6HP Custom Tuning

[bradsm87]

OK here is my grasp of solenoids used for shifts and pressure limitations. If somebody knows otherwise, please correct me.

- Rising yellow EDS solenoids apply more pressure as more current is applied. It is theoretically possible to apply more current than 4.6bar and get a little more than 4.6bar. Exceeding the rating of the solenoid does not sound like something that BMW would do though.

- Falling blue EDS solenoids start at 4.6bar when no current is applied. I doubt very, very much that the TCU could apply reverse polarity to theoretically get a little more out of it. I think it's safe to assume that 4.6 bar is the limit for these solenoids.

Solenoid types used:
1>2 offgoing - falling
1>2 oncoming - rising
2>3 offgoing - rising
2>3 oncoming - falling
3>4 offgoing - falling
3>4 oncoming - falling
4>5 offgoing - rising
4>5 oncoming - falling
5>6 offgoing - falling
5>6 oncoming - rising

This is what does not make sense:
In the stock 335i calibration, with the published XDF's conversion formula in place, offgoing 1>2 commands up to 5.6bar and 3>4 5.15 bar. Also in the Alpina B3 calibration, the 3>4 offgoing commands 5.1bar. These ALL are commands that use falling type solenoids that absolutely max out at 4.6bar! Surely BMW would not put unachievable values in tables!

@RayBan Is there a possibility that the conversion formula in the XDF of 0.001*X is incorrect?

 

[mmm335]

Reading all this is pretty amazing and interesting, yas are doing a hectic job

 

[bradsm87]

For anyone who is interested in custom tuning or getting a custom tune in the future, Justin at Twisted Tuning has generously stepped in and purchased the BMW 8HP and VAG 6HP WinOLS files and he'll be happy to distribute them to any other 6HP table discovery enthusiasts if he can recoup some of the cost from some donations. These files will hopefully hold some clues to finding more tables. Somewhere around US$30-50 is a good place to start IMO . If you can, throw something through to him as "send money to family and friends" so PayPal doesn't rip us. His PayPal ID is [email protected] . He is in the USA (PayPal asks for the country you're sending to). Just put on the notes that it's a TCU OLS file donation. Every bit helps.
Cheers guys!

 

[doublespaces]

Here is an updated XDF with TunerPro bug fixes

 

[bradsm87]

Here is an updated XDF with TunerPro bug fixes

I've merged my breakpoint tables and shiftmap column labels in to this latest revision from XHP and sent it back to Clemens at XHP to hopefully publish on his site and keep in his live XDF moving forwards.

 

[doublespaces]

I've merged my breakpoint tables and shiftmap column labels in to this latest revision from XHP and sent it back to Clemens at XHP to hopefully publish on his site and keep in his live XDF moving forwards.

Are these tables you've discovered yourself which you're trying to get pulled in?

 

[bradsm87]

Are these tables you've discovered yourself which you're trying to get pulled in?

Nah no newly discovered tables. Just defining the axis breakpoints for the existing tables so that they can be changed in TunerPro eg. all stock ongoing clutch shift pressure tables end at 450Nm torque so those figures need to be raised. There is no means to do so with the published XDF.

The other change is I've labelled the columns in the shiftmap tables appropriately. Instead of trying to remember what column 1,2,3,4,5 etc means, the columns are simply labelled by the shift that they are for eg. 1>2, 2>1, 2>3, 3>2 etc.

 

[przemek4net]

Hi,
I'm after 1st mode (N57 245ps e91 330xd after strong mods /hybrid T, big IC, no DPF/) and before correction but for this I have to collect a little more information.
There are several maps and I'm not sure about the sense of these maps
We have:
Shiftmaps
Y: throttle
X 1->2, 2->1, 2->3... etc
Values: OSS
Uphill Ax
like Shiftmaps but for driving uphill??? How ZF recognize such situation?
Timespan ongoing Clutch
Y: nm
X: RPM
values: time
TCC Warm/cold gear
Y: throttle
X: 0,1,2,3 *)
Values: OSS
*)0: Start of Open -> Closed
2: End of Open -> Closed
3: Start of Closed -> Open
1: End of Closed -> Open

TQ Management
Y: Nm
X: RPM
Values: factor for reduction of engine power by TCU??? (like limiter send to the ECU via can?)
Upshift Offgoing Clutch
Y: Nm
X: RPM
Value: oil pressure???
Shift pressure Upshift ongoing
Y: nm
X: RPM
Value: oil pressure?

My goals
1) point of change: not arround 4400 but like 4700 rpm for example
Realization: values from Shiftmaps divide by gear ration and so I have RPM of engine... set proper values and that’s all?
What to do with "TCC Warm/cold gear" ? - leave or adjust like "shiftmaps
2) gear changing faster
- reduce timespan
- what with TQ management (increase these factors? - side effect?)
- what to do with offgoing/ongoing maps?
br,
Przemek

 

[bradsm87]

Uphill Ax
like Shiftmaps but for driving uphill??? How ZF recognize such situation?

Uphill Ax maps are used when in D mode and uphill (or towing I think) driving is detected. I believe it uses expected acceleration or holding of speed for a given amount of torque (or throttle position maybe). If the car is accelerating less than it normally does on a flat road, it must be going uphill or towing, in which case it switches to the Uphill Ax map, depending on the incline of the hill (or towing load). Uphill maps aren't used in S or M mode. If you change your shiftpoints in XE, you will generally want to change the E and the first couple of Ax maps in a similar way, so that all of the modes transition well between each other.

Timespan ongoing Clutch
Y: nm
X: RPM
values: time

This is the target time in ms to complete an upshift. It needs the right values in the shift pressure maps to roughly support that, with adaptation taking care of the finer adjustment. If you get a hard clunk at the end of a shift, the timespan is too short for the shift pressure used, causing it to go "all in" with gear holding pressure because the timespan has been exceeded.

TCC Warm/cold gear
Y: throttle
X: 0,1,2,3 *)
Values: OSS
*)0: Start of Open -> Closed
2: End of Open -> Closed
3: Start of Closed -> Open
1: End of Closed -> Open

The descriptions are not in order but just use the description for the column number. I've always been very happy with Alpina TCC behaviour and have left mine as-is.

Upshift Offgoing Clutch
Y: Nm
X: RPM
Value: oil pressure???
Shift pressure Upshift ongoing
Y: nm
X: RPM
Value: oil pressure?

I think it's target pressure at the end of the ramp-up (or beginning of ramp-down in the case of offgoing) of oil pressure. That's just my theory on it.

My goals
1) point of change: not around 4400 but like 4700 rpm for example
Realization: values from Shiftmaps divide by gear ration and so I have RPM of engine... set proper values and that’s all?
What to do with "TCC Warm/cold gear" ? - leave or adjust like "shiftmaps
2) gear changing faster
- reduce timespan
- what with TQ management (increase these factors? - side effect?)
- what to do with offgoing/ongoing maps?

It's hard to judge OSS figured to RPM because of TC slip. If you know your DDE does not limit RPM before 4700rpm (if it does, you'll need to increase it on the DDE tune) and you know it changes gears at 4400 now, try aiming for 4550-4600 first by taking your shiftmap and anything above 90% throttle, and multiply all those values by 1.04 (4400 x 1.04 = 4576)

I'd start by leaving TCC stuff alone

If you want faster shifts, slightly decrease timespan values and slightly increase ongoing pressure tables. It's a balancing act. If not enough pressure, it'll clunk at the end of the shift. If too much pressure, it will shift faster than the timespan and adaptation will pull the pressure down over time.

Leave offgoing pressure maps alone if you're putting out and reporting less torque than the max torque values in those tables.

If you are putting out and reporting more torque than the max value in any of the shift pressure tables, you'll generally want to increase the last torque axis value and increase the values in that row to follow their original curve/shape.

 

[bradsm87]

The biggest challenge we face with table discovery now is the lack of access to the full version of WinOLS, which is 981 Euro

. We have a CSV map list, but it is not sorted in to category folders. This makes it really hard because I'm searching for phrases like "druck" ("pressure" in German) and there are hundreds of tables containing that name. With folder categories in the original OLS files, only openable by a recent full version of WinOLS, we can simply skip folders not relevant to what we're looking for.

The main line pressure table(s) are very much essential to find in order to know what maximum values to use in upshift pressure and/or in-gear holding pressure tables. No luck yet.

I found the "Modulationsdruck" per gear tables. I'm not sure if this is holding pressure per gear or something else. The axis look like torque and RPM. The infield values are all the same for each gear, with the torque axis values differing between each gear (likely due to different pressure required to hold the same torque for different clutches). It explains why some factory TCU calibrations have a lower 5th gear torque limit than others because it's the only gear that's already at the highest pressure they want to use by 498Nm. Maybe extending the torque axis and a maximum pressure bump is in order for 5th gear. We'll need to know the line pressure first. There's no point increasing the clutch circuit regulated pressure if the main line pressure is already same or below the clutch circuit pressure target.

If anyone knows of a means to access the latest full version of WinOLS remotely via some sort of rental arrangement or if anyone is only using one of their two workstation licenses, please let me know. Also if anyone wants to talk about sponsoring the table discovery efforts by means of contributing one out of your two WinOLS licenses (or purchasing WinOLS for yourself and leaving one of the two workstation licenses for community table discovery), I'd say we could do an official naming rights sponsor for the community-based XDF or something along those lines. I could set up a Windows virtual machine with WinOLS installed and could give people time slots for using it. I'm in Australia and a couple of the guys also doing this are in the US so the times we want probably wouldn't clash. I can host the VM for 6 months at least, or a bit longer if more time is needed to find the tables we want.

 

[przemek4net]

Hi,
thanks for answer!
Could y pls export of mappack into csv? (that is only the question ).. or just screenshots with maplist and addresses
According to my observation: gear change should be (calculation with gear ration) around 4700 and is around 5000 so it is probably like y wrote: because of converter/slip, question: how to reduce this behavior/slip...)
In ECU edc16/17 y have translation of torque into injection quantity and In my case I still operate up to 600nm and mostly people stay with stock axes of torque..It is possible to redefine it to for example 800nm (I made it in another car/ ... maybe I will do this in the future
MAx power in my case is around 4650 and I cut fuel at 4900 so perfect could be defice gear change around +-4750)
br,
Przemek

 

[bradsm87]

Hi,
thanks for answer!
Could y pls export of mappack into csv? (that is only the question ).. or just screenshots with maplist and addresses

You can find the OLS files, bin files, KP mappack files and CSV map lists for VAG 6HP28 and BMW 8HP45 on a thread on another forum titled "Transmission remap - Let's do it ourselves". These are similar ROMs but not the same as ours, so manual searching for tables is required.

 

[bradsm87]

Torque limit per gear table has now been tested and confirmed working. I emailed the latest XHP XDF with the axis breakpoints added, the shiftmap coulmn labels added and the tq limit table added to both xHP and the guy that has the 6HP github account. The XDF should be published soon.

 

[Jeffman]

Brad, Thanks for all of your hard work. As we say here in the states, "You're the man!"

 

[doublespaces]

Torque limit per gear table has now been tested and confirmed working. I emailed the latest XHP XDF with the axis breakpoints added, the shiftmap coulmn labels added and the tq limit table added to both xHP and the guy that has the 6HP github account. The XDF should be published soon.

What values did you test per gear?

 

[bradsm87]

What values did you test per gear?

A user on another forum tested a 300nm limit where he usually runs 560nm. The car was verrry slow!

 

[bradsm87]

It seems a large amount of pressure-related tables are all very close to each other.

There are 17 "Mindestsystemdruck waehrend Schaltungen" ("Minimum system pressure during circuits") tables, most of which have values of 7000 in the Alpina B3 calibration. Neither the BMW 8HP nor the VAG 6HP map lists detail which individual table is for which purpose.

This 7000 figure is the lowest of any "minimum pressure" type tables I've seen. I'm thinking that the upshift pressure tables more than likely work up to at least 7 bar which is plenty enough for a faster shift than you probably should run.

I've linked to my OLS file for the new tables so far below. I left them in German for now. There are some tables that aren't useful in any way for tuning in there. They're just defined to rule out the addresses they use so I'm not looking over the same useless areas all the time. It's created with the WinOLS demo and will open with the demo.

Of course, none of this would be possible without Justin at @Twisted Tuning purchasing the files or _TB_ exporting various stuff out of them. Cheers guys!

https://drive.google.com/file/d/0B7gQHaVxSgteRldDMDFQSGZTLUU/view?usp=sharing

 

[AzNdevil]

don't know if you guys tested these but you can open the files from winkfp with tunerpro
but you need to run them through a hex2bin program first
these will be smaller than the full backup from xhp due to the lack of a "bootloader"

zip file attached contains a xhp backup of a b3 flahsed tcu and 3 bin files from winkfp which i have converted

 

[bradsm87]

don't know if you guys tested these but you can open the files from winkfp with tunerpro
but you need to run them through a hex2bin program first
these will be smaller than the full backup from xhp due to the lack of a "bootloader"

zip file attached contains a xhp backup of a b3 flahsed tcu and 3 bin files from winkfp which i have converted

The BMW Intel-hex to bin converter that I found online works with DME 0pa and 0da files, but not GKE215 TCU ones.

Can you link to a converter that works with GKE215 TCU 0pa/0da files?

 

[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.