[shushikiary]

FBO (DCI, DP's, VRSF stepped 5 inch + CP) minus meth, E30 cobb map running E35 mix, scottaz 2nd inline fuel pump.

I've been having this interesting issue where the car seems to accelerate but then do so even faster while WOT if I was on it partial throttle before. Usually this happens around 4k RPM or so, and sometimes under partial throttle during "spool" at around 1800 RPM. I've been trying to figure out what is going on for a while, thinking maybe its throttle closures as a common problem for the E30 map.

Anyways, I finally got a data log of the event, low and behold, no throttle closure, just a sharp drop in timing for a few 100 RPM and then back to go time. Here's the data log http://www.datazap.me/u/shushikiary/...13-14-15-16-17.

Anyways I wanted input on what you guys might think is going on. Sense its across all cyl and not just one its not knock, no codes, however when I was poking around in access tuner I noticed in the main timing table there is a pseudo drastic drop right at that point (about 180 load actual 4k - 4.5k RPM) where the table becomes "un-smooth" (large derivative) and am thinking that might be the issue.

I've had similar problems with the OTS E30 map with the boost taking long to come in at this altitude (Denver) due to the PID controlling the WGDC and not the base WGDC map so I re-scaled the entire WGDC base table to support a boost setpoint factor of up to 2.6 and a maf requested G/S of up to 450 and smoothed it out to get much more of what I want, I also copied the 335is map's fuel scalar table over and modified it so my LTFT's are under 5% all the time and upped the torque limit tables as they were causing me some issues at times as well..... ok so its not really the OTS E30 map anymore its my own map, but this issue was still there on the OTS E30 map.

Any idea's on anything else I should try are appreciated.

[Aerotest]

It is because your boost is overshooting the requested. Scale the boost factor multiplier or reel in the wgdc base to keep it on target. Timing is connected to boost and you will see a global reduction when boost is significantly overruns the target. Logs look good otherwise. When I experienced this phenomenon the timing reductions didnt show up as corrections (read 0.0). I think timing reverts to the safe table when this occurs.

[Tzu]

^^ This is correct from my testing. Although I got boost actual to be below setpoint, it was still close enoguh that it still causes timing drops, with no corrections. Root cause is my wastegates are abnormally tight.

[shushikiary]

Ok, Thankfully I have MAF G/S and boost setpoint factor on that grab so I can go play with the WGDC base tables to see if I can fix it....

After looking at the WGDC tables I think I'll try setting the D-factor multiplier first as you were saying, might getter a tighter response and reduce the overshoot, or are you refering to the boost limit multiplier?

Thanks!

[Aerotest]

Yes Boost Limit Multiplier.....sorry. IMO D-factor wont fix that overshoot b/c it is already dampened (stable). The PID control cant be effective if it is being pushed upward by too much base. The PID has limited authority. The goal is to have the base table mapped so the PID is close to centered when on target. This allows flexibilty when operating in other conditions than a single gear log pull. IMO if you want to play with PID, P-factor would be the more appropriate value to adjust. If you dont want to hack up your base table, try reducing the WGDC Adder. This is an easier table to work with and I find it quite effective.

[shushikiary]

Ah, so the boost limit multiplier allows the boost error to be larger before it intervenes, and the WGDC adder table adds WGDC based on MAF requested G/S.

Given what you said about how the PID is supposed to be a fine tuner and not the main controller I think I'll leave the P values alone.

I did play with the D factor multiplier and the duration of the negative timing spike got much shorter because the PID reeled in the overshoot much faster, but I agree with what you're saying so I'll put it back and try messing with the adder table first, if that doesnt work then I'll go play with the base table if I have to - it took me forever to get that base table how I wanted it after re-scaling it and I dont want to play with it again if I dont have to.

thanks!

[Aerotest]

Right on. BLM controls the target boost curve otherwise you everything else right.

[shushikiary]

yea I just tried upping the BLM to 2.7 and I was hitting 18 PSI (which is crazy at this altitude) peak around 3.5k RPM and then the timing would pull and drop me back down. It was fun, but I wont want to over spin the heck out of the turbos. Reducing that would make me sad I think, so I think the right course is what we already said... adjust WGDC adder or reduce WGDC base.

[Aerotest]

I run mine at 3.0 at near sea level. Hits 19.2 psi even on cooler days.This is fine for me given air density but agree you should probably not do that if you like your turbos.

[shushikiary]

I played with the PID, I very slightly lowered the WGDC base, but added a decent amount of D factor multiplier and then added 10% WGDC P to + and - 0-0.05 bar ranges and got my self a very nice fast rising not overshooting curve with no more timing pulls when I'm having fun in the corners . (ok no overshoot a initial point, just mid range overshoot during taper, but I'm much more ok with that).

http://www.datazap.me/u/shushikiary/...13-14-15-16-17

[Aerotest]

Nice work!

[shushikiary]

Ok... I've been playing around trying to get rid of, in every case, any timing drops, I was having boost overshoots under some circumstances like downshifting and flooring it, etc, and finally landed on a boost setpoint factor of 3.0 which made the problem go away, but then I get this:

http://www.datazap.me/u/shushikiary/...12-13-14-15-16

Hardly any boost overhsoot, yet I still get a huge timing drop right at the start... not sure at all what is going on... its very strange. Usually it only happens when the weather is cold out.... I wonder if I'm hitting a torque limit or something... Or maybe my load actual is exceeding my load requested by too much?

I also found an interesting article about checking the disable torque intervention by ign check box over on the BB forums.... Though they say only to do that on an MT (thankfully I am).

[CarAbuser]

Originally Posted by shushikiary

Ok... I've been playing around trying to get rid of, in every case, any timing drops, I was having boost overshoots under some circumstances like downshifting and flooring it, etc, and finally landed on a boost setpoint factor of 3.0 which made the problem go away, but then I get this:

http://www.datazap.me/u/shushikiary/...12-13-14-15-16

Hardly any boost overhsoot, yet I still get a huge timing drop right at the start... not sure at all what is going on... its very strange. Usually it only happens when the weather is cold out.... I wonder if I'm hitting a torque limit or something... Or maybe my load actual is exceeding my load requested by too much?

I also found an interesting article about checking the disable torque intervention by ign check box over on the BB forums.... Though they say only to do that on an MT (thankfully I am).

I don't know what's causing the timing drops but I did notice a few things in that log.

Your AFR is really lean at the start of the pull. Is that normal for E30?

Your actual boost isn't following the req boost and is slow to build boost.
It seems like the PID control loop has little effect on WGDC position, it's just the base table that's controlling the boost pressure.

[BQTuning]

Per the last log in post #12

Too much WGDC ceiling and or WGDC Adder in the range where the drop occures. Those timing drops is not a timing correction, its a timing reduction based on a limit or an imbalance in whatever it is your trying to do.

[shushikiary]

In post 12 the log is of me doing a partial throttle starting off in 2nd gear and then flooring it, so that's why the boost looks like its taking so long to build. For AFR at the start its around 12:1 and then it goes to about 14:1 and then back to about 12:1 near the end, its always done that sense I've been on the E30 map.

So BuraQ, you think that even though I didnt boost overshoot the WGDC base table (the adder I have left as stock) is still too high and perhaps its hitting its max WGDC (57 % I believe) so that's why its pulling timing? I could go in and further reduce the WGDC base at those MAF GS / boost setpoint factor points (or adder as you've said). I'll have to give that a try.

OH! You literally mean the WGDC Ceiling (adder) table and the WGDC Adder (airflow) tables.

From the description of the WGDC Ceiling table, it says " Tuning Tips - Raising this value will allow the system to achieve a higher final WGDC. " which seems to suggest I'd want to increase it so I dont hit the WGDC limit, however if I understand you correctly you're saying I should lower the value so I dont hit as high of a WGDC?


Update: I lowered the WGDC Adder (airflow) at the MAF G/S that is having the issue by 15% and no change ... I'll try upping the ceiling and seeing what I get.

[shushikiary]

Oh hey look what I found: http://www.***********.com/showthrea...-Boost-Control (note this is a BB forum post).

Looks like I should increase it as it seems if I am hitting the limit in the ceiling table that might be why its getting so upset, even though the boost doesnt overshoot target, perhaps the requested WGDC is so far over the ceiling that its pulling timing.... I tried increasing the values in there by 5 above the first 57 entry, that seemed to help a little, and I've been monitoring it in real time with my AP, and I do see it get maxed out when the timing is pulled for that MAF G/S.

I'll try adding some more to the ceiling table and see if it helps more.

[shushikiary]

Added yet more to the ceiling... it seems to have helped the issue still pops up sometimes. Especially if I just came off partial throttle, so here is a second gear pull of me down shifting into second, going around a corner then flooring it out of the corner, you'll see that right when I floor it the timing on cyl 1 goes up to 18, but then backs off to about 12 were it should be then drops before I even hit boost target. Note that I upped the WGDC max to 70% as well so it peaks out there during spool.

Interestingly the timing spike happens right when the torque output pops over the requested torque, then it settles out. I suspect that there is something to that. I'm wondering if the load to torque tables (which isnt a limit table AT ALL, but a table that says what torque to report based on what actual load it seems, might be causing me to report over the requested torque and is part of the issue... what's interesting is that I dont get the timing drop every time it happens, but the timing drop always occurs at that inflection point.

http://www.datazap.me/u/shushikiary/...ata=1-2-8-9-10

I'm half tempted to just copy the timing main table over the timing fail safe table just to make it stop being stupid.

[shushikiary]

I feel like I'm half documenting this (posting so many times on my own topic) so maybe it helps someone else (I always hate it when people have an issue but never update the thread as they progressed and found a solution them selves).

I've now been able to show that the issue consistently happens when the torque output exceeds the requested torque by more than 100 n/m and I'm between 2500 and 4k RPM. If I'm above that RPM when the torque output exceeds the requested by more than 100 then no timing drop.

Interestingly I can only really reproduce that situation if 1. I floor it off a partial throttle before hand (this allows the turbo's to spool before I floor it so I get higher power earlier) OR 2. I shift and the boost spikes and when I let the clutch out the RPM goes a decent amount below 4k.

Given that, the problem got a little better when I increased the torque reduction factor (RPM) from 0.898 to 1.0 in 2150 to 7k RPM. From other reading I have done this table, when the values are increased, reduce the amount of response the system has when the torque exceeds the requested torque. Interestingly I got that info from a post from BuraQ on another forum. I'm going to try raising this limit by a lot to 1.6 from 2150 to 7k RPM and see what happens. In the previous attempt I had also copied the ignition main table over the timing safe table to try that as well... which didnt seem to do much so I'm copying it back to the original for this next test.

[Tzu]

I tried a lot of what you're doing, but didn't dive too much into the torque tables. Interested in your results.

[shushikiary]

I adjusted those like I said, it removed a lot of the problem but not completely. So next I tried copying back over the fail safe timing table as well, but I still have the problem every once in a while.

I could try upping the WGDC ceiling yet again, but its already 15 points above what it was originally (i.e. 15->30) and I'm hitting boost target just fine, the adjustments I made to the PID work great.

It seems to happen MUCH less often once oil temp is up to full 230 or so deg, I'm starting to wonder if its some temp limit due to CAT or TB temp, but that's simply speculation based on correlation.

I'll take another look through the tables to see if there is anything else that is suspect as well, it always only happens when I'm partial throttle then go full throttle and the output torque exceeds the requested torque. So maybe I can do something to make the calculated torque report lower as well. I'll most likely email COBB as well to get their input.

I'll keep the post updated, but its possible, as other have found with the post shift bog, that the only way to fix it is to use a piggy.

[WedgePerformance]

Originally Posted by shushikiary

I'll keep the post updated, but its possible, as other have found with the post shift bog, that the only way to fix it is to use a piggy.

I'm not sure why people think the only way to fix it is with a piggy. I have no post shift bog and I probably make more torque than 99% of the cars out there. Most piggy flashes use a low load request and therefor load to torque is calculating a much lower torque output. There are a few tables that impact this. You also have limit tables which have a big impact on this as well. If you can't get it, shoot me a datazap log to look at.

[shushikiary]

Oh, well that gives me hope! I see a few torque limiting tables in there, the only one that looks like it might be the one I want is the Torue Limiting Maps: Requested Torque Limit (driver), which I might want to up. I've only slightly played with the load to torque map, but I could try reducing its values even further to reduce the calculated torque output... Note that I did check the box for "disable torque intervention by ign" as well.

Thanks for the encouragement that this can be fixed without a piggy!

Ok what ever you do DONT do what I just did, I modified the torque limit offset to 700 from its normal value of 389 and decreased the entire load to torque table by 30%... car got very unhappy with me and blew a DSC 4x4 code, wouldnt even hold idle and I had to flash back to the previous map for the car to be drivable again even after clearing codes. Pretty sure its not the table we're to raise, maybe lower? That or the load to torque table really is a limit table and by lowering it I exceeded what it expected so much it triggered the code.

Tried leaving the torque limit offset at 700, that works fine, me reducing the load to torque limit is what made me blow a code. I'm starting to think it really is a limit of some kind based on that.... that or the DME knows if the value is too far bogus.