DIY tuning

Xer0449

Corporal
Jan 30, 2017
174
59
0
I'm really interested in this in particular, because Ive dealt with the false knock as well. Had to dial the car back a bunch to keep it from happening. I'm not too keen on sourcing an M4 flywheel or the MFactory one for a few more psi, but a soft workaround may be nice :)

What kind of safeties are we talking about giving up here?

Made some more progress. Have a work-around for DCT flywheel noise but it reduces some built in safety so I'm still on the look out for an S55 flywheel. Still has a little spike post shift then an over-correction but no more throttle closure. There are some feed-forward tables that I think will help get post shift under control or it could just be that the I term is so wound up because i'm at this turbo's limit in this weather that it won't get any better with a slow reacting PWG. I looked at some stock logs I have and they have a similar spike and over-correct. Think I will focus on spool and vanos tuning now and not worry too much about post shift. It feels great in the car. Also, keep in mind this is with stock load to torque tables so the DCT is shifting phenomenal.

Logs are 2011 DCT 135i with Pure Stage 2, Vader Solutions Stage 2.5+ LPFP, VRSF Race IC and Outlet, custom inlet, Magnaflow full exhaust, custom DP, N20 plugs at .022, running E54 all DI. Ambient temp was 112F (44.44C).

2nd - 3rd - into 4th

View attachment 14233

https://datazap.me/u/sbrach/rev-29-pid?log=0&data=3-4-11-13-17&solo=13-17

3rd - 4th - into 5th
View attachment 14235

https://datazap.me/u/sbrach/rev-29-pid-0?log=0&data=3-4-11-13-17&solo=13-17
 

houtan

Lieutenant
Nov 2, 2017
611
293
0
Ride
135i N55 DCT; PS2
@Sbrach. Have you contacted mhd to see if the pid channels can be added to the mhd app? Or is it not that simple? I would be glad to shoot them an email.
 

Sbrach

Corporal
Oct 2, 2017
224
175
0
Ride
N55 DCT E82
I ordered an S55 flywheel. Even with knock detection and super knock detection thresholds maxed for the 5000-6000rpm range it still occurs.

Houtan, they aren't PIDs. We would need the ram addresses which would require disassembly to find on 98G0B which is outside of my ability. If someone were to find the ram addresses for the PID state variables it would be easy to add them to the logger. That capability already exists.
 
  • Like
Reactions: houtan and Xer0449

bradsm87

Corporal
Dec 15, 2016
162
53
0
Haven’t posted an update in a while.

Dug into the compressor characteristic / turbine power calculations a bit and put together this. Excuse the MSPaint skills ....


KW value from compressor characteristic table corrected with PID based on boost error and exhaust target flow then goes into exhaust flow, temp, and pressure models to get a WG/Turbine distribution factor. This distribution factor is turned into WGDC with the WG Position model and then the spring in the actuator is compensated for with the WGFEDER table. This “base WGDC” value is then blended with the feed forward tables. That value is summed with the altitude correction, pressure dynamics correction, and offset by gear. This value goes into the adaption function, a damping function and is then added to the output of the I controller. Note, this is a second I controller, the first adds kW to the turbine power value and uses exhaust gas mass flow. The I controller most are familiar with is the second one which adds or subtracts WGDC and uses mass airflow and boost error as axes. There are various limiters, most of which are raised in the stock M2 BIN. For the most part I have been successful leaving most of the PID set as in the M2 bin and modifying/rescaling just the compressor characteristic table with the P, I, and D tables zero’d to get boost close to target. Then turning PID back on. I found one of what I thought was my best revisions made about 8psi less with PID turned off. Obviously the compressor table wasn’t right and PID was just adding a ton. Would really be nice to be able to log the individual PID terms like N54.

A couple more notes. WGDC Base log channel appears to be the output of the compressor characteristic table in kW. I don’t believe it is WGDC.

I’m not sure what WGDC after PID is.

In my bin I’ve rescaled the WG position model to a linear 1-100% and zero’d the WGFEDER table (how an EWG car is). I know this is incorrect as the PWG model is not linear because the WGDC does not result in a linear change in WG position but I personally have found it easier to tune this way.

There’s a lot more I’m leaving out I’m sure but just wanted to spark up the PID tuning discussion again.

Spool is great, boost follows target no problem, but it still spikes a few PSI post shift on DCT and AT cars. The WG tables for post shift are all only active on MT. @LamboLover you mentioned your custom boost control strategy on GTR resolves a similar issue. Any insight?

Excellent work! Great method getting boost control sorted by zeroing PID, getting boost consistent at a little below target then re-enabling PID too.
 

Jake@MHD

Major
Platinum Vendor
Nov 7, 2016
1,593
2,058
0
Philly
I ordered an S55 flywheel. Even with knock detection and super knock detection thresholds maxed for the 5000-6000rpm range it still occurs.

Houtan, they aren't PIDs. We would need the ram addresses which would require disassembly to find on 98G0B which is outside of my ability. If someone were to find the ram addresses for the PID state variables it would be easy to add them to the logger. That capability already exists.

Which vars?
 

Sbrach

Corporal
Oct 2, 2017
224
175
0
Ride
N55 DCT E82
Atlri_xeb - I Share (WGDC) - %
Mssol_akzu - Exhaust Target Mass Flow - kg/h
Mssolatl_k - Exhaust Target Mass Flow Corrected for Boost control - kg/h
P_atlr_d_xeb - D Share (Turbine Power) - kW
P_atlr_i_xeb - I Share (Turbine Power) - kW
P_atlr_p_xeb - P Share (Turbine Power) - kW
Pturbine_vst_xeb - Turbine Power from Comp Map - kW (Think this is "WGDC Base" in Logger)
Pturbine_xeb - Turbine Power after PID - kW
Pld_diff_xeb - Boost Error - hPa
Pld_grund - Basic Boost Pressure - hPa
Pld_diff_grd_xeb - Boost Error Dynamic - hPa
F_pldsvld_f_xeb - Charge Factor - (Think this is "Boost Stepoint" in Logger)
 

Sbrach

Corporal
Oct 2, 2017
224
175
0
Ride
N55 DCT E82
Finally got an S55 flywheel.
IMG_4809.JPG


about 17-18lbs lighter.

Look how bad the original one was...


Time to max this PS2.
 
  • Like
Reactions: Jake@MHD and houtan

Bnks334

Lieutenant
Dec 1, 2016
524
340
0
New York
Finally got an S55 flywheel.
View attachment 18122

about 17-18lbs lighter.

Look how bad the original one was...


Time to max this PS2.

So failed DMFW causes rpm discrepancies picked up by misfire logic... DMFW is a really common failure on VW and Audi. Question is... S55 is designed for 400whp/400wtq so is it an actual LONG TERM solution over just replacing stock flywheel? I guess the weight savings is still an advantage even if it fails just as quickly.
 
Last edited:

Sbrach

Corporal
Oct 2, 2017
224
175
0
Ride
N55 DCT E82
So failed DMFW causes rpm discrepancies picked up by misfire logic... DMFW is a really common failure on VW and Audi. Question is... S55 is designed for 400whp/400wtq so is it an actual LONG TERM solution over just replacing stock flywheel? I guess the weight savings is still an advantage even if it fails just as quickly.

No I don't think it is misfire detection. A. the crank trigger is on the crank, not the flywheel. B. The most common result is superknock detection on 5 and 6. I think the two masses of the flywheel just bang together and the noise threshold during the superknock window is very sensitive, especially at high RPM.

As for the S55 flywheel durability; I'm sure not having the additonal 17lbs weight riveted to it will help and Jordan's 1000whp S55 is doing fine with the stock flywheel.

didnt mfactory make a flywheel ?

Yes but single mass. It's sold out until after the holidays and I haven't seen any reviews on it. My concern was that being untested it could have a different harmonic that caused a different issue or the chatter at idle could be really bad. Not saying it is just that it's unknown whereas the S55 flywheel I knew would work...
 

Bnks334

Lieutenant
Dec 1, 2016
524
340
0
New York
No I don't think it is misfire detection. A. the crank trigger is on the crank, not the flywheel.

Right... and it's measuring crank rotation speed for deviations which would generally be indicative of a misfire (dead cylinder), no? As in, mechanical jamming/imbalance of the failed dual mass components...

I was under the impression people complained more about misfires than knock. Both are a possibility... I am not saying otherwise...

and Jordan's 1000whp S55 is doing fine with the stock flywheel.

It's a brand new car, I would hope the DMFW doesn't fail THAT quick even at higher than rated whp. Plenty of people get long life out of the stock DMFW... The S55 flywheel is surely a bit stiffer than a stock flywheel but I questioned by how much... Also, you think he drives that car around at 1000whp?

Yes but single mass. It's sold out until after the holidays and I haven't seen any reviews on it. My concern was that being untested it could have a different harmonic that caused a different issue or the chatter at idle could be really bad. Not saying it is just that it's unknown whereas the S55 flywheel I knew would work...

Engine harmonics are overlooked too much. Since you mention it, are you going to pair the M DCT flywheel with the M harmonic damper? I am sure the two are designed to be paired together... I run a SMFW and it's really not what you want if you care about getting longevity and reliability out of your transmission or rotating assembly lol.
 

Sbrach

Corporal
Oct 2, 2017
224
175
0
Ride
N55 DCT E82
Most people don't know the difference between cylinder deactivation and misfires. These cars basically never misfire unless you have a bad coil/plug.
 

Sbrach

Corporal
Oct 2, 2017
224
175
0
Ride
N55 DCT E82
Well Christmas came early in AZ...
IMG_4991.JPG


IMG_5024.JPG


Got VTTs new kit to test and a PI kit from PFS to support it. Installed the PI this weekend and man what a difference. Rail pressure is a solid 2800-3000psi the whole pull and my HPFP VCV duty is down in the 50-60% range. When testing the VCV changes I've implemented I found that about 65% was the max it would command with stock tables. I increased that and was running 80-90% DC which resulted in much higher flow allowing us to run E50 at 22psi without supplemental fueling with the PS2 but this was right on the edge for my car. I had to be careful to ramp boost up otherwise I would crash rail pressure at low RPM and I'm sure I was putting extra strain on the HPFP, not a big deal since they are relatively cheap now but still. Many people claim PI on N55 is difficult and will misfire. I haven't had any issues but I added tables for max fuel mass, max fuel flow, and max injection time plausibility a while ago as well as increased scalar limits. So far so good. Base map on the AIC and my first wack at the scalar table for PI got it pretty close. Some small changes should get STFT down a couple more percent and then I'll call it good. I haven't tuned for PI before so any tips are appreciated.

Capture.JPG


https://datazap.me/u/sbrach/4th-gea...19-22-23-26-27-28-29-30-31&solo=9-14-19-22-23

The plan is to dial in this tune with PI a bit more, PID is pulling a bit to much out up top when it goes slightly over target, and get it on the dyno to get a baseline with the PS2. Then the new turbo goes on. This thing is pretty. Can't wait to see what it can do.

IMG_5035.JPG


IMG_5034.JPG


IMG_5036.JPG
IMG_5039.JPG


Merry Christmas

IMG_4935.JPG
 

Attachments

  • IMG_4994.JPG
    IMG_4994.JPG
    160 KB · Views: 89
  • IMG_E4996.JPG
    IMG_E4996.JPG
    227.4 KB · Views: 97

houtan

Lieutenant
Nov 2, 2017
611
293
0
Ride
135i N55 DCT; PS2
Very nice early xmas. Looking forward to seeing your ps2 numbers with flywheel vs mine on stock flywheel. And of course results on the new setup.

You are doing e85 with the port setup?
 

Sbrach

Corporal
Oct 2, 2017
224
175
0
Ride
N55 DCT E82
We only have E51-E54 here. It's nice because it is very consistent. I can try some T85 but I don't think there will be much gains over E50 at this power level.
 
  • Like
Reactions: houtan

houtan

Lieutenant
Nov 2, 2017
611
293
0
Ride
135i N55 DCT; PS2
Link to PI kit? I couldn’t find anything made by a company called pfs.
 

Sbrach

Corporal
Oct 2, 2017
224
175
0
Ride
N55 DCT E82
Took the car with the current setup (Pure Stage 2, E54, PI, FBO) to get dyno'd on Thursday. 475whp/476wtq. AZ Dyno Chip. This was at 22-23psi, WGDC in the 90s, 11.5* of timing. Elevation is 1300ft above sea level. Little less peak power than I was expecting with the PS2 but the curve is pretty good. All of the real world dyno charts that I have seen are right around 480 so I think the stock manifold is just too much of a restriction above about 450 wheel. It's at the point where turning it up more doesn't make more power.

Pure Dyno 475_476.jpg


Started taking the car apart and got the PS2 removed.
IMG_5119.JPG


Room for activities
IMG_5120.JPG


Told my wife I was going to install both for twin turbo and she said "Can you do that?"

IMG_5115.png


VTT turbo all setup and ready to go.

IMG_5106.png


Side by side cast v-band manifold vs OEM turbofold. Look at the difference in size of the vband collector vs stock. Even with similar sized wheels I think this is going to make a drastic difference. Keep in mind the stock manifold is double walled with an air gap for insulation. The stock ports through the collector and in the turbine housing are tiny. Speedtech posted a good photo of a cutaway for reference.

IMG_5116.png


stock turbofold.png
 

Attachments

  • IMG_5115.jpg
    IMG_5115.jpg
    276 KB · Views: 101