NefMoto

Greetings,

I've been struggling with one of each of these cars, when tuning for roughly E50 blended fuel mixtures.  The actual adjustment of fuelling constants to compensate for the additional  value is perfectly fine.  The increase in performance is fantastic as all we have available is the equivalent of US 91 Octane.  

MED17.5 Setup

Golf 6 GTI - GT3076R
Stock Injectors
Stock HPFP (I know right, don't blame me) @ 165 bar
Walbro 450 LPH with revision pump controller and DC adjusted with an external FP gauge
Stock MAF
E45 fuel

MED9.1 Setup
Audi S3 with stock K04-064 turbo
S3 injectors
Autotech HPFP at 160 bar
Custom Rail Valve
DW65v Drop in
Custom 80 mm MAF
E45 fuel

As soon as it gets to 5500 rpm and the Median Injection Time goes > +-7.7 ms I get some savage throttle cut on the G6 and mild throttle closing on the S3.  At risk of mistaking correlation for causation, some quick calcs for injection window based on RPM and SOI/EOI:


5500 rpm = 91.6 rps = 0.01091 seconds per rev = 10.91 ms per rev = 5.45 ms per stroke, assuming that there is homogenous injection taking place and that the full intake stroke is being injected in. Now if the limiter for SOI is somewhere around 432 degrees, and EOI is roughly 60, I am reasonably assuming that this is degrees of crank before compression TDC, so 372 degrees of injection as the reasonable limit.  If one stroke takes 5.45 ms at 5500 rpm at 180 degrees, then that should leave 11.26 ms of injection time assuming injection begins 432 degrees BTDC and ends  for the latest at 60 degrees BTDC.  This is excluding any adjustment for the injector reaction time etc.

Am I far off in my assumptions?  Why am I suffering such severe throttle cut?  In both cases, the injection time is around 7.6 ms at roughly the same RPM I see the throttle start to close.  The boost and load request at this point is still high, LPFP is slightly above 5 bar as well.  All LDRLMX limiting maps have been raised, MIRL/MIOP raised and is inversely correct between the two.  Lambda control is around 5-8% additive for both.  I browsed through both the MED9.1 and MED17 FR to see what could cause it and I couldn't find the exact trigger for the throttle intervention (of course it exists, I just couldn't find it).

Both are running around 17 deg BTDC advance at the point of throttle close with minimal (<3 KR).

Is it likely that what I am experiencing is owing to exceeding the injection window, or owing to some kind of torque intervention?  I've maxed out the SOI limit at 432 degrees and still experience throttle cut. I have found that underscaling the MAF serves to make the problem worse.  The MED17.5 vehicle is at around 2300 mbar when the throttle closes, and 2600 mb for the Med9.1 S3.  The MED9.1 vehicle slowly begins closing the throttle and ends up at around 90%.  The MED17.5 G6 throttle cut is a lot worse.

Any help would be greatly appreciated.

Apologies for the crappy log pics, will upload the actual logs a bit later.

Since nobody is replying, I'm just going to continue with my solo rambling:

Concentrating on the MED17.5 Golf 5 GTI (Manual),  I no longer believe the issue to be related to DI injection window, but rather towards a limitation stemming from calculated torque:

I have tried:

Multiple KFMIRL / KFMIOP combinations
All LDRMX maps raised / scaled
MAF underscale -> Made it worse
MAF & MAP underscale -> Still experiencing throttle cut  (before my head gets bitten off, timing and fuelling tables were rescaled too, for testing purposes)
MAFless -> No difference
MAFless + MAP underscale -> Still get throttle cut

With MAFless + MAP underscale, the measured boost was under requested and it still cut the throttle.

I edited the monitoring functions. MIO / MOF / MOX, and with MIO edits it just didn't work at all, lol, I definitely edited where I shouldn't.  when raising the MOF/MOX torque limits only, it requested even less boost and throttle. 
I increased MDNORM as well(I know that this is a normalised torque value cast over CAN, but heck I was desperate).  I think decreasing these values may help??

I've followed the LDR paths in the FR and absolutely nothing is standing out as being the reason for this sudden TI.  The Injection windows have been extended and the Injection time is around 7 ms, timing around 17 degrees, and boost at 1.4 bar flat.  The boost request goes to 1.4 bar in the midrange and drops to 1.2 bar at redline, despite the LDRMX maps essentially at maximum.  Lambda control is around 5% as well, Low pressure has been measured externally at 5 bar.

This ECU is a lot more torque based from what I can see, compared to MED9.1. 

Any advice on where to look would be appreciated.

I'm sure someone with more knowledge on the med17 can confirm but sure an ECU hack is needed to overcome the limitation. I may be talking rubbish mind as never looked at one

Hello, for me the G6 GTI problem boost or knock.

first test has do a mod with advance stock ignition to see.

2nd test log pressure Turbo specified and real.

I was confronting the throttle cut between 3000 and 4000 RPM this comes from the boost

I run E76 on my 2010 passat MED17.5 with no issues at all, juts KRKATE, advance and some minor fixes in "FKKVS"... made some Jetta MED17.5.2, Fusca (New Beetle) and other with sucess... I guess its no correlated to your blend.


Only DP and intake and there is 265~270Whp in those setups..original injectors, HPFP and LPFP (just waiting to discover when they will fail, but 20.000km later and more than 10 months and no problens at all)

Just because I hate open ended threads, the problem was eventually solved, and it was fairly simple.

I just wasn't requesting enough load through mirl/miop, on previous platforms I could get away with a mild increase (210/225 +-) and that would suffice for a flat boost request, but not on MED9.1/17.5 at high timing / high boost.  Raising that to the 300/320 load range resulted in a flat boost request, and the only throttle cut I got was when it actually ended up exceeding the injection window at high boost / high rpm.

I ended up reverting like 8 revisions back and starting with the basics, and got it right that way. MYBAD.png