But, I wouldn't be so quick to do it. My axle seal is leaking at that particular tripod. It might require a bearing/seal assembly, or perhaps I just plain have a leaky seal. Although it works just fine, this issue did come up. See a few posts above for that.
Great news today! I bypassed all conditionals for Quasi-Asynchronous mode, affectively forcing it to run in that mode. Heres a little clip of what the difference sounds like:
you can't hear any change in the quality of the idle, which is great. It was running pig rich in this video, otherwise it would have shut down when transitioning to normal firing mode.
Then, I just went ahead and pulled back the fuel, and added in a little bit of timing.
Now, I just need to figure out which conditional is failing to allow quasi asynchronous pulse, and fix that. I think it's the road speed to disallow QAP, which is set to 255 mph (at L86C2). That may be set to disable QAP entirely, where it APPEARS like it is set to be active all the time.
Either way, It looks like I can probably restore fuel pressure back to stock, as this tuning roadblock is lifted. Now, It will even idle in closed loop, but the corrections are still far to aggressive, I need to figure out which "Idle Proportional gain count" or whatever it is that I can dial that in with. It's making the AFR's go all over!
Fierobsessed
DEC 05, 05:44 PM
It's been a little while since an update. Tuning, tuning, tuning.... That's what my life revolves around nowadays.
However, having done a TON of work on the program, I am pleased to say that I've got the idle sorted out. It wasn't easy, and I burnt a half a tank, JUST IDLING! The bottom line is this: Today, the state of Nevada recognized that a Turbocharged 3.4 DOHC with 60lb injectors easily passed emissions without any hiccups.
Our test is three parts: A 90 second high speed emissions check at 2500 RPM's in neutral An idle emissions test. Visual inspection of all emissions components.
Since I have all the factory emissions equipment installed and functional. The visual was an easy pass, but they did want to see the cars emissions label, which I could not provide,"Oops, sorry, no the label was lost a long time ago" Even though it was simply taped over. They did not know or figure out that it is an engine swap.
This all makes me feel proud of what I have done. I'll get to the chronicle of tuning, and where I stand so far later on.
Fierobsessed
DEC 05, 10:19 PM
The craziness that is tuning, Let's discuss!
Like anything that gets tuned, you start with a base file. Mine is from a 90 Grand Prix McLaren Turbo, It is an $8F mask, ARZC code running in a 1227730 ECM, which is the in-car version of the 1227727 that was used in the 89-90 Turbo Grand Prix's. This ECM has everything I need to do this whole project. Conveniently, it is a very common, and cheap ECM.
I also know that there are three other tunes that contain valuable data for guiding the tuning of this engine.
$DF, BCFA which is the 91-93 Grand prix's with the 3.4 DOHC, and a manual transmission, uses a 16196396 ECM, which is similar to the 1227727, and 1227730, but it has some extra ram. $6D, ATWX which is for a 89 Cavalier 2.8 with a manual transmission, Used in the 1227730. $6D, ARYH which is also for a 89 Cavalier 2.8 with an Automatic transmission, Also used in the 1227730.
The two $6D codes were especially handy for helping show me what GM changed in an otherwise identical car, to make them compatible with a manual transmission. $6D is nearly the same code as $8F, minus the boost stuff, so this little treasure of information was very helpful!
I have a hack for $8F and created most of one for $6D, so it wasn't too difficult to compare the differences to figure out what might be changed for a manual transmission. I have only implemented a couple of the changes to my current code.
The very first thing I did to the $8F AZRC code was change the table called "Base Pulse Constant vs Desired EGR", which were all set to 105, down to 42, which reflects the change in injector size from 22# to 60#, as well as the change in displacement from the 3.1L Turbo Grand Prix, to the 3.4 DOHC, This allowed the engine to actually run. Otherwise, it would have been FAR too rich. It still injected FAR too much fuel to start, so I could only start it with the throttle pressed to the floor, which shuts off all fuel during cranking. (clear flood mode)
I decided to decrease the cranking fuel to 40% of what it was, since that is roughly what I did to the "Base Pulse Constant vs Desired EGR". The table that I used for this was "Crank fuel PW vs Coolant Temperature" I just grabbed all the values and multiplied them by 0.4. Now the engine starts right up.
The next change I found I needed to make was to disable closed loop. It would surge and stall when Closed loop would start to make fueling corrections. This was simple. I changed the "Temperature threshold for C/L determination" from 80.15F to 304.25F, effectively disabling it. This made it run better, but not good enough to drive yet. I also decided that it would be a good time to disable Charcoal Canister Purge "Disable CCP if coolant < this" set to 304.25, and the EGR "If coolant <= this Disable EGR" also to 304.25. I'll deal with those later.
I was having issues where I would hit the throttle and it would stall every time, with the WBo2 reading pig rich when I'd open the throttle. I went "Async Factor Vs Coolant" (under Fuel AE) and multiply them by 0.4. This allowed the throttle to be opened without any bogging. At this point, the car suddenly became drivable. And so I did some breaking in, and had some fun with it.
I have a bit of a dislike for the way That $8F (and $6D, $DF) calculates it's Volumetric Efficiency. It has 3 tables. One is a "Base VE table", which just looks up an RPM and coughs out a simple VE value. Then, it goes to either the "Idle VE table" or the [Open throttle] "VE Table" where it looks up a second number based on MAP and RPM, then adds it to the Base VE table value for the final VE number.
Seems silly to me.
Why not just add the base VE table numbers to all the values on the Idle VE table AND the [open throttle] VE table, and skip having a Base VE table altogether? So that's what I did. Except, I made a new VE table that goes up to 6400 RPM, stuck it in an open spot on the chip, and re-directed the call to it. As well as opened up the RPM lookup cap from 4000 RPM to 6375 RPM so that it could see the whole table. This makes tuning much more straight forward now.
With that sorted, I Imported the main VE table, and the Idle VE table from $DF BCFA, and made the necessary changes so that it would fit the new tables. Then I made a new "Main Spark Table" which was based on BCFA's Main Spark Table in the Naturally aspirated section, and blended in the boosted timing table from $8F AZRC.
This gave me a good basic tune to start playing with.
This seems like a lot, but it is where the tuning STARTS.
My car's inspection was due Dec 7. So that set the deadline to get it running well.
The first thing I tried to do, (and failed at) was to get the engine to idle in open loop, at 14.7:1 on the wideband. All I had to do was decrease the Idle VE table till it would hit 14.7:1 Only, it would die anytime I got near 13.5:1 because the injectors couldn't inject a small enough pulse before it wouldn't inject anything at all. In the end, the adjustable fuel pressure regulator wasn't even enough to get the fuel trimmed back. What I found, was that the engine wasn't running in Quasi Asynchronous Pulse mode, so it was firing the injectors once per engine revolution. I modified the program a little to ignore the MPH disable parameter for Quasi Async mode. This change caused the injectors to fire every other revolution, effectively allowing me to trim the fuel back further. (see video above) Suddenly I was able to adjust the Idle VE table down till I was idling at 14.7:1
With that issue settled, I tried enabling closed loop. "Temperature threshold for C/L determination" from 304.25F to 80.15F. The engine started surging with the closed loop corrections. But I noticed that the INTegrator was also fluctuating heavily with the surges. So clearly the INT was changing too fast, causing the corrections to be too large. I went to this table: "Integrater Update Delay Vs CLFLOW" and multiplied all the values by 3.0. This slowed the INT corrections down enough to keep the magnitude of the corrections reasonable.
One thing that I decided to do to help idle stability, was to import the idle RPM tables from $DF BCFA. This made a nice difference.
I realized that this engine REALLY doesn't like when the cooling fan kicks on. It would stall almost immediately. This one had me scratching my head for a little while. I combed over the logs carefully and noticed that the IACV jumps open a few counts just before the fan is actually commanded to turn on. I found a couple of parameters: "Fan 1 Turn on delay" (0.3 second) AND "Num of IAC steps for Fan 1" (which was 10 steps). So I increased the fan delay from .3 seconds to 6 seconds, to help me see how the IAC steps affected the idle. So I saw the IAC bump open 10 steps. which caused a slight idle surge, then the fan kicked on and the engine died anyways. I bumped the IAC steps up from 10 to 15 and tested it again. It had a slightly larger idle surge, and I set the Fan 1 turn on delay to expire exactly when the idle surge peaked, which was 1.6 seconds. This solved my stalling issue.
This is where I was this morning. It ran well enough to pass emissions. I'm far from done, but getting a car to idle correctly certainly makes me happy, and saves a lot of gas.
One other thing I noticed today that may have a surprise effect on tuning, I've been using a Memcal from a 3.1L N/A aluminum head "vin T" engine to do my tuning. I had a factory Turbo Grand Prix Memcal hanging around that I would re-burn when doing my final tune. I knew It's resistor chips are identical between the 3.1L N/A and Turbo. But it turns out, that the ESC filter on the Turbo memcal is different then the N/A. This is VERY important. This means that Knock was probably being incorrectly reported to the ECM. I was seeing unreasonable knock out of the engine constantly under low boost. I have a feeling that putting in the Turbo Memcal is going to make a big difference in how I tune timing.
Fierobsessed
DEC 12, 07:53 PM
Been having a ton of fun with the car lately. Lots of tuning to do still, I've been working on the main VE table a bit.
I had a couple of problems pop up. I had a fuel leak near the fuel filter. It turns out that the barb that I was using to go from the nylon tubing to the fitting for the fuel filter wasn't aggressive enough, and the fuel was leaking around it. The barb was from a "Nylon Fuel hose repair kit" that you can get from Autozone. So, I'd recommend against those crappy barbs. I replaced it with a barb that I cut off of a metal quick disconnect end, and welded it to the fuel filter fitting. Problem solved.
The other issue was the intercooler pump. I was using a cheap generic pump that came with my intercooler. It was making some funky noises, and I suspect it was possibly leaking too. So, I bought a Bosch "Cobra" pump and installed that instead. Slight difference in size and quality...
On the upside, I did a bit of driving, a bit of WOT fun, and let me say, It's pretty fast at 10 PSI. Leaves me giggling maniacally when I look at the speedometer and it just blows through 40-80 in just a few seconds, then 80-100 in just a couple more. I've only driven a few cars that accelerate this fast. It's quickly getting to the point where it may be the fastest car I've ever driven. But it's not quite there yet...
sleevePAPA
DEC 12, 10:21 PM
I have to see this thing soon lol.
BV MotorSports
DEC 24, 10:52 PM
Any luck on that dipstick?
Fierobsessed
DEC 27, 02:02 AM
Merry Christmas Steve.
BV MotorSports
DEC 27, 07:27 PM
You are a true craftsman, I must pay you for your wares!
Also, that bosch pump can be tricky to prime. I run the same pump for my w2a system. The only thing I dont like its the old school GM fuel injector style retaining clip. Otherwise, its a great bit of kit.
[This message has been edited by BV MotorSports (edited 12-27-2013).]
Fierobsessed
DEC 28, 02:07 AM
It's not uncommon for people to have trouble with those clips. Many try to remove the retaining clip with a screwdriver. If that's your problem, you are not alone. You need to press the center of the retaining clip in and it will release the connector...
The dipstick is a gift. It wasn't a big deal to make, I just needed to get off my butt and do it.
Priming that pump wasn't too bad. But I had a lot of issues getting the system filled though. I found that the Anti-freeze was too thick to pour into the 1/2" fill neck's hose and effectivly have the air come back up through the fill neck. It just wouldnt go! But when I was putting the distilled water in, it worked just fine. But I figured that out the hard way.
Some testing on a fairly deserted desert road. It's quite a bit of fun!
