NefMoto

RS4 engine with K04/K24 buehnbombs (Troys yber'brids for those who know the story), E85, 1000cc injectors, Tonys base M-box file now with super high MIRL (no help) LDRXN stayed at 240. Boost was 1.65peak taper to 1.3. Interpolated MIOP and all the usefull stuff I've picked on the forum ARMD,NWS-smoothing...

I'm running AVCR parallel but it's maxed for now so should be pretty much blocked.

I'm out of ideas of what to try next. Is it knock voltage related? Copied KFKE's from K-box, no help (next revision has 10% on KFKEs)

Oscillation over 5000rpms usually much worse than on this pic, up to 7degree jumps. IAT under 90F, EGT in control although B1 is 1110C so It's running on BTS map until I get my spare but I'm Ok with that.

timing=basar_0
mifa=misol

Surely I forgot some important bits but ask and I'll try to answer.

Here's a pic of cyl 6knock, act load, req load, timing=basar0, retard cyl6
Other pic of related maps


(https://lh5.googleusercontent.com/-D-Z-ZzkZxG0/UDfEorqTXbI/AAAAAAAAB4k/WXVE3zhyTZA/s912/load%2520act%2520req%25205000rpm%2520problem.GIF)

(https://lh4.googleusercontent.com/-Qy3RqOVwXUU/UDfCBzIlGVI/AAAAAAAAB4c/YctHyV4usQA/s912/mbox%2520timing%2520and%2520knock_20120824_192724.png)

A crazy MIRL will really screw up the IOP situation as well. You might want to revisit your IOP.

If your req load goes under actual, you're gonna have intervention regardless of knock or anything else.

You may end up underscaling your MAF to keep act. load reasonable :/

Or you can play selective scaling games like prj did here
http://nefariousmotorsports.com/forum/index.php?topic=2444.msg23124#msg23124

In any case, you will have to post more logs showing all the various torque intervention variables if you want to get to the real cause.

Actually, i take it back.

Your actual load is NOT over requested.

Requested is "load corrected specified" which you did not post.

If "load corrected specified" is the same as ENGINE LOAD CORRECTED then they are the same on that log. LOG-file attached.

I used the Berttos interpolator for the IOP and when the MIRL was lower at 238 I also tried to mod the IOP  based on logs (where loads cross and up) but it made no difference?

I've always had pretty (MRC C-box for gasoline) high MIRL so I can dial the boost with my AVC-R and if possible I'd like to keep that. So it should work with cracy MIRL too if I just figured whats pulling the timing.

I'm back to RS4 airbox so I should get accuratish maf values (rs4 maf and maps) but is it possible it's already restricting  the flow at 400g/s?

(https://lh5.googleusercontent.com/-LJe_yF2TR4A/UDhc6qHyWpI/AAAAAAAAB44/Jx5EG2RN7oc/s912/aaa_mbox%2520timing%2520and%2520knock_20120824_192724.png)

I checked an older log when MIRL/IOP were Tonys base maps and back then act was 190 and crossed the requested late at 6200rpms. I still boosted about the same 1.65 and FATS was the same as it's now. So, maybe I'll just have to lower my MIRL to make the Load monitor happy and just boost away with my AVCR?

Is there a way to raise the "load corrected specified"? I must have some corrections going on the background that I can't spot.

IRL/IOP are torque monitoring, not load monitoring. You didn't actually graph any of the important torque variables to see if torque monitoring is active.

However, your actual load is still above your requested load, which generally indicates that your MAF still isn't scaled right.

(http://i.imgur.com/JcSpT.png)

miist is over misol, but I'm not seeing any real torque intervention.

It's a RS4 engine with RS4 intake component and I used rs4 maps for maf so it should be  pretty close? I could  try my X-1 intake so see  how it reacts?

Post plsol vs actual vs wgdc.

Edit: nvm will get your log, but please zip it next time!

Throw away the AVCR, its useless. You already have an amazingly flexibly PID at your fingertips.

At least sell it.

Disconnect anything controlling boost other than your ECU and run log..

AVC-R disconnected. No change in load or boost as I presumed.

(https://lh6.googleusercontent.com/-L2H38rUFWlc/UDuhHwTTQNI/AAAAAAAAB5M/YycCuUpv3Vs/s912/AA-mbox%2520timing%2520and%2520knock_20120827_190913.png)

Looks fine to me. I dunno what your req boost looks like, but I see nothing obviously wrong in that.

req boost is pretty much actual over 5000rpm.

My LDRNX is 240 and the boost tapers, so do you think I'm at the limit of VE now or could I get more boost up top if I still raised the LDRXN. IIRC I've had over 200 act load at 7000 before I switched to M-box.

edit: found one example
(http://www.saunalahti.fi/~tieksola/fats/66F%2026half%20psi_tiedostot/4runs_compare_graph_v11.15c%20pohja%20XXXXXXXXXXXXXXXXXXXXXXXXXXX_5809_image031.gif)

I would guess MAF scaling.

What's up with your WGDC?

WGDC? Do you mean why it's only at 70%? WG's are not stock K04 ones. Maybe that explains it?

I just tried LDRNX 240-->250 which did not change things over 5000 much.

Why are you trying to raise target load?

You need to get actual load down to where it should be.

Do you have wideband?

Well I'm trying to go faster : )

This car has it in her and I think I still can boost 1.4 bar/20PSI untuil 7000rpms just like before. If I can't then I will try to lover the target to match spec corrected and see if timing gets solid. Keep the ideas coming I'm not done yet trying to get a bit more up top.

Whatever you are trying to "fix", it isn't going to give you more power.

I'd like to know why? A bit more boost or air volume up top potentially has room for more power IMO as I'm not fuel limited yet and the timing is not too much corrected yet.

1. You will be hitting ps_w limit by requesting this much load.
2. The reason for your low requested load is an incorrect KFMIOP.

Thank you. I will try to mod the IOP just for giggles.

Don't mod it "for giggles", do it correctly.

Or even better, post your KFMIRL incl. axes.

This was after last nights tests that did not do anything for actual boost. For next version I raised KFLDIMX a bit to see if that holds some more boost after the spike.

(https://lh6.googleusercontent.com/-_T2wg-lv8cQ/UDyYdOeBwTI/AAAAAAAAB5g/Xq1huFG0Xzw/s912/A-MIRL-IOP.GIF)

The pic can be zoomed hopefully here
https://picasaweb.google.com/lh/photo/6mHIQrsdxvK4JrvresgTj3QrFDJpp-YMyZkhd_Dw__M?feat=directlink

I don't know what has been going on in the meantime, but from the log you initially posted, your mifa_w was going to 70%.

Your KFLDHBN is tuned like crap. Do it correctly as well. Or temporarily set it to 3.00 across the board, because it will be limiting your load.

You also can't just go and request 270 load, that won't work.
Log ps_w. And overall, log less useless things, and increase the samplerate. Your template is a right mess, remove the things that are not needed.

Ok, I will try 3.00 for the KFLDHBN (unmodified so far)

270 was a test so I will back out but 248 only means 1.5bar?

Will shorten the .cfg too and try higher sample. I was getting errors before so I lowered the sample but hopefully cleaning some variables will help.

You will never reach more than 220-230 load, because you will max out ps_w.
More than that, actual load will just cap out and you will go lean.

Good idea to fit wideband as well.

You need to underscale MAF and do some other things. Good luck doing that without wideband.

I have LM-2.  I just tried the PID mod and it kinda worked. Got a dip after spike but I got over 1.45 bar at 6700rpms : )

I went back to 248 MIRL at the same time though and also raised the KFLDHBN to 3.00 so not sure whitch one made the wgdc go up?

ME7Info will provide all the necessary variables for LDRLMX. Stop guessing and log.

How does your target vs actual AFR look up top?

Edit: changing IRL will do nothing if it is being limited by rlmax_w

What made it go up was probably KFLDHBN.

This.

After about 10 versions trying to smooth the PID and deviations I finally tried the RS4 maps nehalem kindly provided for all of us.

It did not solve the deviation problem when spool overshoots and req is lagging but no more dip after the spike. It also gave too much boost as I  still had 248 ldrnx. 1.93bar to 1.5bar was not a lot faster than 1.65 to 1.45 so I will aim for the lower boost and fine tune AFR and timing which pegs the following questions:

1. Should I try to match my WB AFR to requested AFR with FKKVS or KFKHFM or does it really matter? I don't have rpm at the moment on my LM-2 but I will install it next weekend. It's hard to figure AFR from logs with no rpm and I do see some lean spikes that I need to take care of.

2. I'm asking LDRXN 230 at spool but req boost keeps lagging up until 5250rpms and KFLDHBN is not limiting the boost so I get deviation TC. Are  there any maps that I could massage to get req boost up on spool or is it the RS4 PID thats keeping it lower and about linear? For the mean time I widened the deviation delta.

Set the req boost where you want it, then tune the PID.


You should be able to do it all with KFKHFM and KRKTE.

You can't really fix "spikes" per se, only basic trends. If you are seeing spikes, there are likely other issues.


I think you are confusing req and actual. KFLDHBN limits req boost. Devation is the difference between req and actual. KFLDHBN cannot limit actual, only request.


The PID has no effect on request. The PID is there to make actual match request. You need to research PIDs and control systems. Request is the set point, actual is the measurement.


Never do a "mean time" fix. Fix the underlying problem.

KFLDHBN limits via throttle cut if actual stays over the PR setpoint for a specified time. I have mine set to just under the MAP limit and if I spike and hold ~1.5bar for like 1sec I get throttle cut via KFLDHBN.

KFLDHBN does not throttle cut anything. It affects specified and thus requested load.

The throttle cut is due to actual staying over requested for an extended period of time.

Ok, it must have been a coincidense it happened for me after setting KFLDHBN.

Thanks for clarifying :)

No, it may not have been a coincidence, you probably just exceeded requested load.
Please, when you tune understand why you do things and how they are done. Don't just try changing maps for kicks and seeing what happens.
This is best way of wasting lots of time and getting nowhere... especially if the documentation for the ECU is readily available.

Now you lost me.

First you agree with me saying KFLDHBN will end up with throttle cut here:


Then you tell me I dont know how this map is used?


My understanding is that this map will limit the system to run over a specified PR, simple as that. Since this whole ECU is built to output a specific torque in an closed loop, regardless of ambient conditions.

And since boost is controlled in this manner, you could possibly place your car on a dyno ontop of mount everest an try to achieve this requested torque and boost your k03 into pieces, IF it wasnt for this map.

The reason I tuned this map was to hard limit my boost to ~1.45bar while dialing my boost PID on the street, since my block is stock and cant handle any kind of boost spikes. This appeared to me as a very good reason to use this map, but if you have a better way of doing this, please give me some tips.

Bische

KFLDHBN  makes up part of the requested load calculation.

When actual load goes over requested then the ECU can go as far to cut the throttle to bring actual load back in line with requested.

So....

If your actual boost goes over KFLDHBN then you ~can get a throttle cut, but this doesnt mean to say that KFLDHBN controls if you get a throttle cut or not.

If your PID is tuned correctly, you shouldn't get throttle cut, regardless of requested boost. Period.

Based on your statements, I figured that this was the case, yes.
This map only limits specified corrected via rlmax. Furthermore, this map is there to protect the turbos from overspinning. It is not there to protect your "block" as you put it.
It also has as little or as much to do with throttle cut as any other map that affects the requested load calculation.

The fact that you are using it as you are shows a lack of understanding for this map. If you want to limit maximum cylinder pressure for your block, you should use LDRXN together with KFTARX. Boost is not what puts pressure on your block, cylinder filling and combustion circumstances do.
KFLDHBN is for the turbos alone. This is because cylinder pressure will rise and 1.5 bar in -10C ambient and 30C ambient will put very different strain on the mechanical components in the engine, but it will not do all that much for the turbos. In case of the turbos - the main problem is altitude and overspinning them there.

So in closing - yes, I do not think you are using it correctly.
It's your car, your tune, and you can do what you want with it, but saying that KFLDHBN is some sort of hard boost limit that gives throttle cut is incorrect. In fact, KFLDHBN should never cause throttle cut, it will just make the ECU request less load in some circumstances. If it is causing a cut, then your PID is configured incorrectly or the linearization tables are off.

Look at the LDRLMX section in the FDEF and you will understand what I mean.

Ok, I have gone through these sections a couple times now and I was wrong, thanks for getting me on the right track.

I guess I just got wind up as you called me out for altering maps for kicks, as I always try to get the best understanding possible before altering maps new to me. I do understand how this maps affects the system physically, I mean the ultimate function it has and what value to put there and what not.

But I admit I did never lookup how it was done before I tuned it back under the MAP limit in an attempt to hard limit my boost pressure. I was aware I did not use the maps for what it was intended to do, but this is done in many more ways, like the whole fueling enrichment via BTS or KR is all a big hack -but it works for what we intend it to do!

I assumed it would be based on actual boost and possibly drop the WGDC or cut the throttle, and since after I tuned it back down I did get throttle cut when I over boosted, I never looked it up. A lesson learned (again).