Julian's '91 200 20vt

[88a5tq]

That is truly sad and sorry I pressed the matter or even brought it up. I find my line of work has made me more emotionally attached to animals that people so when our cattle dog of 15 years died the year after my father, I hate to say it but the dog hit harder for me. Life can be strange like that.

As for the tune, I'm on my 24th iteration of config files for the new turbo! That's 24 cfg's with extensive changes made for each one. I can imagine the first 19 all had to do with me chasing my tail on the boost PID and only due to my own errors lol. Following the log below, I made further PID/fueling tweaks. I started the evening by fixing my warm up table and quickly found that my fueling was way messed up and lean but mostly at 140 kPa and under, I have brought lambda up to 0.81 at high boost though and I have to say the car really felt CLEAN and strong! Seemed very efficient and pretty quick so I'm happy.

v3.3_u009216-2015.09.01-19.16.58.vemslog

(1.5 MiB) Downloaded 1037 times

Recap: ME -> <- Boost PID

Glad that battle is over

[88a5tq]

They won't be able to save the factory chrome trim around my windshield unfortunately. I have to settle for some generic black s***

Part is NLA they say

[loxxrider]

Things are getting better, but your actual duty cycle is doing some funky stuff still. We need to continue to eliminate variables. I see you've raised D a LOT. That may be the cause of the strange DC response we are seeing.

I suggest you do a few things for variable elimination purposes. Set all of your TPS scaling to 100% (it's more fun this way on bigger turbos anyhow ). Also set your whole boost vs RPM curve to the same thing. 270 will be fine.

Then you need to try PID tuning again. Start with P. Turn it up as high as you can until it starts to oscillate or overshoot too badly. At that point, back it off a little and add D little by little. I will come in later to help correct for steady-state error (it will stop boost from tapering off on the top end).

[88a5tq]

Oh ok, I guess fun would be alright for awhile.

It shouldn't matter I don't think, but my wiring to the n75 is run through a relay's 87a terminal so it only cuts when the relay is signaled by the w/m spray controller in the event of a failure. Failsafe.. for safety!

[88a5tq]

Ok so I'm finally seeing 100% boost DC, but its only for a whole 5 hundredths of a second. Its irritating. So I had to crank P all the way to 175 just to finally hit the 100% at all. I don't think theres an oscillation. I see the taper for sure and that's with I set to 30 even. I followed the other ideas you had as well Chris. I do appreciate your dedication to assisting this forum!

Fueling was changed in some idle and cruise areas. I thought about changing the injector size setting and reverting to my old fuel VE table to see if it was corrected all through the rpm ranges. One step at a time so Ill stay with the boost control for now.

Sadly, my log with the 175 for P and 30 for I didn't have a 100% tps where I hoped. My fault.

v3.3_u009216-2015.09.02-18.41.44.vemslog

Beginning of evening with only P enabled

(252 KiB) Downloaded 1021 times

v3.3_u009216-2015.09.02-18.50.07.vemslog

More Recent

(123 KiB) Downloaded 1011 times

[loxxrider]

OK

This isn't your fault at this point. I'm really disliking this new boost control scheme. It doesn't work well with a stock wastegate. That "PD control (far from target)" setting seems to be the culprit. You can see that as soon as you are -80 kPa from your target, your closed loop control kicks in. I guess we will either have to work around this or have to talk to the VEMS developers/try a different firmware, etc.

That setting is supposed to disable "I" until you are 80 kPa away from the target, but it seems to be disabling all closed loop control (P, I, and D) and defaults to the refDC table values.

One bandaid we could put on this would be to go back to using the 100% until command and set "until" at -60. Then set PD control (far from target) to -60 kPa as well. This should get you your fast spool and then allow PID control to kick in as you approach your target.

Also, don't be sorry about the dog story! I don't mind and find it interesting to hear about things other than cars all the time lol.

[loxxrider]

By George! I think I've got it!

I have made a discovery.

This is an entirely new control scheme which is selectable and not based off of a "simple" PID loop anymore. Whether these things are a good idea or not and are for a user more advanced or more basic in knowledge or not, I'm not sure. In any case, you can change "Reference lookup" to the old style if you so desire (RPM, gear, speed-based). That is how I've been trying to get you to tune it this entire time. I was not aware of the way the new boost control algorithm works, but I am now.

So the bandaid fix I was talking about before is actually the way it is supposed to work. Whether you choose to use the old system or new is up to you. I think you may have success with the new system, at least from a spool standpoint as it's a more stable system by design. I'm still not sure I like it because a properly tuned PID-only controller would be just fine here, but I understand what they're doing and it does make sense.

So either carry on with the "bandaid" or switch to the old system. You'll get the results you desire either way now that we've got this hammered out!

[88a5tq]

I think you hit the nail on the head Im seeing 100% DC now and Ive decided to use the new boost control method you described. I am working on getting a good log still since I was still slamming into boost cut last night. I suspect Ill need to reduce my request back to 250 kPa again for dialing in P, at least for a little while. I had a lot of fun but it was cut short by my laptop battery and me still having issues hitting fuel cut.

Thing is I was using a basically new ASUS Eee netbook (1001PXD) that my boss gave to some co workers Im not going to name. She ended up taking them all back since they were all looking at PRON at work My gain! Rest assured it still gets used for pron hah. JK (like my inititials). SSo this netbook works brilliantly for tuning and its compact. I just ordered a car adapter for it so I never run out of battery again. So happy to have this thing. I did format and set it up how I like though. Its good stuff.

Ill be working on a good log again today. Need to bring P way down where it will behave.

[88a5tq]

It's so strange. I hit boost cut 5 minutes ago hard while tuning and all the sudden my brake light is stuck on full bright as if I have the ebrake pulled.

??

[loxxrider]

Lol can't help you there bud. Brake light switch?

[88a5tq]

I had the same thought. Looks as though that switch is a ground and it went to a gas station to investigate while I was out. Pulled the center area APART! Anyway after all that, the light was still stuck on. No more dimmer when key on (Accessory) and then brighter light on dash when pulling ebrake. I got home and further went through after identifying the yellow and grey wire earlier but this time I went in through the driver side of center console and started shifting wires around searching for it but never found it. A friend stopped in and was curious what I was up to so I turned on the ign to show him the stuck on, bright, brake bulb but this time it was dim and then more lit with the ebrake. I quietly kick myself in front of the guy and say, "nothing to see here." lol no but I wish id thought of it at the time I'm banking on the wire being pinched and grounded before the witch just right though.

So heres a "good pull" log from earlier. I literally could only achieve 21 PSI without any restrictors with 100% DC. Im still attributing this to a lack of vacuum in my intake before the k&n filter. Sigh, my intake is so afficient lol. I sure hope thats the case... So I went back to my welding wire feed tip but I revisited the largest one of 0.045" and oversized it very carefully with a drill bit in a dremel. First one was botched and much too large. That was the one that netted only 21 PSI boost. The 2nd/3rd attempts snapped my bits off inside the pieces. 4th try was just right and produced between 250 and 263 kPa with 60%DC. Now I was in the right place.

This pull log was a result of that restrictor in the pressure line before the n75 and PID tweaks. I did increase P but thats when I hit the boost cut and messed up my brake warning bulb Turned back to what you see now and added I and D. Not too bad for MONTHS AND MONTHS and gallons upon gallons of rage-floored gas.

So thanks again Chris haha!!

PS: 3rd gear pulls only for me on logs anymore. More prone to overboost issues and easier to root out and correct first off.

[loxxrider]

So you moved to the "old" VEMS boost control method I see? Seems like you've got it working mostly well.

From this log, it appears that you mostly just need to add "I" which you said you already did. That's great! Once you get this fully sorted, you might try out the "new" control method to see if it helps with spool. If it doesn't, I'd just stick with this because it appears you have a fairly stable setup going now.

I still don't understand why you are having to put these restrictors in lol. Where is your boost reference coming from? You said you attribute it to lack of vacuum in your intake before the K&N filter? You should get your boost/vac reference from the intake manifold. I'm so confused lol.

By the way, the tune is looking great and GOOD JOB HOLDING THE THROTTLE AT 100%!!

You might add some acceleration enrichment in (dTPS/dMAP). That will get rid of that little lean spike at the beginning of your pull.

Always happy to help

[88a5tq]

Wooot! Lol I knew at least one thing was going well

I wanted to give the old method a chance I guess. I'm glad you brought up that enrichment setting. I was afraid I was overusing it or it was already set well and I was just focusing in the wrong area of the log.

I'll try to clear this restrictor thing up. For the n75 plumbing, I followed the stock tap points. The boost reference comes directly from the cold side turbo nipple (taken from my old k24). The boost reference is also where I've placed the restrictor, but I put it directly in front of the n75 to keep response reasonable. The vacuum source is taken from the front of the cobra head turbo boot, very similarly to stock with the macaroni-shaped intake boot lol. I hate that thing and was glad to lose it with the new turbo. Of course, my concern comes up here because I have 2.5" to 3" intake piping ahead of that cobra elbow and a massive k&n.

So, either the boost reference is disproportionately higher than n75 can make use of or the same can be said for the vacuum source but lower. That's how I've been thinking of it. I'm very open to input haha! Are there better spots to take reference from though? TIA

[loxxrider]

Hah, that setup is... interesting...

Why do you have separate boost and vacuum reference points? Do they both go to the same port on the N75? Maybe a diagram would be helpful. Either way, the N75 should simply be "T'd" into the boost/vac reference line for the wastegate. That line should go from the intake manifold or cold side of the turbo. Period. I really don't know what the heck you're doing with this separate vacuum reference point. it doesn't make any sense to me. The N75 doesn't care about vacuum anyway. Maybe you're calling the vent for the N75 the vacuum source? You can just vent that to atmosphere Holmes.

[DE80q]

Yes, he is calling the vent one the N75 a vaccum reference. I was a little confused myself at first.

[loxxrider]

OK lets make absolutely sure you have the correct sides of the N75 connected to boost reference, wastegate, and vent.

[88a5tq]

Hah Im trying to get pics. Another post to follow shortly

[88a5tq]

Alright in the picture I'm pointing at the restrictor in the boost reference line. The right port is the "vent" line that goes into the cobra elbow. The left port goes directly to the bottom WG port. The 2nd shot shows where the vent line is press-fitted into the cobra elbow.

20150907_112408.jpg (2.51 MiB) Viewed 24910 times

20150907_112434.jpg (2.34 MiB) Viewed 24910 times

The electrical plug is facing the camera in the first shot.

One last shot unplugged:

1441639778948258740490.jpg (3.36 MiB) Viewed 24910 times

[88a5tq]

It be worth noting that my wiring for the n75 only looks suspect lol. It's fully functional. I don't think polarity matters, but that's also correct. When I went back to the n75, I grabbed the one that completely prevented air leakage from the vent port. The other had a very slow and minor leak. That's all I can think of that's pertinent.

[loxxrider]

Dude! That's a real special setup you have going there haha.

From your description "The right port is the "vent" line that goes into the cobra elbow. The left port goes directly to the bottom WG port." and based on the picture, it would appear that you have the vent line and wastegate line backwards. The long tube coming off of the N75 is for venting. The short one coming off of the side goes to the bottom of the wastegate, and the bottom port is boost reference which you definitely have correct.

If this is not how you have them arranged, then do yourself a favor and get rid of that restrictor (the N75 already has one built into the bottom port), buy new vacuum hose so you can get rid of a few of those zip-ties (lol) and hook it up per my recommendation above. Bottom to compressor or intake manifold, long side port vent to atmosphere, and short side port to the bottom of the wastegate. Vent the top port OF THE WASTEGATE to atmosphere (which it looks like you are doing already).



And yes lol the wiring does look... special... but I don't suspect that it is part of the problem either.

[DE80q]

I was thinking the same thing about the long port. I tried to find a pic Elaw posted in my thread showing how they should be hooked up, but no such luck.

[88a5tq]

DOH!! Sorry, I reversed them when I typed it. Short RIGHT port goes to bottom WG. Long LEFT port is vented to intake elbow. Sorry, I learned my rights and lefts yesterday lol.

[88a5tq]

Oh goddamit! My brake light is still on after startup

[DE80q]

OK, so you do have them hooked up properly? I finally found the pic I was looking for.

[88a5tq]

Thanks Dave. That's a great illustration too. I used the one on Scott mockery's (sp?) site when I reverted back. Yep all that stuff looks good.

I know this is not most of you want to hear, but I'm gonna pull the dash bulb lol. Maybe it'll burn out first tho