Julian's '91 200 20vt

[88a5tq]

For the love of Jesus!! I'm about to pay you to do this lol.. No cuz then id have to change the sticker to iTried

I see what youre saying. Apparently those PID #s are quite LOW. I guess I was having a difficult time at first trusting the PID response time. I really loath the boost control part of this mainly because I have to find appropriate areas to smash on it all while "going somewhere" in town in order to keep my coolant and MAT values in the normal range. This year I accidentally cut almost all my hair off so I cant very well pull it out lol. I'm on top of this. Thanks for constantly stopping in to drop a few lines. I'm really hoping the newer guys like Kevin or whoever will be able to use the accumulated advice in the future.

And yes, I slammed into boost cut... in front of a cop bahahha! It was the only thing that kept me just 2 miles over the speed limit too. I did tweak the table again after that and happened to not be logging when it occurred. Educated guesses commenced.

[loxxrider]

Haha, I'm happy to help and educate on this topic. It seems not many understand it well and it's something I really do enjoy teaching. I used to be a T/A for the controls lab at the University of Florida

[Aktapod]

And I'm most certainly watching! I'm trying to retain what I can, but even then, this thread will serve as a great reference source when the time comes, so I cna hit the ground running!

[88a5tq]

Cool that's good! I was sure hoping for a similar resource when I started haha.

I've found that my 2 bulbs closest to the grill don't light up. On the 5k they would come on with the first click of the headlight switch. A different set of bulbs do instead though. Is my switch to blame, or more likely, the bulbs are fried?

[pilihp2]

Please do continue. It helps those of us, like me, who don't know as much as we should know, learn stuff that we should know.

[88a5tq]

So you want me to continue to not know what I'm doing because its for the "greater good." Can't be too difficult

guinea pig

[Aktapod]

Regarding your lights, I'm fairly certain that's due to the difference in the Euro and DOT lights. Your 5000 should have the parking lights next to the grill, whereas E-codes have the running lights in the main reflector housing, along with the main H4 bulb. I think the Euro's use that centermost reflector solely for the high beams, but don't quote me on that.

[88a5tq]

Also from S'ingTFA's it looks as though nobody has swapped the Apikol short shift parts into a 3B for the 016. Only have seen where it was used after a 6 spd 01e swap. Oh well. Guess I'm tearing the 034 shifter out of the car I don't drive anymore.

Thanks Kevin. The euros are so new to me. I will check their bulbs soon enough

[88a5tq]

Very interesting. Only when the headlights are on will the inside bulbs light up if pushing the high beams forward to on. With headlights off and the high beams clicked forward to on, NO lights will illuminate in the headlights. With headlights still off and pulling the switch back to flash high beams..... they flash lol. I'm only one person so I can't say WHICH bulbs turn on when flashing them. With the headlights on, flashers still work the same. I'm convinced this is the intended operation since it seems too perfect to be a glitch hah. I find the part about clicking high beams to on the most interesting though. Only get the full effect when your headlights are on.

I have been slamming into boost cut all day! Chris called it. Of course I disabled "100% until" altogether. Probably something to do with this.

[loxxrider]

It's because of your refDC table. The values you have during boost onset are too high. 75% DC is OK at high RPM because the turbo is falling off up top, but at low RPM it's way too much (the turbo wants to boost it's balls off at that point).

[88a5tq]

Yep I just continue the 60% trend down further to mid range. Logs don't lie. ~80% was where I was pealing my face off the wheel

[88a5tq]

Unbelievable. It took me this long to figure out how to fix my boost control in 20 minutes hahahah! Based on this pull below, I made changes to PID values which I included to the right again. I have had D at 30 for ever since I can remember but basically, as per your suggestion many moons ago, I was set to zero and P was adjusted up or down based on whether I was hitting boost cut. When I got within 3 kPa of my target, I was increased for fixing the errors. Whew! You'll also notice that 100% until setting is still off.

Boost PID.png (202.74 KiB) Viewed 26326 times

I'm quite impressed with the sampling speed. It must be pretty damn often for the system to be able to react while I'm still in first gear. Yeah and I'm not exactly hyped about that 0.75 lambda during the run. I went looking for my paper with the w/m mixture equation I figured out so that I could calculate the exact amount of water to pour into the reservoir to achieve a meth mix of 25%. Its gone , likely happened during my move 2 years ago. I may just pour the rest of the 1/2 gallon of water I have sitting here and see.

Also my idle has been overly rich ever since I dropped it from 1000 to 850 RPM as an experiment. A fleeting vibration was the result and I don't particularly like it that low anymore. Got too used to the higher idle after all this tuning.

[Mcstiff]

Did I miss the issue with 100% until? Your refDC table is doing the same thing for everything under 200 kpa.

[88a5tq]

It's cheating. Chris says haha. He mentoned the choppy nature of the DC table and that it needed revamped since it, combined with my minute PID, were not helping with boost "control." I may turn it back on to see how spool is effected.

Actually I think I just was using everything wrong and I needed to go back to basics. My ref DC table was too precise and my PID values were too small to be allowed to control anything. Now that my table is improved and the PID is actually effective, I could probably try toying with that 100% until parameter again. Will revisit tomorrow.

[88a5tq]

Oh yeah I guess instead of turning that setting back on, I basically just set it at 100% anyway
I have the ability to be an idiot on demand.

[loxxrider]

lol yeah... There is no issue with having the 100% until on if you're using it correctly, but like you said, it is a good idea to get back to basics and get the boost control working properly without that. Then apply that setting as the icing on the cake if needed.

Ed is right that you were already doing the 100% until thing with your refDC table lol. The other issue is that you can't really tell if your PID has any authority over anything when you have these other settings going on. The idea is to get away from bandaids like "100% until" and the refDC table in general if possible. It looks like it isn't really trying to do a whole lot still (the PID values). Anyway, I'm sure you'll get it figured out. You're on the right track now.

Let me put it this way. If tuned properly, the PID control system will ask for 100% duty cycle while the turbo is trying to spool up. There would be no need for 100% until.

[88a5tq]

That last part was pretty helpful. I was never sure it would make that decision itself.

So I cracked my windshield while trying to muscle the MFing 034 shifter in..

20150826_183301.jpg (2.11 MiB) Viewed 26264 times

[88a5tq]

Oh god dammit!

$200 deductible. Filed and appointment for Friday morning. F### me I'm stupid when I'm in the zone

[DE80q]

Um... How did you break the windshield trying to change out the shifter?

[88a5tq]

I'll give you 3 different options. Hint - it's not always the most detailed scenario:
A) I meant to
B) I was laying on the roof kicking my feet while trying to reach the shifter inside through both front windows. It was at that point where you have to figure out how to get the lowest, spring-loaded, round part of the shifter back into its socket. You don't WANT to take the long way but fear you might NEED to in the end but you've convinced yourself you have a shot anyway. Suddenly, and without warning, it cracks before my foot even hits the glass again.. !!!
C) same as B above except I was in the front of the car with one foot inside each foot well and I'm applying my entire body weight to the shift knob. That's not cutting it so I brace myself against what feels like the roof. And, like giving birth, I push with all I've got... right into the windshield and the sound of it cracking makes me die inside. Cursing commences.




But, let's be frank, I did it on purpose. I have no use for low insurance payments in the long term or $200 in the short term.

[88a5tq]

Also, after I finished with tweaking the shifter, I went to taco bell on a whim. Problem was I left my wallet in the house before I even started on the car. As I returned home, I decided to drift out at the plaza's light. Got my wallet, returned, and as I'm walking into the bell again, I notice a state trooper creeping up on the plaza light and go into stealth mode. Got food, go to light, read Dave's post on my phone right quick, forget to signal, redline out and AWD drift the hell out of there and he lights me up and writes me a $600 ticket for all kinds of crazy shit!



Nope actually I put the phone down and signal at the last second before leaving and see the 2 man team of troopers sitting to my left ready to light me up for lighting 'em up. Lol couldn't resist. I'm a liar but I always tell the truth eventually

[pilihp2]

wtf

[DE80q]

I actually started to feel bad for a second.

[88a5tq]

Dude this whole thing feels akin to breaking a tail light while wiring a stereo head unit

Lol ya but I bet they woulda pulled me over just to high five me

[loxxrider]

Lol you're crazy. That windshield thing is very common with BMW owners because the shift knobs are push-on rather than thread-on. I don't know who the hell decides to make a push on shifter. Thread-on is so much better!