So I have been working on this for about a year. I am going to start selling these. This is the new TCM-2800 from PCS.
Biggest changes are current monitoring and amperage targeting vs duty cycle and the whole system is a torque based tune. So derived from torque we target a amperage for the fill pressure and overlap control and then also from the torque we target a desired shift time. The shift time dictates amperage for the ratio change part of the shift.
We end up with a two stage pressure control for each shift. Much more refined and a lot more control of the more difficult shifts. We also have separate maps for upshift and downshift. So they can all be tuned independently.
There are a lot of other changes as well. Inertia tuning and many new setup features.
The other big one is we are on the factory CAN bus as an option. So this can be installed into a factory car as a replacement for the Factory tcm. This is only on ME 2.8 ecus and up. So pretty much any car from 2002-2003 up to the end of production. The ME CAN ids for 312 etc.
But this takes us from the early 2000s to the end of facory opted 722.6 cars. As well as the Dodge cars.

This will also be coming out with a package that I am developing for factory ecu and standalone tcu setups. For engine swaps. I will be starting the engine swap of a om648 very soon into my w123 wagon. Using the factory cdi module and a standalone DAS setup and the tcm-2800.

W:

What is the source of the torque information, and the form of that information as an input signal to the TCU?

We can do this a few different ways. But basically it is either receiving torque via CAN from the ecu. Ie factory setup. Or we use TPS vs RPM to build a torque map. The later is much harder to do. Unless you have a full 3d torque map from an engine dyno.
What I have done is taken a tune that was torque based via CAN and then dumped that into a starting point for a TPS/RPM setup. Then tune the torque map to the transmission tune. This gets you pretty close. Then you start tuning the pressures from there if you cannot do it with torque. This works well with similar transmission setups. Pressure is pretty linear to torque input for any given setup.

Well, the guy at the next desk wins the jawbreaker, I wagered for MAP vs RPM. I think that if the option were available, MAF rather than MAP would be the choice. Your reference to the factory setup suggests that for many recent engines torque information will be MAF derived. (Although some recent installations are using both MAF and MAP load inputs to the ECU.)

If one aspect of the control protocol is modulating fill pressure, and another aspect is rate of shift valve movement (via solenoid current control), is there an opportunity for these two to conflict?

You could use MAP or MAf if you want. It just is not really necessary. Many engines today still just use MAP.
The factory uses a lot of data to come to torque calculations. Air, fuel mixture, timing, reciprocating loses, accessory drive loads. Etc. etc,. There are so many things that you could never come up with that you could not accurately measure. But I have built TPS/RPM torque maps via monitoring CAN messages in facory apps and it is within a few lb/ft one way or the other.
I do not know what you mean by shift valve movement. There are no valves that we are regulating a rate of movement for a shift. There are regulated valves for position for a certain pressure. If you are asking if we are doing apply rate then not at this time. But we might. Just more complicated.

The other big one is we are on the factory CAN bus as an option. So this can be installed into a factory car as a replacement for the Factory tcm. This is only on ME 2.8 ecus and up. So pretty much any car from 2002-2003 up to the end of production. The ME CAN ids for 312 etc.
But this takes us from the early 2000s to the end of facory opted 722.6 cars. As well as the Dodge cars.

Is a sentence mission here between the 312 and "early 2000s" ? This stops making sense.

Also, the .6 was used in a Jaguar or two and I remember a Jaguar message board posting about tcu issues so they might be interested in this conversion.

When I read this quote in the opening post:
"and then also from the torque we target a desired shift time. The shift time dictates amperage for the ratio change part of the shift.
We end up with a two stage pressure control for each shift. Much more refined and a lot more control of the more difficult shifts.", I took it to mean that the current thru the shift valve solenoids was being modulated to change the apply rate.

Thanks for the clarification; the feedback is much appreciated.

Early cars used a different CAN message structure. Ie 1996-2000 ish. I have seen some variance. They did not use the id 312. Later cars did. So if you knew the messaging structures in the cars it would make more sense. It is hard to narrow it down exactly so some CAN sniffing would be best if you did not know for sure which system you had.
Dodge used basically the identical CAN structure. So there are no changes necessary to run between Mercedes and Dodge.
Jag and Porsche are different setups. I have no idea what they used for the CAN. But I have no intention of changing any of it for a small number of people.

Frank,
So sort of yes. We regulate pressure to control the firmness and speed of a shift. What we are doing now is targeting a pressure//amperage. What we might switch to is a pressure ramp rate. So as the shift begins from the fill pressure and ratio begins to change we will switch to a ramp rate over time to complete the shift. The steeper the ramp the faster the shift. The way we are doing it now is simply targeting a 2nd fixed amperage value to target a shift time. It works very well and is a little less complicated. The 2nd pressure stage though is controlled via a desired shift time. So we are targeting a shift time or clutch apply rate.

How long until you guys come out with a 722.9 7g-tronic controller?

What he said ^^^^?

__________________
To see my 129 parts for sale visit:
http://stores.ebay.com/The-Mercedes-SL-Store
John Roncallo

The 722.9 may get done some day but not in the same way as the 722.6. The .9 has a mechatronic tcu embedded in the transmission. So it really cannot get replaced. So what we do instead is make the factory tcu operate as though it is in a stock vehicle. The problem with Mercedes on this is it is part of the theft system. Also this requires a precise torque map of the engine to operate correctly. So swaps to cars that do not use a factory tq map are difficult to get right. Not something that we really sell out to the public but rather do in house. A similar project to a 6l80e swap with GM.

I wouldn't wish that trans on anyone. Why anyone would willfully swap for it is beyond me. Mercedes can barely make that thing shift reliably and consistently. Trying to tune it in a standalone environment just sounds like a supreme exercise in frustration.

__________________

90 300TE 4-M
Turbo 103, T3/T04E 50 trim
T04B cover .60 AR
Stage 3 turbine .63 AR
A2W I/C, 40 LB/HR
MS2E, 60-2 Direct Coil Control
3" Exh, AEM W/B O2
Underdrive Alt. and P/S Pulleys,
Vented Rear Discs, .034 Booster.
3.07 diffs 1st Gear Start

90 300CE 104.980
Milled & ported head, 10.3:1 compression
197° intake cam w/20° advancer
Tuned CIS ECU
4° ignition advance
PCS TCM2000, built 722.6
600W networked suction fan
Sportline sway bars
V8 rear subframe, Quaife ATB 3.06 diff