KAMRF Variances Bank1 vs. Bank2 (Intro/Long Post)

[Coz_KB95]

Hello All, First time posting despite having been actively consuming information from some extremely brilliant folks on this forum for 5 years or more. I can't count how many times I've read Decipha's write-ups ... almost feel like I could recite them. At the point now, after tinkering with a dynotune I had done 7-8 years ago, that I am ready to start from scratch and tune the car myself. The main reason for wanting to do this is that the dynotune was done in a way where the tuner opted to have me run timing 'locked-out' as he wasn't confident he could control timing in the tune with the 94/95 CBAZA strategy. He's much more comfortable with the older ECU strats. SO ... 19degrees, set at the distributor, spout removed ... that's the way I've been driving the car for 7-8 years. Granted it's been safe and drivability has been relatively good (after I tweaked injector pulsewidths on his tune for warm and hot start fueling to get hot starts in order) but I feel I'm leaving alot on the table by not having spark control ... and I suspect that some of my temperature creep at cruising speeds/lower loads is related to the engine wanting more spark advance and not getting it. The car isn't overheating at all ... but I suspect heat is building as a result of fueling and timing based on how the car is tuned.

Where I am at today ... I've set up the T4M2 base tune, using all of the known inputs and have everything set up with TunerproRT and my QH for emulation and datalogging. Fired the car with my initial startup tune just the other day and was delighted that it fired beautifully on the first crank ... hit stable warm idle, no surging or hunting. Collected my first datalogs and overall, felt really good about my first go. At this point I am really getting comfortable with understanding the numbers and how everything connects and relates. I have a computer science/math background so it's not completely foreign which helps. Other than the car idling ~150 rpm above DSDRPM (which I gather is the setting on my throttle stop screw) everything else looks really good. What I do have an immediate question about is the KAM's I'm seeing at warm stable idle, specifically Bank1 being above 1.0 and Bank2 being Below1.0. I have my AEM 30-4110 analog output wired to the correct pinout on the ECU to log WBo2 output. Here is what I'm seeing, as I follow Decipha's example for Fuel using a warm stable idle example:

Avg. WBo2 reading from the AEM-UEGO: 0.968

Avg. Lambse1: 1.008
Avg. Lambse2: 1.010
Combined Avg. Lambse: 1.009

Avg KAMRF1: 1.023
Avg KAMRF2: 0.952
Combined Avg. KAMRF 1&2: 0.987

The Car holds a stable warm idle in and around 175 ad counts for MAF. Based on what I am seeing my MAF Flow correction for fueling at that MAF point should be (0.968/1.009)*0.987 = 0.948 ... which gives me a flow correction of 1.185 (down from the 1.25 from the known value for my SCT 2400 MAF)

Now all of this makes perfect sense to me ... however my question is ... what is the reason for this variance between my KAM corrections from one bank to the other? For one bank to be corrected toward lean, and the other to be corrected toward rich is a bit concerning. But of course ... at 5% or less either way this might be insignificant. I don't know what I don't know. So is there cause for concern here? Is there a problem I should start looking to track down? Or am I splitting hairs over something that is not really something to worry about?

Hope this isn't too much of a 'rookie' question. The bit of research I did online for similar conditions went in many directions and I didn't feel there was anything glaring I should be looking at. I smoke tested for leaks a few weeks back and already corrected a leak in the Bypass piping for the Kenne Bell ... all else seems good, at least to me.

Thanks in advance for reading and for your responses.

Coz

[decipha]

yep perfectly normal, thats the whole reason why henry tossed in two hegos and made each bank independent.

Usually once you load it and get up higher in the imaf counts your banks will come together, thats when a variance becomes an issue.

At idle and low load if you have a 10% variance or higher id investigate. Below that it doesnt make any difference.

No such thing as a rookie question

[Coz_KB95]

Thank you so much for the quick response, I got a big sigh of relief as i read it. I'll be spending some time over the next few days getting the imaf curve confirmed dialed, along with idle air and cold start fueling. Then onto Load and Spark.

Decipha, the resources and tools you have put together on this site are nothing short of remarkable ... as is your knowledge and that others I've seen posting on here. T4M2 as a whole, the k-constant, and the implementation of PERLOAD ... just brilliant.

I'm sure more questions will flow soon enough ...

[Coz_KB95]

Ok, I'm back ... after doing a whole slew of mechanical work the last several weeks to ensure my exhaust is leak free so that I could get through the tuning process without any question marks about skewed results and the like. So I am back to it, and deep into what I feel is the steep part of this learning curve and trying to understand some results and fueling behaviors to understand if I have a mechanical problem, or not.

I'm at the point where I have been dialing in the MAF using closed loop to get my fueling errors and creep up on that 1.01-1.03 sweet-spot that I've read about in some of the threads the search function led me to. In the attachments is a snapshot of the fuel error histogram from my last drive.

Not fully there ... but on it's way. Or so I thought ... until I started looking at the log data in more detail, specifically to understand how my wideband was matching up to commanded LAMBSES. Here's some relevant information ... I have not yet started dialing in IDLE AIR as I think there is a bit of refinement to warm stable fueling still to be done. This means I have not yet touched FN875N or ITHBMA. I have set the throttle screw and by and large the car seems to idle great other than a stall condition that happens occasionally coming to a stop with the clutch depressed. I am running longtubes on the car, I have not adjusted HEGO delay or bias or anything of the sort as the fueling corrections made sense while I was starting to dial things in. I have an AEM 30-4110 WBo2 in the Bank2 (driver-side) pipe ... not more than 6-8 inches downstream of the Hego in the collector of the longtube. Here is what I'm seeing as I try to digest the numbers in the logs:

1) Sitting at Idle, closed throttle, my fuel error variance between banks seems to climb up from 0.01 to 0.13/0.14 before the Hego's seem to catch up and bring them closer together. My WBo2 readings hover within about 3-5% of LAMBSE2 from what I can objectively see.

2) Part Throttle acceleration looks really good ... KAMS are registering at 1 most of the time. It's Decel that has me wondering, particularly from mid RPM ranges. When I take my foot off the accelerator and am into DASHPOT and Decay ... The LAMBSES head in opposite directions, Bank1 veers rich, I've seen it getting into the 0.700's (a few times on those occasions where it dies coming to a stop) while Bank2 demands lean, into the 1.20's all while the WBo2 continues to hover around Stoich 0.97-1.03.

I'm going to attach the data log I captured, feel free to review and critique.

KB95_Log09_14_2025.xdl

(2.37 MiB) Downloaded 2 times

I'm trying to read, learn and process as fast as I can ... but don't want to waste time chasing what I think might be a mechanical problem when it could just be a normal behavior or nuance that folks with experience would we well aware of. I've actually spent the last few nights staring at Excel and wondering if my HEGO's could be crossed over ... which I think is pretty much impossible.

Thanks in advance for your thoughts and opinions!

Coz

[Coz_KB95]

Note: and it's visible in the log of course ... fuel variance gets as high as 30-35% in some of those decel conditions ... which is a big part of my concern.

[decipha]

i didnt check the log but if one bank goes lean and the other rich on decel then that confirms your either getting cross scavenging or the hegos are swapped. Thats troubling that your wideband reports near stoich though.

Your maf histogram looks perfect seems as though your fuel is dialed in so its odd youd be having that issue.

instead of swapping the hego connectors simply swap them in the tune.

set cyl 1 through 4 to bank 2 and set cyl 5-8 to bank 0 in the output port table to reassign the hegos to the other bank and see if that makes her more consistent.

[Coz_KB95]

Thank you! I'll give that a go and see what shakes out.

Few points to mention, since you brought up cross scavenging. I did swap out an H-pipe midsection for an x-pipe midsection ... but given the placement of the HEGOs in the longtube collectors, I wouldn't expect it to have that significant an impact. Could it in your experience.

Otherwise the only other thing I can think of is that I do run an actual J4J1 ecu( from a 95 Cobra) in this car. However from what I know the Banks were not identified differently between the GT's and the Cobras in that year.

I'll post back with results once I've had a chance to tweak the tune and log another drive ...

[Coz_KB95]

@Decipha ... just confirming. Cyl 1-4 should be set to 2 and cyl 5-8 should be set to 1, correct? Comparing with the base T4M2 ... the only values used are 1 & 2. In your post you mentioned "set cyl 5-8 to bank 0 in the output port table" however I didn't see any of the ports set to 0 when comparing to any of the base tunes. Am I missing something?

[decipha]

if you have an x pipe then theres your problem

you have to disable hego 2 for that very reason

if your not cross scavenging with an x then its a pretty crappy x and not doing its job, sounds like its working perfectly

all 94-95 gt and cobra ecus are the same

[Coz_KB95]

Got it! For some reason I thought this was more of an issue for cars with Fore and Aft sensors ... not the case here but I suppose the change in flow clearly has an effect, particularly on decel. I'm inferring that this means my fueling is likely off by some margin, since I've been applying corrections based on the averages from both banks measured independently.

I'll disable HEGO 2 and see how she likes it with the current MAF curve. Since my WBo2 is in Bank2 I can use that to keep an eye on fueling on the side where the HEGO is turned off.

Couple of additional questions that I haven't been able to find direct answers to using the search:

1) For dialing in PERLOAD (Which I have set to 0) ... which histogram has the corrections that would apply for CBAZA?

2) My guess thus far (based on some interpolation and reverse math from my logged WOT runs) is that LOAD_FG would be the ACTUAL load the ECU is registering and using to calculate PERLOAD. Is that correct?

3) I currently have the Kenne Bell bypassed, running a short belt, and dialing in fuel, load and spark with it NA before I dive into Boost. in NA form during my WOT runs Load_FG remains between 0.600 - 0.700.

From a number/validation perspective, would that be normal for an N/A - stock bottom end HCI 302? Nothing really radical about the engine, Edelbrock Performer heads, fairly tame Crane 2031 cam, ported gt40 lower intake. Of course the Kenne Bell is sitting on top of all of that, I was wondering if running that compressor bypassed is hindering the engine's ability to ingest as much air as it can. I can't recall if my vac/boost gauge was registering 0 during the runs, I was much more focused on the WB readings, but it should have been 0 at WOT unless the bypassed supercharger is restricting airflow to that extent.


Thanks again for your help and invaluable insight!