722.6 build - Page 5 - PeachParts Mercedes-Benz Forum

duxthe1 · #61 01-22-2016, 08:16 PM

Not just any PCS tcm-2000 but a tcm-2000 specificly prepped for the 722.6 by Sound German Automotive. Whipplem104 has spent a tremendous amount of time and effort with PCS developing their controller for the 722.6 and AFAIK is the only source for it. Whether its respect for forum rules or just modesty, he doesn't much toot his own horn about it. As a highly satisfied customer I'll take the opportunity to plug it for him. It's a great product and his support after the sale has been top notch.

whipplem104 · #62 01-22-2016, 10:14 PM

Thank you,
I will say that it is only as good as the end user. That, I have learned from the years selling these. I could set them up to bang shift and lock out the tune but that would defeat the point of the system to me. The adaptability and to work with just about any setup imaginable is the best thing. So far I have sold to people behind military diesel snow cat personnel transporter to and Ultra 4 ls3 jeep. My latest prospect is a LMP style car with a Mercedes E55 drivetrain. But it is the daily driver's that is the most fun because those are the people really using the system to it's fullest potential.

300SD81 · #63 02-12-2016, 12:56 PM

Quote:

Originally Posted by duxthe1

Not just any PCS tcm-2000 but a tcm-2000 specificly prepped for the 722.6 by Sound German Automotive. Whipplem104 has spent a tremendous amount of time and effort with PCS developing their controller for the 722.6 and AFAIK is the only source for it. Whether its respect for forum rules or just modesty, he doesn't much toot his own horn about it. As a highly satisfied customer I'll take the opportunity to plug it for him. It's a great product and his support after the sale has been top notch.

Have to agree here, excellent product and very helpful. Would recommend it to anyone.

duxthe1 · #64 05-21-2016, 01:04 AM

Thought I'd throw out an update on this one. I've got about 5k mi of daily driving on this trans now.

I haven't had the laptop on the car for over a month so a couple days ago I decided to datalog a few drives ands try to fine tune the last few quirks. The two biggest things were the 1-2 upshift under light load and 3-4 upshift under moderate to heavy load.

The 1-2 was too firm of a shift under light load esp. given the mechanical advantage of low gear. I noticed in the datalog that it was taking .8 seconds for the RPM to start dropping after the shift was commanded. Looking at the hydraulic flow of the trans it appeared that I didn't have enough line pressure to stroke the command valve in a timely fashion. Adding line pressure got the shift starting in .4 seconds with a much better feel. Still a bit firmer than I'd like in about 10 % of the 1-2 upshifts but even then just a bit firm no longer jerking.

The 3-4 upshift has only ever had a good feel under light load and coasting upshifts. 3-4 is by far the most difficult shift to nail down in this trans. Under moderate load it flares about 200 RPM then bangs the crap out of 4th. The datalog of this behavior doesn't point to a clear resolution. I had tried just about every combination of line and accumulator pressures shy of running the line pressure off the high end of the scale. So I gave that a try along with pushing the accumulator pressure almost as far. Sonofa***** if that didn't finally give me a quick and smooth 3-4 upshift.

The only other issue I've had has been the 2-1 downshift. The problem here is that the downshift occurs before engine RPM can climb enough to match the lower gear. This gives a real herky-jerky feel to the shift. The workaround is to tune so that the 2-1 shift is locked out above about 12-13mph. In trying to tune the 1-2 I noted that when accumulator pressure is at a minimum the shift is delayed enough that the RPM can climb to match the lower gear. I've been considering using an input to skew the accumulator pressure during this shift but this isn't a pressing issue.

All in all I'd call this tune 99.9%. The objective of a custom install is seamless integration. Having nailed down the tune solid I feel like the only indication of a custom installation is the 5 speed shifter gate and a count of the upshifts.

Technicals aside, I have to say this is an awesome upgrade. It makes me love the car even more that I already did. 5th gear transforms the car. Not to mention it will now hold a ridiculous amount to power. I'd do it again in heartbeat.

duxthe1 · #65 12-25-2016, 10:26 PM

I did an oil change on the coupe this past Friday, which marks 10K miles on this trans. I checked the fluid and it still looks brand new. The level is still spot on and no leaks what so ever. It's been pretty darned solid. I don't remember when I last had the tuning laptop on it. Occasionally I'll have a shift that maybe isn't perfect but I'm done chasing the current tune. I'm not going to mess with the tune until I add a mass airflow sensor to the car and use that as my main load indicator. Any shortcomings in the current tune are artifacts of the limitations of a MAP based tune, especially considering I'm at altitude with a low vacuum engine. I'm in no hurry to re-engineer my induction system to include an air flow sensor, though.

whipplem104 · #66 12-26-2016, 11:40 PM

Definitely give some feed back on using MAF. I am interested in your results. I wired one up on my diesel a while back but it did not really give me the results I was looking for. I think on a gas engine it would be pretty great though.

duxthe1 · #67 12-27-2016, 01:26 AM

Considering that my MAP goes to ambient pressure between 15 and 20% throttle, I'd expect to have a much more representative indicator of load using MAF. Its a testament of my willpower that I tuned MAP into as good of a tune as I did. While I'm my own worst critic, the MAP tune is very good. I catch myself driving 722.3 cars and comparing shifts to my car and almost always I judge my car's shifts the better. I do that with 722.6 cars, too, but its less often that I judge mine the better. It does happen but without having the engine ECU give a shift cut its hard to match OE shift quality.

duxthe1 · #68 07-09-2017, 10:18 PM

Well the car this trans was installed in met its end, but another has taken its place. The transmission and controller + harness were pulled from the red coupe this past week. I've started installing the controller into the blue coupe in preparation for the trans, in the near future.

I'm planning to change things up a bit on this install. For whatever reason I had it in my mind that I would not be able to use the airflow potentiometer signal as a load indicator. I recently measured that signal while deflecting the plate and contrary to my assumption, it is a relatively linear 0-5V signal. This should provide a much better load indicator than a MAP signal. Consequently I expect an even better tune than the previous installation

duxthe1 · #69 07-31-2017, 11:31 PM

Finally installed the transmission in the blue coupe. 3.5 hours to swap the trans, cooler lines, flywheel, shifter, speedo, and wire up the lockout, reverse light, and shifter illumination. All other signals were set up prior to the install.

Since I'm now using the airflow potentiometer as my primary load signal I just took my old MAP tables and blended them for what I expected the difference would be. I can say that it is pretty close already with only the four miles driven home from the shop. I can tell already that my 1-2 is much better modulated in the lower load ranges. 3-4 can use a little work but that shift is the most difficult to tune in this trans.

1-2 downshift needs to be played with also. I need to figure a way to dump accumulator pressure on that downshift. It grabs 1st before engine RPMs come up and its pretty herky jerky. With the better load resolution I'm hoping to improve on it. I had previously blocked that downshift after about 12 MPH but I'm really wanting to be able to wind this thing out on the downshift.

duxthe1 · #70 01-04-2018, 01:34 AM

Thought it time to update this thread. I've recently taken my painstakingly developed tune and tossed it out the window! What!?! You read this right, the very good performing tune is out. Take a breath, because I'm test mule-ing for Whipple104 who is working with PCS developing the next generation software for this controller. A couple weeks ago I was given the opportunity to flash my TCM2000 controller to a beta version firmware that uses most of the tuning/operating strategies from the TCM2800 controller. The software is still very raw but the tunability is mind blowing.

The first major change is that the pressures are now based off of a torque map for the engine. This map uses throttle position v/s RPM to model engine torque. I had always thought that the pervious software lacked an RPM component to the shift pressure tuning. Bam!, torque map in your face. Admittedly my map is essentially based off of a little info and a lot of educated guess but should be pretty close to actual engine output.

Next major change is two stage pressure control during the shift. You have modulator and accumulator pressures to control how the off going clutch releases that then switches to a second stage separately controlling those pressures during the application of the on coming clutch. This is some really hardcore shift tuning here. Admittedly the learning curve is very steep but now that I've wrapped my head around it, the new tune is coming together exceptionally well. Unfortunately there is no manual, no guide, just a few hints from Whip along the way.

Another major change is separate pressure controls for upshifts, downshifts, and "garage shifts" ie N-R and N-D. Previously, you had to tune for a happy medium between the upshift and the downshift. This was one of the very small holes in my previous tune. No longer do compromises have to be made in this regard.

In addition to all of the new stuff, all of the important old stuff like programmable PWM outputs is still there. Those settings ported right over no problem. Cal A and Cal B is still present, though I don't use the function much. A new feature I'm using it is being able to select Cal B when I drop to 4 from D. CAN integration is pretty well developed for this controller now but I'm not using any of those features in my preCAN car. There is also a bunch of new stuff like transbrake control, adaptations, and diagnostics to scratch the surface.

Right now this is very much "in beta" kinda thing. I don't expect it to be released for public consumption until it is very well sorted. I'll apologize for the tease, but rest assured there is some awesome stuff on the horizon.

97 SL320 · #71 01-04-2018, 05:52 PM

Quote:

The first major change is that the pressures are now based off of a torque map for the engine. This map uses throttle position v/s RPM to model engine torque. I had always thought that the pervious software lacked an RPM component to the shift pressure tuning. Bam!, torque map in your face. Admittedly my map is essentially based off of a little info and a lot of educated guess but should be pretty close to actual engine output

.

I wonder if using actual engine load would be better RE: MAF or MAP + throttle position, this would automatically compensate for altitude, increased air filter restriction , engine wear.

In the 80's / 90's GM used a throttle cable and vacuum modulator on the front drive 440T4 while it was full hydraulic shift. There should be a tech manual outlining why this was done.

Quote:

Next major change is two stage pressure control during the shift. You have modulator and accumulator pressures to control how the off going clutch releases that then switches to a second stage separately controlling those pressures during the application of the on coming clutch. This is some really hardcore shift tuning here. Admittedly the learning curve is very steep but now that I've wrapped my head around it, the new tune is coming together exceptionally well. Unfortunately there is no manual, no guide, just a few hints from Whip along the way.

Something to consider on the two stage pressure that would not be so apparent when you were running single stage. If you double fluid pressure, the flow increases by less than double. This will give you a non linear pressure vs shift firmness graph.

Are you looking at fluid temp? Fluid is thicker when cold so valves react slower and orifices appear to be smaller as well.

duxthe1 · #72 01-04-2018, 10:12 PM

My previous tune was MAF based. It was an ok load indicator but was lacking an RPM component. So far using the torque map is working really well. The clutches are transferring torque so tuning the pressures on the basis of torque is a simple elegant solution. I think most modern OE electronic transmissions use this strategy.

The way the two stage pressure works is that during the release phase, the modulator and shift pressure solenoids each have user defined duty cycle value based on torque. These determine how quickly the off going clutch drains and releases. As it physically releases the software detects the ratio change and switches to the apply duty cycle values defined for modulator and shift pressure solenoids. During apply, modulator pressure is based on torque but the shift pressure that actually applies the clutch is defined from another separate table based on the desired time (in ms) for the shift to complete depending on the torque and RPM during the shift.

None of these values are related to the others. You can define any duty cycle for either solenoid for either part of the shift. This means you have to have reasonable values entered for them all to be able to complete a shift. It's not idiot proof at this stage either. For me the steepest part of the learning curve was trying to figure out what is wrong when a shift won't complete. Since there are 4 valves that have to cycle in the valvebody for a shift to complete, its not always immediately obvious what isn't happening properly.

Thankfully the datalogging software tracks all of this which makes tuning relatively simple once you understand how the values interact and how the trans responds. The older strategy of one modulator and one shift pressure value for the whole shift can do a surprisingly good job, but the new strategy is some next level.... stuff

. This is OEM level control. In fact there are some OEM applications that use PCS TCM2800 based controllers, which this software was originally developed for. In a couple weeks of very steep learning curve I have a better shifting trans than I had after a couple years tuning the old setup and I feel that there is still a considerable distance between the present tune and a perfect tune with all that is possible in the new software.

A cool feature that I haven't even begun to play with yet is an inertia modifier. A 50 NM shift at 2000 RPM won't be the same as a 50 NM shift at 4000 RPM due to crankshaft inertia. There is a new table that accounts for it and allows you define a duty cycle modifier. I haven't had any cause to use it yet but when putting the finishing touches on the tune it may find use. These inline 6 crankshafts do have significant mass and store considerable energy when spun hard.

There is temp compensation, to increase pressures in a cold transmission, with cal A and cal B maps. The last few days tuning has been getting that dialed in as its been pretty cold here in the mornings. I typically come home for lunch so that affords me the chance to datalog the same drive twice, once stone cold and then again with a warm trans. I feel I'm close but want to datalog one more blisteringly cold drive to make sure the far end of the map is good.

whipplem104 · #73 01-05-2018, 12:29 AM

Glad you are having fun and not cursing me. Just one correction. This is all currently available on the new tcm-2800. We just did this as a back date sort of for the tcm-2000. As you said it really is hard to justify when you have something that works so well already. And since I no longer sell or work with 2000 units really anymore I will not be building any calibrations for this. I have a full on race car that does no street driving local to me that we will do this on and one other car. All my stuff has been moved over tot he 2800 almost 2 years ago. Everything I am doing now on the 2800 is amperage based and there are a few other differences. So it is just to much work to go back and rebuild all the old cals.

duxthe1 · #74 01-05-2018, 02:00 AM

Cursing.... no. It's been fun but there has been plenty of frustration too... funstration. After really wrapping my head around the new regimen, its gotten a lot more fun and much less frustrating. Taking something that was as good as I could make it and being able to re-engineer it into something much better is pretty gratifying. The amount of new control is stupefying. Yet as early as I am into tuning my setup with it, its already well ahead of where I left off with the 2000 software. Its pretty easy to bang a shift, but to get a consistent smooth but positive shift through different loads and speeds is a significant challenge. The new firmware is a spectacular toolkit designed to meet that challenge. A manual would be nice though.