NefMoto

I'm having issues with trying to match ps_w vs boost and mshfm vs msdk_w via WDUKGDN. Does one affect the other?
The car is heavily modified (1.8T forged, big turbo, RS4 MAF, 80mm Hemi throttle, I.E. manifold, 1000cc injectors, WMI) so I have recently decided to go back to basics as I wanted to get the fuelling much better in order to make the most of the WMI.
So I turned the boost down to spring pressure, turned off WMI and got the attached graphs.
It's all good up to 5000RPM, but then ps_w increases while the boost stays constant. This completely ruins my fuelling above 5000RPM (super rich).
Previously I increased WDKUGDN at 7000RPM (my next point after 5000RPM) to match msdk_w when I was running much higher boost. Is this not the right approach? Are we meant to instead edit KFWDKMSN/KFMSNWDK to get mshfm match msdk_w?

The MLHFM is a direct copy from RS4 file. I guess another option is to underscale it, but this doesn't seem like the right approach.
How exactly is ps_w calculated?
 

ps_w is calculated through inverse VE model (KFURL/KFPRG/KFPBRK(NW)).

It also tries to keep the models close, if you airflows drift too far (msdk_w vs mshfm_w) it will kick in a correction to get closer to msdk_w, this is probably what is happening.
WDKUGDN is not the only thing you have to modify, your KFWDKMSN and KFMSNWDK are probably wrong and no idea how you gonna get them correct with a hemi throttle short of flowbenching one, much easier to use a VAG throttle with a map already existing in ME7.

What is happening in your case is most likely msdk_w is much higher than mshfm_w and you get a correction factor.
You can force that factor to be 1 always, by setting the max and min limits, look in the FR for that, I don't remember off the top of my head what the parameters were called. Something like FKMSsomethingMX/MN.

I'd probably start with 80mm VAG throttle maps (e.g. rs6?), then as prj mentions - get VE in line.
Then polish up the result.

Basically:
1) Verify MAF & boost sensor are calibrated correctly, set WDK maps close enough.
2) Bring ps_w closer to pvdk_w by adjusting VE. Don't exceed sane values. If it doesn't want to go there, stop and think. Observe AFRs.
3) Bring msdk to mshfm by adjusting throttle maps.

It's essentially a hack, but if done properly - should give a good enough result.
MAF health is crucial with this approach. If it's off - you're going to screw up more, than fix.
And well - car should be absolutely perfect mechanically. No intake leaks, etc.

Awesome thanks to both.
I took your advice and copied RS6 map for KFMSNWDK from and .ols file floating on this forum.
I then used Excel + interpolation to calculate KFWDKMSN precisely.
Finally I used using the equation here to populate WDKUGDN:
http://nefariousmotorsports.com/forum/index.php?topic=9221.15
mAir[kg/h] = Vcyl[cc]*rpm[1/min]*rhoL[kg/cc]*60*nCyl (theoretical 100%rl) and multiply it by 0.95 taking PSPVDKUG into account.

This gave the attached result, which is much better than before (and fueling is more consistent).
There is no VVT, so I have set all VVT maps to 18 degrees and disabled it using CDNWS, then set all KFPBRK(NW) to 1.
I have also set FKMSDKMN and FKMSDKMX to 1 as per suggestion from prj.

Now, which is the correct map to modify to get rid of the inconsistency in ps_w at 2,500 RPM and then again at 6,000-7,000 RPM?

URL/PRG to adjust VE.
Will likely have something wonky in them around 2500ish.

Good stuff, think I am getting my head around it.
The latest attempt looks good, on the right track but some further refinement needed to bring ps_w down to correct level in boost.
I've used WDKUGDN to raise ps_w and msdk_w at the same time. I think now I need to lower KFURL to bring ps_w down (independently from msdk_w) and then fine tune wdkugdn further by raising it slightly above 5000RPM.

WDKUGDN should not affect ps_w unless you have a MAF error.

Ok, well I looked through the FR and I think this is the section that is screwing with my fueling and ps_w calculations:
"with positive load jumps, the alpha / n signal can be more dynamically accurate than the HFM signal, especially if it is areas where strong pulsations occur. In this case, it is possible to exceed a threshold value to switch to the alpha / n signal for an applicable time TDPSLGRD."
Basically, my AFR is way below target on WOT (10.8 instead of 12.5) but lambda control is consistently adding over 10%!

So it looks like it is more important to get msdk and mshfm to match first before addressing ps_w and pvdk_w?

That's why I told you set FKMSDKMN and FKMSDKMX to 1.
It makes sure that this does not happen.

But I did set them to 1!
What else would cause my fueling look like this (lambda control 1.175!):

rlroh_w, rlzuhfs_w
If they diverge - something's triggering switchover to alpha-n and you investigate it.
If they don't - can also log rkukg_w (fueling transients compensation) and check that.
If none of the above look suspicious - you adjust HFM corrections and VE, I guess. (It can get really messy here if you're unsure about what you're doing - hence my note about "sane values" in the previous post).


EDIT: right, re-read your post. It looks like it's not really about alpha-n vs hfm anymore.
Frankly, I've never seen such behavior previously. Have no idea, why the eff would frm_w be like that when you're already overfueling.

EDIT2: it almost looks like your controller is disabled under WOT (and just keeps the latest val?)

well my inline 044 was leaking, maybe it tripped some ECU monitors.
I got another higher flow pump on the way but I am away over Easter so will have to wait until April before further progress.

Reading further, the culprit might be LALIUSMN, in my file this was set to 0.8, I will try to lower it and see how it goes.

well, new inline pump and lower LALIUSMN has sorted frm_w.

WDKUGDN sorted out WOT msdk_w (few minor adjustments still needed), KISRM dialled in the dynamic msdk_w, MSLG for idle msdk_w and KFWDKMSN/KFMSNWDK for low throttle msdk_w and finally msdk follows mshfm pretty well at all times.

KFURL + KFPRG sorted out ps_w.

Still dialing in FKKVS for the WMI as there are a few areas where frm_w is working hard but everything should come together soon.

Hello BoobieTrap,

I'm having trouble calibrating bigger TB.
The TB is from 3.6 FSI N/A VW Tuareg(almos the same size as yours) The values from the software of Tuareg are not working correctly  because the ECU is MED9.1 and instead of RPM axis they have Pressure ratio values. 

As you mentioned above, you copied KFMSNWDK from RS6 file and calculated KFWDKMSN in excel. Could you please tell, did you copied only Z values from KFMSNWDK of RS6 or axis data as well? I assume you have just copied Z values and therefor you calculated KFWDKMSN in Excel using interpolation?

There is one more thing that is unclear to me. How did you calculated WDKUGDN? I have used the stock values and calculated air mass(kg/hr) for WDKUGDN but my calculation of % is not matching the stock value. Could you please tell how did you converted kg/hr to TB%?

Thank you in Advance,

 

I don't exactly remember the details as this was over 2 years ago, but as far as I can tell I copied most of the map KFMSNWDK from RS6 file, then fine tuned it based on log files.
I also remember that I had to add some really fine fine values around 0 to allow for better idle control, my first throttle plate values were 0.78, 1.18 and 1.57. Those I had to tune manually based on log files. This is because the difference in flow rate between 0.78 and 1.18 throttle opening was significant at idle RPM (I eventually managed to get stable idle at 800 RPM even with 1000cc injectors), but it would have probably been much easier to just raise the idle to 1000 RPM+.
I removed some of the higher values from the RS6 map to allow for this because my tiny 1.8 motor could never reach such high flow rates anyway.

Regarding WDKUGDN, once you have the values in kg/h, you need to use the KFMSNWDK to estimate the throttle opening at the calculated kg/h and RPM (hence KFMSNWDK is important to get right).

My advice (which is the advice given by prj at the beginning of this forum) is to use one of the VAG throttle bodies with well-known maps.
If you don't have access to a dyno, road tuning a custom throttle body is a very long and painful process unless you like the challenge and want to learn more about the inner working of the ECU. The problem is that once you get into adjusting throttle body/VE model etc (based on other mods too such as bigger MAF, different intake manifold, CAMS, injectors etc), you no longer have a fixed known variable, and as I found out, all these variables are so interlinked that you end up with a lot of guesswork.

Thank you for your reply Boobietrap,

Actually I have access to Dyno. Could you please describe the process of TB calibration on the Dyno?
Currently I'm working on a tool that will look into logs and based on the actual data generate TB maps.
I'm little confused because when I log  wdkba_w begrenzt (Minimalauswahl zwischen wdkba_w und wdkugd_w) it is fluctuating like crazy. Could you please tell what is the correct variable to log?  ???

Thank you in Advance,

I am not well equipped to give advice on Dyno tuning, since I do not have access to one.
I would guess that if you have a stock MAF (or a MAF that you are confident is linearized correctly in your tune), you can set constant throttle and RPM and for each point you can compare the msdk and mshfm and adjust KFMSNWDK so that the two match each other.

However, prj or someone else with extensive dyno experience and understanding of ME7 would be able to give much better advice than me.

This, and use gear and dyno load variations to hit every RPM/load point you can.

To calibrate the throttle you don't need a dyno, you need a flow bench!
You must measure the maximum flow through the throttle at every throttle opening. PR is not too irrelevant for a throttle as long as it's > 1, but of course if you have higher PR, it's better.

Using a car and a dyno for that is well... dumb. And it's impossible to get all the data, especially for the smaller throttle openings, as the engine will stall.
It's the best to lift the data from an ECU where the throttle was used.

Thank you for your advice prj,

I have attached a picture of maps for TB I'm using.
Could you please tell how to adapt this maps to ME7.5 as this data is from MED9.1 of 3.6 liter N/A VW TUAREG. Instead of rpm axis there is a quotient of intake pressure/pressure before TB. 

Thank you in advance,