[bradsm87]

One of the more challenging ones to pinpoint, but an absolute must for petrol engine owners is finding the ramp rate at the end of gear shift intervention tables. N54s with 6HP have been battling post-shift knock-triggered timing corrections forever. There have been workarounds with varying success, but ultimately, timing advance needs to be re-introduced a little more slowly. It comes in way too fast currently.

These are going to be one or both of the tables with these names:

Dauer Rampe ME bei Schaltungsende - Duration ramp circuit at the end of ME

Dauer Rampe ME bei Ende Gangeinlegen - Duration ramp ME at the end of gear engagement

ME = Motoreingriff, which translates to "engine intervention" when put in to context in other table names.

Unfortunately, there are 26 of these consecutive tables in the VAG 6HP map list, they are all single cell and they mostly share the exact same name with nothing to differentiate what each individual one does. The other issue is I can't find any surrounding maps that are same between the VAG one and ours. I'll have another go tomorrow!

[335dsl]

Can I get a little help please? I would greatly appreciate it!

I did the backup process and purchased the license/stage 3 flash. Once I go back it says 'create a backup'. Do I copy another set of the factory flash on the Backup folder or somewhere else?? (Which means I would have 2 factory flashes on the Backup folder?)

[_TB_]

Anybody have inputs on: "Wandlerverstaerungskennlinie".
Google translates it into: "Converter characteristic curve".

I can see that it differs from the Alpina B3 to the original BMW. Alpina D3 and Original BMW 325d has the same curve - but the B3 differs. Does the B3 run a different converter?

[_TB_]

Anybody have inputs on: "Wandlerverstaerungskennlinie".
Google translates it into: "Converter characteristic curve".

I can see that it differs from the Alpina B3 to the original BMW. Alpina D3 and B3 differ from the Original BMW 325d. Does the B3 and D3 run a different converter?

[bradsm87]

Originally Posted by _TB_

Anybody have inputs on: "Wandlerverstaerungskennlinie".
Google translates it into: "Converter characteristic curve".

I can see that it differs from the Alpina B3 to the original BMW. Alpina D3 and Original BMW 325d has the same curve - but the B3 differs. Does the B3 run a different converter?

Interesting. It's a bugger that Alpina cars don't show up on RealOEM for comparing parts in their cars. It'd be interesting to know.

The DME also has a torque converter characteristic curve table.

[bradsm87]

I think I've found the end of intervention ramp tables. The only way to know which one is which without fancy decompression and disassembly of the code is to test them. I'll test them one at a time with logs by putting extreme values in and seeing which gear shift they apply to. If it's nothing that I can notice in logs, I'll set it back where it was and move on to the next one. They're 8-bit single cell tables at 60933-6094C.

For a long time with N54s, we've had to make compromised band-aid fixes for knock-induced post-shift timing drops. If these tables do what I hope they do, I think they'll be the best fix yet.

[bradsm87]

I'd like concrete confirmation regarding shiftmap throttle axis. This is the weird thing:

There are 21 shiftmap tables and just before them, 21 consecutive 13x1 throttle axis tables BUT in the XDF, a lot of the axis are shared, but not all. There is no pattern to it (It'd make more sense if all of one category of shiftmap shared an axis, but it appears random). The first 3 shiftmaps in the XDF do have what I'd expect for the axis, the first 3 consecutive axis tables. I would assume this trend should continue.

Does anybody know if the shiftmap throttle axis addresses in the XDF were a result of actual IDA tracing work with the disassembled code?

It would seem more logical to me that each one of the 21 axis tables apply to one of the shiftmap tables, in the order that they are in the rom

[DWR]

Originally Posted by _TB_

Anybody have inputs on: "Wandlerverstaerungskennlinie".
Google translates it into: "Converter characteristic curve".

I can see that it differs from the Alpina B3 to the original BMW. Alpina D3 and Original BMW 325d has the same curve - but the B3 differs. Does the B3 run a different converter?

_TB_, since you have limited access to a full WinOLS, I would like to request that you included screen shots of the map, scalar, etc and the surrounding structure (location) in which it is found, when possible.

My understanding is that the converters are indeed different.

That characteristic curve has influence on parts of the calibration such as torque multiplication of engine torque, stall speed to determine torque reduction during "brake boosting", and more.

[DWR]

Originally Posted by bradsm87

I'd like concrete confirmation regarding shiftmap throttle axis. This is the weird thing:

There are 21 shiftmap tables and just before them, 21 consecutive 13x1 throttle axis tables BUT in the XDF, a lot of the axis are shared, but not all. There is no pattern to it (It'd make more sense if all of one category of shiftmap shared an axis, but it appears random). The first 3 shiftmaps in the XDF do have what I'd expect for the axis, the first 3 consecutive axis tables. I would assume this trend should continue.

Does anybody know if the shiftmap throttle axis addresses in the XDF were a result of actual IDA tracing work with the disassembled code?

It would seem more logical to me that each one of the 21 axis tables apply to one of the shiftmap tables, in the order that they are in the rom

I do know that in initially testing, to determine which map corresponded to which "function", it was not possible to achieve them all by driving. Some disassembly was used, at times, to determine what to log for parameters internal to the code. I think it reasonable to assume that more disassembly work was performed after xHP was formed.

As we sometimes found out, the cal did not always match up with our "logic". Usually it had more to do with our ignorance than a flaw in the cal. Remember, just because there is structure and facility, does not means it always needs to be used. Still, I appreciate your comment and have those same questions. As a practical matter, experimentation can answer the vital questions - it just takes much longer.

[rjahl]

Originally Posted by DWR

I do know that in initially testing, to determine which map corresponded to which "function", it was not possible to achieve them all by driving. Some disassembly was used, at times, to determine what to log for parameters internal to the code. I think it reasonable to assume that more disassembly work was performed after xHP was formed.

As we sometimes found out, the cal did not always match up with our "logic". Usually it had more to do with our ignorance than a flaw in the cal. Remember, just because there is structure and facility, does not means it always needs to be used. Still, I appreciate your comment and have those same questions. As a practical matter, experimentation can answer the vital questions - it just takes much longer.

I wish I was not so slammed at the office this week. The GKE211 is so similar and we are working towards similar goals, I just don't have time to even properly analyze the logs I took this morning.

It does appear that the changes I made to the toque reduction maps had no affect on the target torque values through the gearshift. I did raise the peek RPMs to the targets I was looking for but they are not consistent. I think it depends on how fast the RPMs are raising. I'm close so I feel good about it anyways.

The excel map list for the VAG was great are you planning to post the BMW version?

[bradsm87]

Originally Posted by rjahl

The excel map list for the VAG was great are you planning to post the BMW version?

Here is the BMW 8HP45 one:

https://docs.google.com/spreadsheets...it?usp=sharing

[bradsm87]

Another interesting one. Looks like a "Max pressure per clutch" table. These are all quite a bit lower than the maximum values in the "Modulationsdruck x.Gang". The modulationsdruck tables I would assume are system pressure but this one is probably max pressure per clutch circuit.

[bradsm87]

It appears to me that the maps at the very end are the ones that are used by BMW custom logic and are not in the VAG 6HP OLS. Also I noticed in the BMW 8HP OLS that all of the maps near the end have completely abbreviated names. I'd assume that's because the English and German descriptive names are ZF tables and the abbreviated ones are from BMW.

I would assume BMW would get huge manuals with detailed descriptions for each table from ZF. Does stuff like this get leaked often? I'd think if the DAMOS gets leaked, more detailed descriptions would get leaked also.

[DWR]

Originally Posted by bradsm87

One of the more challenging ones to pinpoint, but an absolute must for petrol engine owners is finding the ramp rate at the end of gear shift intervention tables. N54s with 6HP have been battling post-shift knock-triggered timing corrections forever. There have been workarounds with varying success, but ultimately, timing advance needs to be re-introduced a little more slowly. It comes in way too fast currently.

Is this knock truly post shift? In other words, the engine rpm has completely dropped to the new gear ratio.

If the knock is happening as described, that most likely means the cylinders are experiencing heightened temperature. The inertia phase does put additional load on the engine, especially if there is insufficient reduction of torque. Knock has a load vs rpm relationship. Basically, it becomes less likely as rpm increases. The converse is the situation discussed here.

Sounds like you are trying to extend the spark advance ramp-in to allow cooling time for the cylinders. I wonder if the shifting strategy will allow you to simultaneously shorten the shift and extend the torque reduction control beyond the shift. So many parameters are interdependent in the control strategy, that seems like it might be problematic. Everything I have seen indicates the endpoints of torque reduction are contained within the shift timings. The ramp rates affect the full reduction duration within the total time. But that is all conjecture - there is only one way to find out for sure. Good Luck.

[DWR]

Originally Posted by bradsm87

Another interesting one. Looks like a "Max pressure per clutch" table. These are all quite a bit lower than the maximum values in the "Modulationsdruck x.Gang". The modulationsdruck tables I would assume are system pressure but this one is probably max pressure per clutch circuit.

Axis values?

[bradsm87]

Originally Posted by DWR

Is this knock truly post shift? In other words, the engine rpm has completely dropped to the new gear ratio.

If the knock is happening as described, that most likely means the cylinders are experiencing heightened temperature. The inertia phase does put additional load on the engine, especially if there is insufficient reduction of torque. Knock has a load vs rpm relationship. Basically, it becomes less likely as rpm increases. The converse is the situation discussed here.

Sounds like you are trying to extend the spark advance ramp-in to allow cooling time for the cylinders. I wonder if the shifting strategy will allow you to simultaneously shorten the shift and extend the torque reduction control beyond the shift. So many parameters are dinterdependent in the control strategy, that seems like it might be problematic. Everything I have seen indicates the endpoints of torque reduction are contained within the shift timings. The ramp rates affect the full reduction duration within the total time. But that is all conjecture - there is only one way to find out for sure. Good Luck.

It's more the ram effect when the rpm gets suddenly pulled down by the transmission but the air in all the piping still has heaps of inertia from redlining it. It happens right at the moment the timing gets increased at the end of the shift. It's common with older cars to have audible knock on hard gear changes for this reason (and to a lesser extent, higher cylinder temp left over from the high rpm). People have had some success with advancing the intake cam more than ideal just so it's less influenced by the extra cylinder filling from the ram effect from the fast moving air through all the piping etc. but increasing the torque increase ramp time would be better. I run richer than necessary at very high RPM just to keep it as cool as possible leading up to the gear change. I'm on a phone now but i'll post up some logs showing the issue later.

[bradsm87]

Originally Posted by DWR

Axis values?

It's just 0 1 2 3 4 5 6 7 8 9. The title of the map describes which cell does what.

[Unklejoe]

Hello,

I've merged in the latest XDF changes sent to me by bradsm87. The XDF can be found here.

Also, I've attached a Linux executable version of my 0da/0pa to BIN conversion program.

For those who are interested, you can find the source to the program here. You can compile it yourself via the following command under pretty much any Linux machine:

"gcc -o bmwhextobin bmwhextobin.c ihex_copy.c ihex_parse.c ihex_record.c"

To use the program, use the following syntax:

"./bmwhextobin 0x100000 xxxxx.0da yyyyy.0pa" (replacing xxxxx and yyyyy with your 0da/0pa file names)

You may also omit the 0pa file from the command completely, since it's not really necessary for just comparing maps.

NOTE: I had to add a ".pdf" extension to the executable so I could attach it here. It's not really necessary to remove it before execution, but I would.

Thanks,
- Jake

[DWR]

Originally Posted by bradsm87

It's more the ram effect when the rpm gets suddenly pulled down by the transmission but the air in all the piping still has heaps of inertia from redlining it. It happens right at the moment the timing gets increased at the end of the shift. It's common with older cars to have audible knock on hard gear changes for this reason (and to a lesser extent, higher cylinder temp left over from the high rpm). People have had some success with advancing the intake cam more than ideal just so it's less influenced by the extra cylinder filling from the ram effect from the fast moving air through all the piping etc. but increasing the torque increase ramp time would be better. I run richer than necessary at very high RPM just to keep it as cool as possible leading up to the gear change. I'm on a phone now but i'll post up some logs showing the issue later.

Sorry, I tend to use terms from an engineering perspective. Load increase means air fill increase. If you can log what you have just described, the ECU can usually respond in time.

The start or end point of the timing parameters are defined by speed measurements, directly or indirectly. If we are talking about the TCU, increasing the ramp down rate of torque reduction will delay its onset. That means decreasing the ramp time. The ramp rates of the torque reduction are intended to mirror the increased torque due to ratio change inertia. I would wager it is entirely possible to increase the rate of inertia reduction faster than the engine can respond to the torque reduction. As I understand it, there is a parameter for that too.

[DWR]

Feeling completely incompetent at the moment.
Cannot seem to align the 8HP45 CVS file with the 8HP45 bin.
Anyone know the address offset?
bradsm87, you are too clever not to have figured it out

[bradsm87]

Originally Posted by DWR

Sorry, I tend to use terms from an engineering perspective. Load increase means air fill increase. If you can log what you have just described, the ECU can usually respond in time.

The start or end point of the timing parameters are defined by speed measurements, directly or indirectly. If we are talking about the TCU, increasing the ramp down rate of torque reduction will delay its onset. That means decreasing the ramp time. The ramp rates of the torque reduction are intended to mirror the increased torque due to ratio change inertia. I would wager it is entirely possible to increase the rate of inertia reduction faster than the engine can respond to the torque reduction. As I understand it, there is a parameter for that too.

I definitely don't want to delay or change the ramp down of torque that happens at just the right time. I just want to delay the torque increase at the ent of the torque intervention.

The ECU is very fast and that's not an issue. It's more just x manifold pressure during an in-gear pull has y cylinder filling but x manifold pressure immediately after revs have been pulled down has more than y cylinder filling due to the extra inertia of air. This is just a theory of why it happens but it does happen. It could just be extra cylinder heat rather than inertia. Either of those causes would both benefit from a slower onset of timing/torque after the shift.

Here are a couple of recent multi-gear WOT pulls of mine (on the far left and the far right). The little boost spikes on this particular log aren't the issue. Post-shift timing corrections happen if they're there or not. Ignition timing target is reduced quite a bit as soon as boost is over target anyway. You can see the log point where timing has been restored to normal is the exact logging point where the knock-induced timing corrections happen. I think just need a little slower onset of timing after a shift to stop it:

http://datazap.me/u/bradsm87/rev42-t...5&zoom=424-866

[bradsm87]

Originally Posted by DWR

Feeling completely incompetent at the moment.
Cannot seem to align the 8HP45 CVS file with the 8HP45 bin.
Anyone know the address offset?
bradsm87, you are too clever not to have figured it out

I haven't looked much at the 8HP one because the VAG 6HP one seems to have big amounts of exact same tables in the exact same sequence, sometimes in big chunks of many tables before a difference pops up. The tables that aren't in the VAG 6HP one like Shiftmaps, I've looked for in the 8HP map list but they're not there either. It seems the VAG 6HP one is by far the most useful and shares most of the tables with ours. I'll prob just wait till we get access to full WinOLS to play more with the 8HP stuff.