[rjahl]

Well,

This has been a long time coming, I've been watching the Diesel crew and taking notes for more than a year and I am finally at the point where I have made enough progress to share.

Background, I have an N52 powered car (E85) with the ZF 6HP19 automatic transmission, first generation. So the Alpina flash or the new xHP tools will not help me.

I have upgraded the engine to a 3 Stage Manifold and with the help of others learned how to create my own software / tunes for the MSV70. During the process, my data logs have shown that the Automatic transmission was really letting me down, sort of an embarrassment to BMW.

In the early stages I learned how to get the DME to partially ignore the torque reduction request given during a gear shift. These can cut power by 35% + for nearly 3/4 of a second or longer. This helped but really did not solve the problems, clearly a transmission tune is needed.

More than a year ago, I started watching the diesel members create a thread called "Transmission Remap, let's do it ourselves". Fantastic effort from several members to log and learn how their transmissions worked. The project turned into Xhp Tuning and it looks like a really promising and well priced product for all second generation ZF 6 speed gearboxes. Unfortunately, the first generation 6 speeds have been left out and frankly from a commercial perspective, I can't blame them.

However, I did not come away empty handed. I've been watching, learning and gathering notes all this time. Sort of an electronic pack rat. Now thanks to the RSA private key contributions provided by Tarraphatm, I am able to produce my own transmission tune and sign it with the correct RSA signature, convert to an 0da calibration file and flash into the transmission using a standard D can cable and WinfKp. No Boot sector coding or backup nonsense, just a 3 minute flash. Very similar process as how I've been writing tunes for the DME.


My goals with this tune is the following:

1. Reduce the time spent under limited engine output, torque reduction
2. Increase Engine RPM limits to 7,100-7,200 RPM
3. Increase the firmness of the high loaded shifts
4. Activate the gear display in the Combi for all modes
5. Turn off full throttle gear kick-down when in the manual mode.

The project begins.

[rjahl]

The following is an example of the manual mode shift map.

Notice the shifting patters 1>2 and 2<1 are based on OSS (Output Shaft Speed) and throttle position. I suspect the 110 percent is the kick down switch.

To make this easier to read I created another chart converting this to engine RPM using the reduction ratios in the gearbox. This are not very accurate as the torque converter will always provide some slippage.

Notice the down shift columns, these are the minimum engine RPM the gearbox will allow before shifting to the next lower gear. In manual mode if will never shift to first without being commanded.

The column at 110 percent is the one I'm going to target to eliminate the kick down switch in manual mode. It annoys me to be in manual and have the gearbox downshift when I push on the throttle too hard.

This is just one of about 18 shift maps found. There is economy modes, towing modes, downhill, etc.

Also last thing to think about on this post is the throttle position is the driving factor on these maps, not engine output. Therefor anything you do that change the engine output for a given throttle position, will affect these maps. Not this specific manual map but the ones we drive on most of the time.

[rjahl]

Originally Posted by BimmerGeeks

Here we go...I can see this thread becoming the next "I cloned my msv70". Haha.

Keep it up!

Only if the same caliber of people are interested and contribute.

I took two swings and the gear display in the Kombi today and the both failed. I still have a few other possible switches to test.


One of the hard parts is identifying the maps in the calibration files. I have found most of the interesting maps but sorting them will be tough. The manual shift map in my example was easy as the shift point were obvious. I think Testo is going to be a valuable tool again.

With no roadmap / damos it's going to be challenging.

[Eagle1oh7]

So do you have a test car/engine-tranny so if the tranny blows up you're not left without a car?? I have a GM tranny, but it's still very interesting!

[Terraphantm]

Originally Posted by rjahl

Only if the same caliber of people are interested and contribute.

I took two swings and the gear display in the Kombi today and the both failed. I still have a few other possible switches to test.


One of the hard parts is identifying the maps in the calibration files. I have found most of the interesting maps but sorting them will be tough. The manual shift map in my example was easy as the shift point were obvious. I think Testo is going to be a valuable tool again.

With no roadmap / damos it's going to be challenging.

Kombi gear display might not work in an e85.

[rjahl]

Originally Posted by Eagle1oh7

So do you have a test car/engine-tranny so if the tranny blows up you're not left without a car?? I have a GM tranny, but it's still very interesting!

I don't plan on killing my transmission but if I do, I can live with it. This is my daily driver but I can afford to fix it or buy a replacement. I should have traded this car last year but I'm having too much fun with it.

Wife had been pushing me to buy another lexus for years now.

[rjahl]

Originally Posted by Terraphantm

Kombi gear display might not work in an e85.

Please explain.

[Terraphantm]

Originally Posted by rjahl

Please explain.

Electrically, E85 is basically an E46. CAN messages are completely different than what you find on the E6x/E9x (BN2000). I would not be surprised if those Kombi gear display functions only worked in the BN2000 cars. I've never seen any of the cars with the old format have a similar readout, even with Alpina flashes.

[rjahl]

Originally Posted by BimmerGeeks

If you have what you thought should have worked for the gear display, I'm glad to test it out for you on my e90. You have my number and email.

I need to look at your calibration file again. I know you are running an e60 file but I can't remember the number.

[Terraphantm]

Probably best to get confirmation from him, but from what I can tell, it would be A7564227.0da

[rjahl]

Here is another example of a shift map, this one I believe is a sport mode version.

Notice this up shift points for 33 and 48 percent throttle and think about what some tuners and modifications will do to the engine output at any given throttle position.

So if your new tuned 3 stage manifold car needs only 33 % throttle for a decent take off instead of the 48% it needed when stock, it's will now shift into second gear some 600 RPMs sooner. Maybe not what you hand in mind when you purchased a tune from AA or BPC.

Again, please note that these RPMs are calculated with zero slip in the torque converter.

[rjahl]

Another interesting detail is the idea that we have a peak torque map in the calibration file. If the transmission is looking at throttle values, what is it doing with torque curves?

Attached shows the map values for the 160KW and 190KW variants. Notice the huge difference in torque between the two below 2000 RPM. I don't think the DME or the engines are that different in those ranges, maybe I'm wrong.

160KW is the 215HP 3.0SI = 325
190KW is the 255HP 3.0SI = 330

[Terraphantm]

Originally Posted by rjahl

Another interesting detail is the idea that we have a peak torque map in the calibration file. If the transmission is looking at throttle values, what is it doing with torque curves?

Attached shows the map values for the 160KW and 190KW variants. Notice the huge difference in torque between the two below 2000 RPM. I don't think the DME or the engines are that different in those ranges, maybe I'm wrong.

160KW is the 215HP 3.0SI = 325
190KW is the 255HP 3.0SI = 330

Are you 100% sure it's throttle rather than percent of maximum torque?

And hard to know how the cars actually vary below 2000RPM since pretty much no one actually dynos that low

[rjahl]

Originally Posted by Terraphantm

Are you 100% sure it's throttle rather than percent of maximum torque?

And hard to know how the cars actually vary below 2000RPM since pretty much no one actually dynos that low

No, i'm not 100% sure its the throttle. Everything I've read from ZF mentioned throttle rather than load, The INPA test pages use throttle and the xHP maps use throttle. Certainly worth checking again. If it is engine load, it's hard to imagine the logic behind the gear kickdown values.

We could look at the toque maps inside the various engine maps for an idea.

[Terraphantm]

Kickdown is a command sent by the DME isn't it?

It very well could be throttle, just seems like an odd choice to use when torque is available too.

[rjahl]

Originally Posted by Terraphantm

Kickdown is a command sent by the DME isn't it?

It very well could be throttle, just seems like an odd choice to use when torque is available too.

This idea surprised me as well and but I could not find any evidence to contradict the throttle theory. Keep in mind that we also have down hill and trailering shift maps to consider. Would you not require throttle input from the driver to properly hold a lower gear?

To my understanding the DME does not have any control of the current gear selection. Have you seen this in the programs?

Those shift maps do show what looks like a clear downshift command at the 110% setting. I have one that forces a 5>4>3 down shift that drove me nuts. At certain times a full throttle command will cause a downshift to 6,000 RPM's and then force another up shift shift at 6,800 RPMs, each with it's own torque reduction and time delay. If you are trying to pull into a faster traffic lane and this happens, you feel like your are driving a four cylinder rice burner and your life is about to end.

3 Stage manifold and engine tone helped with this I now have power at those RPM's, the original engine tune just fell on it's face through this whole process.


I've never logged anything more than 100% load from the DME.

I open to ideas, so I'll look at this again.

[Terraphantm]

I don't think the DME commands any gear changes directly, but as far as I knew, the way kickdown works is the DME looks at the throttle position, if it hits a certain value (which just happens to be the value that you depress the "button" at), it sets the kickdown flag on the CAN-bus. When the transmission sees that flag, it does its thing.

I wish we could decompress the code, dump some ram, and just see what's actually fed into the routines.

[anjuna]

Just stirring the pot here.

In the previous generation cars that I tuned, the throttle pedal was basically a "commanded torque" value for the car, which could be changed based on the engine speed (3d table).

So 100% throttle at 3000rpm would command X torque
while 100% throttle at 6000rpm would command Y torque
This sort of logic could play into the shifting logic in the tables and help explain a lot.

I would think that, since the DME is moving all sorts of stuff dependent on throttle position (VANOS, valvetronic, etc) it would make sense to me that the DME is outputting some other value to trans, whether its load or not.

Remember that VANOS is a HUGE player in emissions and torque at highway speeds.

If i was looking at this, I'd want to consider shift points, hysteresis, and the inputs the trans needs (load vs engine rpm vs gear 3d map)

[rjahl]

OK,

This should clear up a few things regarding the megatronics control.

I did not know about the "Upshift only up to the gear required" feature. I'll have to try that on the way home. Not sure when I would actually use it.

[hassmaschine]

Glad you made some progress here - no reason to leave so many out of the fun.

I haven't seen a kickdown / throttle constant, but that doesn't mean it doesn't exist. I wouldn't be surprised if the throttle position is broadcast over CAN - torque definitely is (for DSC). I never looked to be honest.

In the case of MSx70 and MSx80, there's no 3D throttle mapping. It's not even 2D - it's just a % change per voltage of throttle position. There is a "driver requested torque", but I'm not sure of the interaction between that and the AT computer.

[Terraphantm]

Originally Posted by hassmaschine

Glad you made some progress here - no reason to leave so many out of the fun.

I haven't seen a kickdown / throttle constant, but that doesn't mean it doesn't exist. I wouldn't be surprised if the throttle position is broadcast over CAN - torque definitely is (for DSC). I never looked to be honest.

In the case of MSx70 and MSx80, there's no 3D throttle mapping. It's not even 2D - it's just a % change per voltage of throttle position. There is a "driver requested torque", but I'm not sure of the interaction between that and the AT computer.

Kickdown threshold is c_v_pvs_thd_kd

In non-BN2000 cars, kickdown status is broadcast on ARBID 0x329 (DME2) Byte6 Bit 2 (0 = off, 1 = on)
In BN2000 cars, it's broadcast on ARBID 0xAA (Torque 3) byte 6 I think (0xB = Kickdown on)

[rjahl]

Originally Posted by Terraphantm

Kickdown threshold is c_v_pvs_thd_kd

In non-BN2000 cars, kickdown status is broadcast on ARBID 0x329 (DME2) Byte6 Bit 2 (0 = off, 1 = on)
In BN2000 cars, it's broadcast on ARBID 0xAA (Torque 3) byte 6 I think (0xB = Kickdown on)

Ya, the kick-down switch seems to be something monitored by the DME, At least it's part of the INPA jobs for the DME and not the GKE211.

I've logged the throttle angle and I can not get a 110% reading from anything. Max I see is the 100% all the way up to and past the kick-down switch.

I also used my Dash Commander to log throttle angle, engine load and RPMs on the way home tonight. Seems pretty consistent that a normal or slow take off and run through the gears uses a 25% throttle angle but shows up as a 55% engine load. Each up-shift happens around 2,000 RPM. This appears to follow one of the economy maps based on throttle more than engine load.

Even the kickdowns seem to follow the throttle angle concept as it's really easy to push the DME to near full load with 50% throttle and not force a downshift.

I need to run proper logs to really settle this. I'm doing to much from chicken scratch notes.