95 LT-1 Idle Cell Comparison - Humidity?

[kur4o]

Prepare to start all over with cyl trims.

I made some progress on the first cold start. Still needs some fine tunning. While experimenting it turns out that running around 20-35*C coolant at startup it idles best at 14* advance and a liitle fatter mixture.[Tuned by sound and feel.]
I will let you know the best setting I came up when it gets fine tuned.

You can monitor the rpm rise after startup to tell if it needs more or less fuel. If the rpm rise alot like upto 1500rpm it has more unburned fuel left on the walls from prime pulses.

[kur4o]

Another vote for the adjustent cylinder fuel robbing.
The lt4 settings are

Cylinder Multiplier
1 1.05 3 0.96
2 1.05 4 0.97
5 1.01 7 0.98
6 1.02 8 0.96

Somehow my trims closely resembles this pattern with the exclusion of #2 1.039 and #4 1.024.
Next on the list is to test it on running engine and starts to cut #4 to around 0.98 and see what happens.

The fuel robbing feature seems to be more pronounced on cammed car than to be related with ports size and headers lenght and style. Interesting to know the stock size idle air feed holes on a GM production lt4 manifold.

Can you source some pics on the vette air induction system like filter box, passages, maf mount and so on. Are you running some form of CAI kit. I suspect that the air induction system is the reason there are different scaling in MAF tables across platforms.

[spfautsch]

I suspect that the air induction system is the reason there are different scaling in MAF tables across platforms.

I also suspect it, times ten. In fact it's widely published that OEMs generally mock up the entire front of the car on a flow bench to develop their MAF transfer curves. Google: "helmholtz resonator". OEMs have been using them for > 10 years to smooth airflow in the intake ductwork for this specific reason (more accurate MAF transfer curves).

Here's the stock y-body intake. I believe all MAF equipped LT-1s got this intake setup. I'm also fairly sure the ZR-1/LT-5 (read: 450 hp off the showroom floor) also got an identical airbox and MAF but with a slightly different shaped throttlebody bellows.





This might be closest to a ram air Pontiacs. The only anomaly in this tract is that the Y-body engine is slightly offset towards the passenger side of the car. Anything that came with an elbow in the air ducting will likely have vastly different MAF curves.

[kur4o]

The recommendation from GM is at least 4 inch of straight pipe before the maf sensor, to get reasonable air flow readings. That is for the newer style slot type of mafs but I guess it can be applied to older setups too.

Thanks for the pictures. It looks like a funnel, before and aftre the maf and is vastly different from the f-bodies. The vette maf curve reads lower from 2-3% at low airflow to 5-6% at mid airflow upto 10-11% at high airflows, compared to stock f-body settings.

I am running the slp version of CAI and there is some problem with lower map readings at high rpm, high load at WOT, especially on third and forth gear I am wondering if it is airflow restriction of some kind. Have you measured the map drop at wot pulls under high rpm and load. What are the max AFGS you have reached with the new setup.

[spfautsch]

I haven't been doing much logging or tuning at WOT with this setup so unfortunately not much to report. I will try to pick through what logs I have but as I recall the few times I've gotten into it hard while logging were mainly corner exits or two lane highway stops / takeoffs where I didn't have much WOT time due to traffic. The majority of the hard throttle passes I've made thus far were a few un-logged joy rides I owed a couple guys that borrowed me tools and such where it broke the tires loose around 5500 rpm in 2nd and immediately bounced off the rev limit.

What I've read on the subject of MAF tuning (primarily Banish's stuff) has me thinking it will be something I touch as a very last resort if the wideband is still showing lean during WOT passes after I've removed as much as I'm comfortable with from the injector constant. The accuracy of the instrumentation I have to work with is inferior to what GM was using in the 90s by orders of magnitude, and I have a lot of unknowns like the modded 42lb mystery injectors.

[spfautsch]

Initial start + warmup after cleaning intake and enlarging #1 idle feed orifice didn't show any considerable improvement. But after an initial warmup I started working with individual trims and unlike before was able to see responses in closed loop BLMs after moderate trim changes. This is what I'm running currently for individual trims.

1	131	83
8	127	7f
4	134	86
3	126	7e
6	136	88
5	127	7f
7	125	7d
2	135	87

The last log I made (attached) had left trims at 131 and right at 135 albeit for a short time, and at higher air density / lower IAT numbers than previous logs. This is the best cell 16 split I recall logging with the current build. Idle felt and sounded good. Unlike 2 weeks ago I feel like there's potential for improvement simply from continuing to fine tune individual trims. I also removed a small amount from the prime pulsewidth multiplier to account for the trim changes and had good results.

I think there's a lot of knowledge to be had from putting an oxygen sensor in all 8 header tubes. There is definitely charge scavenging happening and I feel like we're just scratching the surface of learning the airflow dynamics happening here.

[spfautsch]

Here's a log with a WOT pull starting at timestamp 1120.0. [link]

As you've noticed it seems like it's making a small amount of vacuum at 100% TPS, but this was going downhill in 2nd so very low realized load. MAF peaked just under 320g/s @ 5900rpm. I could have pushed it another 1000 RPM but my second cup of coffee hadn't quite taken effect yet and I figured it best not to. Notice my wideband must have gotten too cold and jumped back to it's warm-up routine.

Results from logging today was promising. The idle split was back as bad as ever this morning, but BLM splits in the main driving cells were much better then before pulling the intake. I have no idea if enlarging the #1 feed hole or eliminating the oil getting sucked into 1 & 3 were both factors or only one. Whatever the case, this seems to be improvement in pretty much all the driving BLM cells.



I have one disassembly question that I would fault no-one for ignoring. I've been tinkering with the IAC adders and multipliers trying to eliminate the surge when the a/c clutch engages and disengages. It seems like it's improved in all areas except idle, where the IAC adder seems too much and the RPMs will flash up slightly before the clutch locks completely. I'm trying to figure out what the row labels should be for the table at 0x26A3. The cell values seem to be multiplied by 0x2691 and I believe the lookup is comparing to VSS but not 100% positive. Here's what I've done with the disassembly.

855A    15 89 08        @32    bclr    l_0089_, #%00001000
855D    B6 01 56        @33    ldaA    l_0156_NTRPMX_25/bit    ; rpm scaled in 25 rpm/bit
8560    13 87 20 21         brclr    L0087, #%00100000, @35  ; a/c request?
8564    B1 26 71            cmpA    L2671            ; d8 in OE cal (5400rpm?) 
8567    24 48               bcc    @39
8569    F6 01 9C            ldaB    L019C            ; one MPH/bit road speed variable
856C    C1 30               cmpB    #$30            ; 48mph?
856E    23 02               bls    @34
8570    C6 30               ldaB    #$30            ; if vss below 48mph load 48mph into B
8572    54              @34    lsrB                ; else shift register B right
8573    54                  lsrB    
8574    54                  lsrB    
8575    CE 26 A3            ldX    #$26A3            ; load pointer to 7 byte table to X
8578    3A                  aBX                ; add registers b & x ???
8579    A6 00               ldaA    0, X            ; load something into register a ???
857B    B1 02 4C            cmpA    L024C
857E    23 26               bls    @38
8580    7C 02 4C            inc    L024C
8583    20 56               jr    @44

Disassembly isn't my cup of tea - if anyone has any pointers it would be greatly appreciated.

Edit: I see a major mistake I made around the bls jump at 0x856E - I'm thinking this means this tables rows top out at 48mph?