NefMoto

Noob Zone => Noob Questions => Topic started by: carlhook on November 30, 2022, 02:40:16 PM



Title: MED 17.5 custom PID vs KFTVLDST approach
Post by: carlhook on November 30, 2022, 02:40:16 PM
Kinda a noob.

Have my CCZB DSG6 Scirocco with stock turbo, that I just re-built, but installed a non-stock wastegate + arm.
Meaning it has a different actuation force/linearity compared to stock. (ordered wrong parts, realised with turbo off the car already, so busted out my grinder, drill press and TIG to make it work)

With stock file this leads to overboost and Throttle shut.

I have purchased a patched bench file that created a KFTVLDST map and enable/disable switch for it.

I have messed around with it a bit, and can make it almost stable-ish at WOT, but tuning all part throttle stuff looks like fun...
Most issues are weird ones, where sometimes it will undershoot the requested boost by a lot, but then the next run match it, and 3rd run overboost, but 4th run again undershoot.

I heard MED17.5 ECU's need time to adapt and that might mess with the KFTVLDST map?
I noticed the KFTVLDST map now introduces some negative Boost Pressure Control Adapt (VCDS log value)
 

So I am still returning to the idea of somehow tweaking the stock PID to accept my slightly different wastegate setup.

What is the correct way to go about this? I assume a similar setup when using a K04 turbo or any other big turbo upgrade?

Im after a very progressive pedal feel, just like stock or Stage 1/2 but with stock wastegate.


I have attached the bench bin thats patched and my Stage 1.ols


Title: Re: MED 17.5 custom PID vs KFTVLDST approach
Post by: carlhook on November 30, 2022, 02:41:29 PM
bench read with the patch.
If someone who knows their stuff could have a look and see if its done fine,that would be great.

One issue is that VCDS does not report the N75 as the duty specified in the map, looks more like an inverse of it.


Title: Re: MED 17.5 custom PID vs KFTVLDST approach
Post by: carlhook on November 30, 2022, 02:51:14 PM
the request/actual weird behaviour can be seen here.

I am slowly setting up Vehical logger, but first need to understand what im looking for.


First run
(https://i.imgur.com/T90MOfX.png)


Second
(https://i.imgur.com/2lmmG8C.png)


Third
(https://i.imgur.com/g39SSvp.png)


Title: Re: MED 17.5 custom PID vs KFTVLDST approach
Post by: IamwhoIam on December 01, 2022, 12:59:17 AM
bench read with the patch.
If someone who knows their stuff could have a look and see if its done fine,that would be great.

One issue is that VCDS does not report the N75 as the duty specified in the map, looks more like an inverse of it.

Are you sure you didn't plumb your N75 wrong? That's the first thing that comes to mind when people mention inverse N75 duty


Title: Re: MED 17.5 custom PID vs KFTVLDST approach
Post by: carlhook on December 01, 2022, 12:07:26 PM
Are you sure you didn't plumb your N75 wrong? That's the first thing that comes to mind when people mention inverse N75 duty

Yeah, you did make me think twice, just checked, and N75 is plumbed correctly, as stock.
I am also getting lean errors since I installed Unitronic intake.
Have sprayed all joints soo many times now with both brake clean (hoping to hear rpm dip/jump) and foam spray.
Last resort is smoke test.

But all that aside, how does one go about retuning PID?


Also I did some dirty fix tests before by dropping all I wg map by 10-30% on a otherwise stock file, and this managed to get the request to actual perfectly in sync.

Yet I have not played with it anymore on a Stage 1 or 2 files.


Title: Re: MED 17.5 custom PID vs KFTVLDST approach
Post by: ThomasHH on December 02, 2022, 04:55:52 AM
Why not ask the patch supplier for more information on how to set all part throttle stuff with just one map?  :D

Please attach the original bench read.


Title: Re: MED 17.5 custom PID vs KFTVLDST approach
Post by: ThomasHH on December 02, 2022, 06:40:44 AM
Attached is one translation of the MED17.5 LDREG function by Joshua Laborde.

https://damosfiles.co/product/med17-1-5-ldreg-tuning-manual/?fbclid=IwAR1gNzRu4iIPNY372ONFIi_iuKaCAQSbiiM24W-gu_LOSNpoV7GDA6Txl0E



Title: Re: MED 17.5 custom PID vs KFTVLDST approach
Post by: prj on December 02, 2022, 08:19:18 AM
Why not ask the patch supplier for more information on how to set all part throttle stuff with just one map?  :D

Please attach the original bench read.
I am not the patch supplier, but this info is here:
http://nefariousmotorsports.com/forum/index.php?topic=12352.0title= (http://nefariousmotorsports.com/forum/index.php?topic=12352.0title=)
The tuning method applies to any ECU with a RPM x target pre-control map.

But hey, reading was never your strong suit.


Title: Re: MED 17.5 custom PID vs KFTVLDST approach
Post by: carlhook on December 02, 2022, 05:21:31 PM
Why not ask the patch supplier for more information on how to set all part throttle stuff with just one map?  :D

Please attach the original bench read.

I did, not much response.
He also tried 3x to fix the N75 VCDS logging issue with no success.

Attached stock bench read and patched one.

I didn't find the LDRREG doc as I didn't know what to look for, and it did not show up while searching for PID or pre-control maps.




Title: Re: MED 17.5 custom PID vs KFTVLDST approach
Post by: carlhook on December 02, 2022, 05:24:39 PM
I am not the patch supplier, but this info is here:
http://nefariousmotorsports.com/forum/index.php?topic=12352.0title= (http://nefariousmotorsports.com/forum/index.php?topic=12352.0title=)
The tuning method applies to any ECU with a RPM x target pre-control map.

But hey, reading was never your strong suit.

Yeah, this one I saw, but thought its non MED17 related as it states - "All MED17 ECU's remedy this."

But thank you both so much for this, Ill read both and try and apply it to my setup.
I assume PID maps must be kept stock, as they sacred?




Title: Re: MED 17.5 custom PID vs KFTVLDST approach
Post by: carlhook on December 02, 2022, 05:29:09 PM
On a non PID note, I found my Lean Codes culprit.

0 leaks.

It's caused by a combination of 2 things - someone screwed with MAF housing, and cut-out all the stock fins.

STOCK
(https://i.imgur.com/hxaoHPQ.jpg)

Mine
(https://i.imgur.com/GbJOPNM.jpg)

So it worked fine with stock airbox, and even works fine with POD filter (Unitronic) straight on the MAF.
But the second I introduce the Unitronic 90 degree bend after pre-MAF, it spits garbage and Lambda Correction goes to +25%
(https://i.imgur.com/zNSJWPd.jpg)

So new MAF is ordered!


Title: Re: MED 17.5 custom PID vs KFTVLDST approach
Post by: IamwhoIam on December 03, 2022, 04:45:26 AM
Charging $75 for 4 pages worth of poorly translated OEM FR and 15 lines of sort of a "how to" is preposterous... but hey it's done by the God of all things tuning himself LOL


Title: Re: MED 17.5 custom PID vs KFTVLDST approach
Post by: carlhook on December 04, 2022, 02:39:05 PM
Ok,

MAF ordered, but to be able to do some tests now, CAD'ed up a quick airflow stabilizer and it seems to work ok with full intake now.
(https://i.imgur.com/995hmRU.jpg)

so I need a little bit of guidance as all manuals/guides are ME7 based and MED17.5 has different map names?

So KFLDRL of ME7 is KFTVLDST on my MED17.5 right? Thats the new DC WG map.

KFLDIMX is the ME7 PID I-map? So KFIWG on my MED17.5 right?

I think mines Y axis a tad screwed?
(https://i.imgur.com/wkA34Q6.png)

What and where is KFLDRAPP?
As I need to set that to different values to do the linearization runs, right?

Or do I do not do any linearization runs and just fix the KFIWG to display correct and then make it linear to utilize the patched KFTVLDST map?

If so, can someone guide me in proper axis data/addresses for my file? (KFIWG)
I thought I had a good map pack, but maybe not.


To add, weirdly, my stock file and also patched has the original KFTVLDST map.
But the patched has 2 of them?!?

(https://i.imgur.com/AhblqMp.png)




Title: Re: MED 17.5 custom PID vs KFTVLDST approach
Post by: prj on December 05, 2022, 12:52:52 AM
Ask patch supplier.
Nobody knows what was patched and what was bypassed and how.


Title: Re: MED 17.5 custom PID vs KFTVLDST approach
Post by: ThomasHH on December 05, 2022, 03:35:10 AM
Charging $75 for 4 pages worth of poorly translated OEM FR and 15 lines of sort of a "how to" is preposterous... but hey it's done by the God of all things tuning himself LOL

Yes, you and prj are the same here now. You talk a lot and are very "famous" but nothing in return. Sorry for you!


Title: Re: MED 17.5 custom PID vs KFTVLDST approach
Post by: carlhook on December 05, 2022, 05:03:45 AM
Ask patch supplier.
Nobody knows what was patched and what was bypassed and how.

Yes, I will, but I would like to understand what is the correct procedure for MED17.5 to begin with.



Title: Re: MED 17.5 custom PID vs KFTVLDST approach
Post by: quattro85 on December 05, 2022, 09:16:05 AM
I don't think someone could guide you into this, because now your boost control works in a way nobody knows(except patch provider).


CWMDAPP.4 sets B_ldsafw, which makes WGDC to run on TVLDSTAPPP.
So set CWMDAPP.4, leave TVLDSTAPPP=0 and make a run while logging;
Set TVLDSTAPPP=10, make a run while logging;
Set TVLDSTAPPP=20, make a run while logging;
...
Set TVLDSTAPPP=90, make a run while logging;

you get it - right?
Now you have data for the pre-control table.
Clear B_ldsafw

Now search FR for AWGTV function.
With a single look on the first page I believe you can find how to switch pre-control from working from model, to a table based on rpm and rl with or without PID.