DIY tuning

Sbrach

Corporal
Oct 2, 2017
224
175
0
Ride
N55 DCT E82
BMW/Bosch didn't design the stock boost control strategy for stock turbo. They designed it for *any* turbo. The point of the system is that they can change the turbo and just drop in the new characteristic tables and everything should work. The logic is the same for 335/M2/M3 etc. If we can get a handle on the models and create reasonably accurate characteristic tables for a PS2 or a BW 7670 or whatever then tuning would just be dropping in those tables and PID takes care of the rest. A lot more work up front but much less work down the road.
 

all4bspinnin

Corporal
Jun 12, 2017
178
98
0
Ride
135i TT
If thats the route we want to go then we'll do that then. I have no issue building a comp characteristic table. Its not difficult. But like you said, more work up front.
 
Last edited:

LamboLover

Corporal
Apr 6, 2017
238
242
0
Ride
Everything
Tim seems to be resigned to it or embracing it for Ford (ie dropping in the new compressor and/or turbine tables) I couldn't tell which.
 

cookiesowns

Private
Nov 13, 2017
26
7
0
Ride
2010 535i ZHP6 + 2017 M2 6MT
No this was before the new tables but the new tables won’t help if you are starving the HPFP. Around 17-18 psi was the limit on E30 if I remember correctly. You would think a 450 would be plenty. Might help to find additional tables for fuel mass request sent to EKP.

On the F series at least, the EKP module has various coding values for various levels of a "DME" tune. I don't believe the request is directly from the DME but rather coding on DME and also coding on the EKP module. Overcurrent values can definitely be manipulated via manual hand coding of the caf/ncd.

BMW/Bosch didn't design the stock boost control strategy for stock turbo. They designed it for *any* turbo. The point of the system is that they can change the turbo and just drop in the new characteristic tables and everything should work. The logic is the same for 335/M2/M3 etc. If we can get a handle on the models and create reasonably accurate characteristic tables for a PS2 or a BW 7670 or whatever then tuning would just be dropping in those tables and PID takes care of the rest. A lot more work up front but much less work down the road.

Correct. I think most tuners don't spend the time experimenting and understand the boost control tables on these cars... There's alot more that I still too don't understand. But on stock turbo or even PS2 it's not too difficult to tune if you take the time to understand the tables and make small changes to experiment. I understand most tuners don't have this luxury.. but honestly while commanded WG like on N54 might make it simple on the N55.. given the nature of the small PWG turbos i don't think it's going to be very good.

I see quite a significant difference between WGDC needed to achieve target boost between gear pulls depending on the car. Hell even F10 N55 vs F30 N55 has slightly different WG actuator characteristics that's defined in the WGPOS model table.

That said... I've been seeing alot of tunes out there that rely on the ceiling of the WGDC output/compressor table to adjust boost, and leave the deviation between target and boost actual above 3-4psi..... they also seem to raise the PID adder completely to 99.99% and rely on PID ceiling along with other limit tables to achieve the desired boost of the tuner.

The boost control on these cars are very, very, very precise and can account for a significant amount of variances.... if you fight it it fights back hard.. but if you work well with it, it rewards you quite well from the limited experience I have.

Tim seems to be resigned to it or embracing it for Ford (ie dropping in the new compressor and/or turbine tables) I couldn't tell which.

Question... have you seen differences between early E series N55 ROM's and late F series PWG roms? There seems to be wildly difference variances in tune performance from these. In skimming through the public E series n55 rom's it seems the table definitions are very slightly different too.
 
Last edited:

Jake@MHD

Major
Platinum Vendor
Nov 7, 2016
1,593
2,060
0
Philly
On the F series at least, the EKP module has various coding values for various levels of a "DME" tune. I don't believe the request is directly from the DME but rather coding on DME and also coding on the EKP module. Overcurrent values can definitely be manipulated via manual hand coding of the caf/ncd.

DME definitely sends a L/hr request. It's very probably maxed at the high end though, as is on N54.
 
  • Like
Reactions: cookiesowns

LamboLover

Corporal
Apr 6, 2017
238
242
0
Ride
Everything
Question... have you seen differences between early E series N55 ROM's and late F series PWG roms? There seems to be wildly difference variances in tune performance from these. In skimming through the public E series n55 rom's it seems the table definitions are very slightly different too.

I've not looked enough, I've been deep into comms, memory management, automation, circuit design/firmware dev etc. The stuff I do these days is too abstract from the tuning, so for example if someone doesn't like the flex fuel tables I've blended, I just change or add a few rows in a spreadsheet, it grabs the named maps from the xdf, writes the code to whizz them up and recompiles, I was just going to leave the tuners to get on with it and give them features.

I'd like to understand the electrical output characteristics of PWG though. I have not yet found any info on how it works, is it PWM, what frequency, does it have end stop switches from which it learns, does it measure drive current to work out physical resistance etc.
 

Jake@MHD

Major
Platinum Vendor
Nov 7, 2016
1,593
2,060
0
Philly
I'd like to understand the electrical output characteristics of PWG though. I have not yet found any info on how it works, is it PWM, what frequency, does it have end stop switches from which it learns, does it measure drive current to work out physical resistance etc.

You mean EWG? The pneumatic wastegate cars use the same solenoid as N54 (300hz pwm, vac operated).
 

Sbrach

Corporal
Oct 2, 2017
224
175
0
Ride
N55 DCT E82
DME definitely sends a L/hr request. It's very probably maxed at the high end though, as is on N54.

Yeah it requests L/hr and max is 254. 255 is error. Coding the EKP with NCS for flow characteristic curve 51 instead of 50 helps run more ethanol. I doubt it changes anything in max mode but maybe biases the middle of the curve for a bit more flow. It makes a big enough difference that on E50 we were starving the HPFP with a Fuel it S2 but it was fine on the same tune and a stock pump with just the EKP coding change.

I'd like to understand the electrical output characteristics of PWG though. I have not yet found any info on how it works, is it PWM, what frequency, does it have end stop switches from which it learns, does it measure drive current to work out physical resistance etc.

I haven't looked at it much but the EWG has a min and max position mechanical stop and adaption for those positions as well as adaption for transit time, current and duty cycle. It appears to calculate a required current to move the gate and then derives a duty cycle from that like the VCV. It also has a temperature model and limits.
 

cookiesowns

Private
Nov 13, 2017
26
7
0
Ride
2010 535i ZHP6 + 2017 M2 6MT
DME definitely sends a L/hr request. It's very probably maxed at the high end though, as is on N54.

Perfect... I wonder why some of the tune platforms out there haven't exposed these tables yet. Either way, the EKP probably won't honor these requests if the module isn't coded for it, at least on F series yeah?

I haven't looked at E series cars via NCS..

Yeah it requests L/hr and max is 254. 255 is error. Coding the EKP with NCS for flow characteristic curve 51 instead of 50 helps run more ethanol. I doubt it changes anything in max mode but maybe biases the middle of the curve for a bit more flow. It makes a big enough difference that on E50 we were starving the HPFP with a Fuel it S2 but it was fine on the same tune and a stock pump with just the EKP coding change.


I haven't looked at it much but the EWG has a min and max position mechanical stop and adaption for those positions as well as adaption for transit time, current and duty cycle. It appears to calculate a required current to move the gate and then derives a duty cycle from that like the VCV. It also has a temperature model and limits.

Do you mind sharing some of the coding values from NCS for the E series EKP? I can post the decrypted NCD/CAFs of my F series N55 EKP too. Curious to see the differences.

F30 N55, and M2/X4m40i N55 all have different coding values for the EKP. Different "power levels" are defined by coding changes that I can tell on both DME and EKP.

However how it actually related to the ROM/calibration on the DME alludes me... which I'll leave to you guys here.

But either way... it sounds like we have very smart people working on DIY tuning on N55 MEVD.. which is awesome.. and probably some of you far surpass my knowledge about the actual DME innerworkings.

You mean EWG? The pneumatic wastegate cars use the same solenoid as N54 (300hz pwm, vac operated).

heh.. I had to look this up myself. Someone reputable somewhere mentioned that N54 and N55 pressure converters aren't interchangeable... and I believed him. Silly me.

https://www.realoem.com/bmw/enUS/pa...335i&mg=11&sg=45&diagId=11_4472&q=11747626351

Indeed they are the same.

Currently we have the WG Position table. Do the X columns account for PWM% output to the solenoid? On N63/S63 PWG rom's it seems like the columns scale to 100%, vs capping at about 45% on N55 PWG.
 

Sbrach

Corporal
Oct 2, 2017
224
175
0
Ride
N55 DCT E82
F990BC7B-B7AE-46D0-94E4-2D92B1A5DA1D.png
Update:

Added some more torque tables and set all the tq limiters I could find to 1000nm. Now I’m running my DCT car with stock Load to Torque tables with no issue even with report tq at clutch over 700nm and shifts seem better/faster.

Only issue I see is that from 6500-7000 rpm it drops requested load a bit. My load target is high enough that it isn’t an issue but I will try to figure out what table I’m missing. Load max doesn’t drop. Only target.

https://datazap.me/u/sbrach/rev-41-2?log=1&data=3-7-12-14-15-16-17-18-19-21-22&solo=3-12
 

houtan

Lieutenant
Nov 2, 2017
613
296
0
Ride
135i N55 DCT; PS2
Mid rpm 3rd gear stab stock flywheel and no timing corrections?! That is badass brother. Is this on commanded wgdc?
 

Sbrach

Corporal
Oct 2, 2017
224
175
0
Ride
N55 DCT E82
Mid rpm 3rd gear stab stock flywheel and no timing corrections?! That is badass brother. Is this on commanded wgdc?


Yes still on commanded. Did some testing today with PID but I have more reading to do. It’s closer than my last attempt but still doing things I don’t understand.
 
  • Like
Reactions: houtan

houtan

Lieutenant
Nov 2, 2017
613
296
0
Ride
135i N55 DCT; PS2
super jealous you are running E50, no meth. Hoping I can do the same when everything is ready for release.

And per that last log, definitely hoping to try the PID approach.
 

bradsm87

Corporal
Dec 15, 2016
162
53
0
I don't mean to be a leach but my situation is very different these days compared to when I had the N54. I have zero time, motivation or petrol money to be doing heaps of table testing like I used to. Hopefully my N54 sharing and contributions can sort of justify my information leeching this time around.

I still haven't so much as put the downpipe on my N55 PWG F11 or even flashed it once. I'm kind of just waiting for more info to come about and 8HP TCU flashing so I can just raise the 3rd and 6th torque limits in the TCU so I don't have to under-report torque on the DME.

Can anyone specify which torque limit tables are verified as needing to be raised when increasing load target? I'm really not a fan of modifying tables that turn out not to need modifying. I just really want to find out which ones need raising to not get torque limits.

Can anyone share some more specific insight into which tables are best to modify to get boost under control after switching to a catless downpipe? I believe there is definitely a more appropriate table than the power to drive compressor table that is most commonly used. It seems a few people are modifying others on BM3. What are the others and how to they differ in practice to each other? I think it's important that we find the right way to get boost control under control using tables that apply to the turbine side, not power to drive compressor which should not need to change.

I am assisting with table discovery when possible but tuning/testing I just can't do. All I can really do at this stage is get all the info and hopefully get the tune sorted within 10 or so revisions. I did over 120 revisions on the N54 testing various tables and features and sharing the info but I'm unable to do the same again.

To be honest I'm a little frustrated in general with the N55 tuning scene. It's all Chinese whispers.
 

Sbrach

Corporal
Oct 2, 2017
224
175
0
Ride
N55 DCT E82
So not sure if your frustration is directed at me but I think I have been sharing pretty openly. I can't just post every revision of my xdf and bin because a) there are people like EMP out there and b) this is dev work and much is untested.

For the torque stuff I am now running 100% stock load to torque tables. The tq limiter tables - KL_MDMAX, KL_MDKMX_ATL_***, KF_MDKMAX_GANG_MOD1, and KF_MDKMAX_GANG_MOD2 are raised. Additionally, some codewords for various tq intervention functions are disabled.

It is hard to test individual parameter changes on the street with a flash required for each iteration. So while I agree that it is best to only change what absoluteluy is required it isn't necessarily always practical. At least until real-time tuning is released.

For the DP you should look at KF_AUSY_TURB - Flow characteristic turbine.
 

Sbrach

Corporal
Oct 2, 2017
224
175
0
Ride
N55 DCT E82
More work on PID boost control.

I'm starting to wrap my head around it and it seems that maybe some of the confusion is that most (me included) have attempted to tune it like an N54. We assume P and D apply to the WGDC but that doesn't appear to be the case. As far as I can tell; Compressor Capacity -> Power-based controller PD share -> Commanded WG Switch -> Conversion to WGDC -> Duty Cycle Correction -> I Share -> Adaption -> Duty Cycle Limit

All the P and D tables use Target Exhaust Flow Pre-Turbo rather than Target Boost and the Z data is in kW. So running stock P and D tables, M2 Compressor Characteristic Map with some tweaks, and raised limits on the I-component I get this.

https://datazap.me/u/soyelmaui/multi-pulls-64?log=1&data=4-5-13-22-32-33-39-40&solo=4-5

I think my next step is to lower target load to get around a 17psi boost target and see how well it controls. I can't hit a 21psi target at redline when its 110F (43C) out.