[dave205t]

Originally Posted by Mik325tds

After 1 year, 3 months and 26 days I'm incredibly proud to announce the first US based 335d is driving an Alpina B3 cal!
We are now able to read data from our TCU, modify the calibration section and write it back to the TCU.
Thank you to all that have contributed to this effort. But it wouldn't have been possible without the contribution of dave205t! I cannot tell you how thankful I am for your hard work!
Cheers!

It never hurts to help

[Iaro]

Great work

[lnxguy]

Originally Posted by Whitbread

Wohoooooooooo!!!!! Where are we all meeting for celebratory beers?

Can we makes those Bells Kalamazoo stouts?

Nice to see you posting over here, looks like some of the TDI guys have big boy cars now, lol me included.

[Nadir Point]

Originally Posted by Mik325tds

...shift to 3rd is a bit harsh but otherwise pretty quick shifts in D mode. Note the Dx in the dashboard.

... rev matching not working properly. I think they work quite well. Too bad it's only in manual...

One lightly annoying behavior is that from standstill letting go of the brake will introduce a little jolt followed by small drive train oscillations. I didn't notice it that pronounced with the other cals. In DS mode that is completely smooth.

It sounds to me what you are saying is it mostly doesn't run right. Are you going to continue driving it like that? I'd be curious to see how long it lasts.

[iaknown]

We can see how valuable dave205t has been in this endeavor, his rep count has just surpassed his post count

[Mik325tds]

Originally Posted by UberArchetype

It sounds to me what you are saying is it mostly doesn't run right. Are you going to continue driving it like that? I'd be curious to see how long it lasts.

You are correct. I don't think the Alpina B3 cal is going to be the solution. I'm going to see how it changes as the adaptations move. The shift into 3rd has gotten better but it is still noticeably stronger than it should be.
Most interestingly, the adaptations are now completely different than they have been before. Not sure what to make out of that.
I'm thinking about going back to the A7610591 cal and play around with the D-shiftmap that DWR identified.

[Nadir Point]

Originally Posted by iaknown

We can see how valuable dave205t has been in this endeavor, his rep count has just surpassed his post count

What was the "endeavor," again? Looks like somebody unlocked a "map" of the entire U.S. with nothing but secondary roads on it. Now the task at hand is figuring out the best way to get from NYC to LA.

Good luck mapping that.

[Poosik]

one step at a time

[DWR]

Originally Posted by UberArchetype

What was the "endeavor," again? Looks like somebody unlocked a "map" of the entire U.S. with nothing but secondary roads on it. Now the task at hand is figuring out the best way to get from NYC to LA.

Good luck mapping that.

Not everyone is looking to get to the same destination. It is actually unlikely that this effort will not benefit someone. That would seem to be reason enough to try.

[DWR]

Originally Posted by Mik325tds

Most interestingly, the adaptations are now completely different than they have been before. Not sure what to make out of that.

I think we can make the analogy with engine related adaptions. Adaptions are made in the presence of errors between a desired setpoint and the actual. For example, ECUs have had short and long term fuel trims for decades. By looking at the magnitude and direction of AFR error, tuners can adjust maps to improve the tune.

Similarly, understanding the adaptions being made by the transmission should point the way to altering maps and scalars. Mik325tds, can you share the differences you are seeing?

[Mik325tds]

Originally Posted by DWR

I think we can make the analogy with engine related adaptions. Adaptions are made in the presence of errors between a desired setpoint and the actual. For example, ECUs have had short and long term fuel trims for decades. By looking at the magnitude and direction of AFR error, tuners can adjust maps to improve the tune.

Similarly, understanding the adaptions being made by the transmission should point the way to altering maps and scalars. Mik325tds, can you share the differences you are seeing?

Sure. The first one is with the A7610591.0da after completing (most of) an adaptation drive cycle as described in the attached .pdf + a little bit of driving. The second one is after completing (most of) the adaptation drive cycle with the Alpina B3 cal.

The question that comes to mind is "Where are the desired values stored and what are they"? If those values are stored in the application part of the program then it would make sense - the adaptation is obviously trying to slow down some of the shifts. But why would they? To me, adaptation targets should be part of the cal as well.

[Mik325tds]

Here is a log I took a few days ago in comparison to earlier logs of the A7610591.0da cal. I'm not sure why I'm getting so slow data updates through test-o even though the data rate (time stamps) is almost double of what I had before.
However, it looks like the shift 1>2 starts earlier while the shift 2>3 then is about at the same rpm. But we can also see how abrupt the 2>3 shift is.

[DWR]

Originally Posted by Mik325tds

The question that comes to mind is "Where are the desired values stored and what are they"? If those values are stored in the application part of the program then it would make sense - the adaptation is obviously trying to slow down some of the shifts. But why would they? To me, adaptation targets should be part of the cal as well.

So, let's see if we can come to concensus on the purpose of adaptions. The use of adaption in passenger vehicles really started as a means to meet emissions requirements throughout a predetermined duration, typically mileage. It comes in 2 forms: adjustment to production tolerances and adjustment to aging/wear. Those same two reasons apply to the electronically shifted transmission.

The way ZF6HP transmissions shifts gears, a clutch is released synchronous with a clutch apply. This requires precision timing of events within the transmission, that cannot be accomplished with a static calibration. The TCU is essentially self tuning. Just like an O2 sensor gives feedback for an ECU to correct AFR. The TCU monitors input and output speeds of the transmission to look for bind up or flare during shifts. It also has shifting speed targets. All that information forms the basis for corrections, also know as adaptions.

Have you compared Alpina B3 adaptions with RayBan81? If a cal is put into the transmission that was originally for a different engine/vehicle combination, I think it is very likely the adaptions would be quite different. Obviously, the magnitude of the variance is what is interesting.

This is the first time seeing the adaption values in ISTA (I have to get mine back up and working again, perhaps I'll pay you a visit soon and get some help ). The lack of adaptions in the Alpina B3 cal, along with 2 large deviations, is most surprising. I'll have to refresh my memory on the clutch sequences to see if the characteristics you described are in line with the adaptive corrections.

Thanks for sharing!

[Mik325tds]

Originally Posted by DWR

So, let's see if we can come to concensus on the purpose of adaptions. The use of adaption in passenger vehicles really started as a means to meet emissions requirements throughout a predetermined duration, typically mileage. It comes in 2 forms: adjustment to production tolerances and adjustment to aging/wear. Those same two reasons apply to the electronically shifted transmission.

The way ZF6HP transmissions shifts gears, a clutch is released synchronous with a clutch apply. This requires precision timing of events within the transmission, that cannot be accomplished with a static calibration. The TCU is essentially self tuning. Just like an O2 sensor gives feedback for an ECU to correct AFR. The TCU monitors input and output speeds of the transmission to look for bind up or flare during shifts. It also has shifting speed targets. All that information forms the basis for corrections, also know as adaptions.

Have you compared Alpina B3 adaptions with RayBan81? If a cal is put into the transmission that was originally for a different engine/vehicle combination, I think it is very likely the adaptions would be quite different. Obviously, the magnitude of the variance is what is interesting.

This is the first time seeing the adaption values in ISTA (I have to get mine back up and working again, perhaps I'll pay you a visit soon and get some help ). The lack of adaptions in the Alpina B3 cal, along with 2 large deviations, is most surprising. I'll have to refresh my memory on the clutch sequences to see if the characteristics you described are in line with the adaptive corrections.

Thanks for sharing!

I completely agree on why there are adaptations and what they do. Maybe I didn't express myself correctly before. My question was, if there is a speed target for a shift and a target value on what is considered a bind or flare. Where would those values be stored? Application of Calibration section? But it would be interesting to see how RayBan81 adaptation values look. It would also be interesting to see if another 335d would behave the same with the Alpina B3 cal.

[DWR]

Originally Posted by Mik325tds

The question that comes to mind is "Where are the desired values stored and what are they"? If those values are stored in the application part of the program then it would make sense - the adaptation is obviously trying to slow down some of the shifts. But why would they? To me, adaptation targets should be part of the cal as well.

The fill pressure and fill times are targets within the cal. Ideally there would be no adjustment, but that is just impractical/unrealistic.

The EDS solenoids control fluid flow. Therefore, they control fill and release times. One of the differences in the Gen 1 vs Gen 2 ZF6HP is the solenoids. The Gen 2 shifts faster because it has higher flowing EDS solenoids to targeted clutches. Interestingly, Ford has various bands of flow designated to EDS solenoids. In the graph below, we can see ZF uses the fastest solenoids. The flow/speed is controlled via PWM activation, so the ZF strategy gives the most flexibility of control.


Pressure controls the firmness of the shift. It is modulated on 2 levels. First, the overall main line pressure tends to increase with engine torque, to increase holding power of clutches. Second, because there are fluid accumulators of some sort for every clutch, the flow of fluid controls the rise and decay of local pressure. It is a little more complicated than that as there are pressure targets that initiate other flows - but let's keep it simple for now

The B3 is a cal intented for an trans/eng combo that has a different torque curve. That means the pressures could be off. If that is true, that also means the flows will be off. And the transmission will try to compensate via adaption. Here's a comparison of engine outputs:

Standard 335d remapped
BHP = 350
Torque = 540

Alpina B3 BT
BHP = 360
Torque = 368

Standard 335i remapped
BHP = 348
Torque = 360

So, we can see why a B3 cal in a remapped 335i works very nicely. It might even work in a stock 335d.

Next post I'll talk about the inertia phase of the shift ... think the diesel engine has different inertia than the gas engine? Me too.

[DWR]

Originally Posted by Mik325tds

My question was, if there is a speed target for a shift and a target value on what is considered a bind or flare. Where would those values be stored? Application of Calibration section?

I was kind of reinterpreting your question. Just wanted to bring others along in the conversation.

Can you identify such a section? I think that's why in our private communications we agree that some sort A2L/Damos would go a long way. We may need some $ donations to buy such a thing.

Looking for scalars in a large cal file is beyond my capabilities. I say that and then immediately I want to ask if you can get me a 335i cal that matches cronologically with the B3 cal. Lol, never say never.

[rjahl]

Originally Posted by DWR

I was kind of reinterpreting your question. Just wanted to bring others along in the conversation.

Can you identify such a section? I think that's why in our private communications we agree that some sort A2L/Damos would go a long way. We may need some $ donations to buy such a thing.

Looking for scalars in a large cal file is beyond my capabilities. I say that and then immediately I want to ask if you can get me a 335i cal that matches cronologically with the B3 cal. Lol, never say never.

Have you started to map the calibration file yet? With my 6HP19, I'm feeling a little left out and I'd like to take a look the various calibration files and see if can start to make any sense of them. Do you have anything you can share? I'm hoping many of the map definitions will be similar.

This is in anticipation that we get support for the 6HP19.

[DWR]

Originally Posted by rjahl

Have you started to map the calibration file yet? With my 6HP19, I'm feeling a little left out and I'd like to take a look the various calibration files and see if can start to make any sense of them. Do you have anything you can share? I'm hoping many of the map definitions will be similar.

This is in anticipation that we get support for the 6HP19.

Well, we have located many maps. Even think we know what some of them are. We just got so that someone can flash a modded file. That is to say, no, we aren't sure what does what. I think that is what you are asking.

This will take some time. Changing shift points is going to be more straight forward. Changing shift feel and speed is going to be much more difficult without some sort of "template".

Files are too big to attach here. I'll send some to you.

[DWR]

Ok, let's talk about the 3 phases of a synchronous shift, as they occur in the ZF6HP.

By definition, a shift is a change in the relation between engine speed and transmission output speed (and internal speeds). What is know as the torque path changes (sometimes called the power path, but that is somewhat misleading). Because of the change in engine speed that accompanies this torque flow path transition, such shifts include a torque phase and an inertia phase.

I've removed my previous description of how that works because I wanted everyone to get it from the horses mouth (no jokes, please ). Here's a description from a Ford patent that would pertain to the 6R80, Ford's licensed version of the ZF6HP26.:

In the case of a synchronous upshift, a first torque establishing element, referred to as an off-going clutch (OGC), is released while a second torque establishing element, referred to as an on-coming clutch (OCC), is engaged to lower a transmission gear ratio and change the torque flow path through the transmission. A typical upshift event is divided into a preparatory phase, a torque phase, and an inertia phase. During the preparatory phase, the OCC is stroked to prepare for its engagement while the OGC torque-holding capacity is reduced as a step toward its release. During the torque phase, which may be referred to as a torque transfer phase, the OGC torque is reduced toward a value of zero or a non-significant level to prepare it for disengagement. Simultaneously, the OCC torque is raised from a non-significant level, thereby initiating engagement of the OCC according to a conventional upshift control strategy. The timing of the OCC engagement and the OGC disengagement results in a momentary activation of two torque flow paths through the gearing, thereby causing torque delivery to drop momentarily at the transmission output shaft. This condition, which can be referred to as a “torque hole,” occurs before disengagement of the OGC. A vehicle occupant can perceive a “torque hole” as an unpleasant shift shock. When the OCC develops enough torque, the OGC is released, marking the end of the torque phase and the beginning of the inertia phase. During the inertia phase, the OCC torque is adjusted to reduce its slip speed toward zero. When the OCC slip speed reaches zero, the shift event is completed.

The rotating mass of the M57 is much greater than an N54. That is an inertial effect that is not part of the B3 cal. That will tend to effect the down shift more than the upshift. In an upshift, the vehicle is the mass resisting motion. That is relatively close for either application. In an accelerating downshift, the massive torque of the M57 compensates for the increased rotating mass. In a braking downshift, no such compensation exists.

[Mik325tds]

Hah. I've been reading pages 9 and 10 yesterday to find out where we left off before the RSA problem.
We have a pretty good candidate (see below) that may be the D-XE shift map since it closely matches the downshift behavior under 0% throttle.
The strategy to confirm this was to modify the the downshift 6>5 to be a constant value so it could only be triggered by a specific speed threshold and not via throttle input.
I modified my binary accordingly and corrected the checksums with Dave's tool. It is ready to go in the car but I want to drive the Alpina B3 cal a little longer to see if the adaptations change any more.

[Mik325tds]

Originally Posted by DWR

Ok, onto the interia part of the shift.

By definition, a shift is a change in the relation between engine speed and transmission output speed (and internal speeds). What is know as the torque path changes (sometimes called the power path, but that is somewhat misleading). Because of the change in engine speed that accompanies this torque flow path transition, such shifts include a torque phase and an inertia phase. To make the shift smooth, the torque phase is completed quickly by boosting pressure at the beginning of the shift. Just before the torque phase is completed, the pressure is lowered so the inertia phase bump is reduced. There is also a pre-boost phase to get initial engagement of the clutch.

The rotating mass of the M57 is much greater than an N54. Fortunately, massive torque tends to be an equalizer for acceleration of the rotating mass. There is no such compensation on deceleration. That is why the mismatch in down shifts will be more noticeable than upshifts, for the B3 cal.

Wow. How do you know all that? Have you been working for transmission manufacturer before?
That completely makes sense!

[DWR]

Originally Posted by Mik325tds

Wow. How do you know all that? Have you been working for transmission manufacturer before?
That completely makes sense!

Ain't my first rodeo.

Way back, after they had just invented the wheel, I was modifiying the valve bodies of Ford automatics. But they worked different then. It was easier.

Isn't that why we initially teamed up? I'm seasoned and your smart.