Honcho's Long Road to Imola Type-S Zero

[Honcho]

I am so happy for you to finally, finally get to the point where the car is drivable. A couple more hurdles to clear but you are almost there. Congratulations.

Glad to hear the engine feels good..Funny how the things you don't pre-worry about give you the hurdles..

Honcho car looks stunning out in the wild! Imola is such a nice color for the NSX! Amazing work!
Congratulations!

That’s lovely!

That's how it's supposed to be, at least with OEM clutch.

I can relate to the feeling

I can't even imagine the feeling you must have right now. Awesome!

Awesome that it's up and running! Cant wait to see some driving vids.

Your car looks awesome! Looks like you are in for a fun driving season this year! It has been along haul watching this process, I can't imagine the effort and skill it took to actually make it happen. It has been a privilege to be able to watch this from start to finish.

Thanks guys. It's such a relief and I was floating all weekend!

 

[Mac Attack]

Congratulations to you, Mrs. Honcho, and your team! You guys should be proud of yourselves! Thanks for all the time spent documenting your work.

 

[ekhonda87]

Wow, such an inspiring build! It must feel amazing to finally be in the home stretch.

 

[jazzmann]

I'm so happy to see this, Honcho, and so happy for you. What an EPIC project.

 

[docjohn]

Hope to see the Imola S in Georgia!

 

[Honcho]

Congratulations to you, Mrs. Honcho, and your team! You guys should be proud of yourselves! Thanks for all the time spent documenting your work.

Thanks- it has been a long road indeed. But on the upside I basically have a brand new gen1 NSX in 2023...the only other two people I think who can claim something similar are @cmc140 and @serialNSXer. That's pretty nice company!

I'm so happy to see this, Honcho, and so happy for you. What an EPIC project.

Epic is definitely the word. How's my baby?

Wow, such an inspiring build! It must feel amazing to finally be in the home stretch.

It does. Most of the next couple months will all be driving-related tweaking, which is so much fun!

Hope to see the Imola S in Georgia!

That's the plan....

 

[Honcho]

Re-Assembly 110

Decals, etc.

It took months, but I was finally able to create the proper color labels. I'm using weatherproof polyurethane decal paper with a laser printer. The decals now match the interior and exterior colors. Hopefully it will last.

 

[Honcho]

Re-Assembly 111

Security system, front lip, radiator duct, etc.

It occurred to me that I never tested the security system. Now that all of the panels and hatches are back on the car, I thought I'd give it a try. Success! One thing I noticed was that the keyless transmitter does not lock/unlock. It came with the car and is very old, so I assume it is the correct fob for the SCU. Does anyone know the method to get it to "relearn"? Or maybe the battery is too weak?

View attachment 20230202_175820.mp4

Got the restored original 91-96 front lip installed using the stainless bolt kit from @MITA Motorsports. Very nice.



Also managed to get the rest of the snap clips and body bolts installed under the radiator cowl and at the two bumper reinforcement struts. The work under the car is complete.

 

[Honcho]

Test Drive 2
Emissions, fuel, etc.

Today the temperature rose to almost 50F, so it was a good opportunity to go to the "land inspection office" (in T3TEC parlance) for emissions testing and official licensing. This would also give me the opportunity to run the engine and transmission at road speed. It was a total trip of about 14 miles, with bright sun and dry roads (with wet spots from melting snow). Feedback:

• The steering alignment is incredibly nervous. Even the slightest bump makes the car feel like it wants to veer sideways off the road. I was struggling and fighting the whole time to keep it straight, coasting through turns, being gentle with throttle and brake input, etc. The car needs an alignment badly.
• The transmission is flawless. Smooth shifting, no whines or grinds. Clutch feel is like butter, even with the NSX-R damper delete and lightweight flywheel.
• The light flywheel still has enough inertia to get the car going on an incline comfortably. I got to test this about 100 times waiting in the 45-minute line at the inspection facility. Highly recommend the Jun flywheel.
• The engine is strong. Like really, really strong. It feels like a different NSX under there. Even light throttle input resulted in a lively and torquey pull.
• I'm going to have to do something about that exhaust. From 2000-2500 (the danger zone for all NSX drone, regardless of exhaust), it's unbearable. I'm certain the mufflers are blown out.

The inspection facility was an adventure. I spent most of the time nervously monitoring the temp gauge as the car sat there and heat soaked for 45 minutes while waiting in line. Thankfully, no issues. The guy behind me was in a lime green Viper ACR with the big wing and everything. We spent a good bit of time chatting about the cars. The inspection techs spent more time taking pictures of the NSX than testing it lol. When they hit the throttle on the rollers, the Viper owner was like "Oh man listen to that! I need a NSX!". LOL

I am running on Mark's RDX tune, which is a mashup of John's edits, the Prospeed tune and Mark's adjustments based on his logs. It's not that great and I am going to start from scratch with a new tune next. However, Mark's NSX passed smog so I was hoping his tune on my car would pass too. It did, but barely.

Here is the emissions report from the car when it was the automatic on the stock tune/injectors. Very clean, even for 137,000 miles!


Here is today's report. Note the HC, CO and NOx values. WAY higher. The tune is pig rich and the car is pulling fuel all over the map. It looks like the car can't pull enough fuel to hit 14.7 (the limit is -30%). In addition to not being good for the cats, it's putting me too close to the limit for comfort. I have a lot of work to do on the closed loop fuel cells.




After the inspection, I drove to the gas station and filled up the tank.



Now I can finish the little interior pieces and get that Type-S badge installed. Then, it's off to the alignment shop next week. Even in its nervous condition, it felt so nice to finally be driving.

 

[Big McLargeHuge]

The steering alignment is incredibly nervous. Even the slightest bump makes the car feel like it wants to veer sideways off the road. I was struggling and fighting the whole time to keep it straight, coasting through turns, being gentle with throttle and brake input, etc. The car needs an alignment badly.

That should be the considerable toe-out on both front wheels . Taking a WAG, I'd wager close to 1.0 degrees? Would be interesting to see the initial alignment printout.

The weather's looking up for Feb so hope you can get out and do some more test & tuning driving.

 

[MotorMouth93]

Yeah the only real changes I made to that tune were the closed/open loop transition parameters. I left the jacked up fuel maps alone...I didn't really learn how to dial in fuel maps until I installed the ITBs and had to start from scratch and didn't have O2 feedback to save me.

 

[ekhonda87]

Do they salt the roads in Colorado? Any concerns about salt? I am itching to drive my car, but the roads are heavily salted here.

 

[docjohn]

The Bee has seen salt for 26 winters and the undercarriage looks pretty good imho..

 

[Honcho]

Do they salt the roads in Colorado? Any concerns about salt? I am itching to drive my car, but the roads are heavily salted here.

No, they use magnesium chloride, which does not corrode iron. It is a bit caustic to paint, so I always make sure to rinse the car exterior periodically.

That should be the considerable toe-out on both front wheels . Taking a WAG, I'd wager close to 1.0 degrees? Would be interesting to see the initial alignment printout.

The weather's looking up for Feb so hope you can get out and do some more test & tuning driving.

It's really, really nervous and squirrely. I wouldn't doubt it's 1.0 or more.

Yeah the only real changes I made to that tune were the closed/open loop transition parameters. I left the jacked up fuel maps alone...I didn't really learn how to dial in fuel maps until I installed the ITBs and had to start from scratch and didn't have O2 feedback to save me.

There is a layer of black soot on the rear bumper- much more than normal. It's clearly running rich for sure.

 

[Superhatch]

If it's that rich but also feels really good/strong hopefully dialing in the tune will make it even better! Good news generally though so far. Are you bringing it to Atlanta?

 

[MaxGeek]

Good to hear the JUN flywheel feels good. I ordered one as well, but won't be using it until I need a clutch change.

 

[Honcho]

Test Drive 3

After the long drive to the inspection station, I checked the coolant level after cooldown. It looks like a little more air was purged. Topped off to the max line and I will continue to monitor.



I plan a small test drive today to trial a revised ECU tune that fixes (hopefully) many of the bad problems in the Prospeed original tune, as well as testing a history table function that could potentially greatly simplify the tuning process.

 

[Honcho]

Alignment 1

The S-Zero is at the alignment shop. This is the same place that was willing to take on the crazy project of pressing out my suspension bushings. They have real mechanics who know what they are doing and a very clean facility, so I decided to trust them with the NSX itself. They mostly work on lifted 4WD "bro-dozers" and jeeps (it's a big thing in Colorado), but they have skill, which is what I'm after. I can teach them what they need to know about the NSX-specific stuff.



They have a commercial grade pressure bleeder, so they will run another liter of the ATE brake fluid through the system. My efforts gave a reasonable brake pedal feel, but a pressure bleeder will get ALL of the air out. They will also fill the A/C system with refrigerant and test for leaks.

The drive down to the shop was terrifying due to the alignment. The front toe settings are downright scary and at 50 mph, the car feels like it is wiggling with the slightest bump and could just jump sideways at any time. Thankfully I made it safely and they will have the car for a few days.

Upon returning home, I noticed this on the garage floor:



Sigh.

I'm pretty sure I know where it's coming from. When I drained the engine block to re-do the coolant, I couldn't get much torque on the rear engine drain plug due to the awkward access. It's as tight as my index finger and thumb can make it, but probably not at the 22 lb/ft spec. I'll take a look when the car is in the air at the shop.

 

[Superhatch]

That's not too bad of a drip to deal with. If the shop is going to have the car up for alignment maybe they could tighten it?

Always nice to find a good shop that does good work. I've been taking all my D/B series builds to an old school muscle head shop for a while and I remember when I brought them my first D series block and head and they asked what hp I was going to make with it, they were all shocked when I said 400.

Since then they've actually taken on a lot of work they wouldn't have and developed a way to drill out the oem valve guides and press in a bronze insert as they found pulling out OEM guides and putting in good aftermarket ones (supertech, etc) could lead to some play.

Also, that paint is amazing.

 

[The Kid]

Definitely one of my top 5 NA1/2's in existence. My only issue I always had with my Zanardi was the lack of color options, shame we never got lime or orange and Type S trims here. Your interior is simply unreal and the whole build is nothing short of amazing. I have the same engine cover and used the no drill kit too. Just curious, what did you do for audio? I must have missed that portion.

Did you PPF the car? I'm hoping to see it in Georgia.

 

[Honcho]

Test Drive 4

ECU tune, etc.

The drive down to the alignment shop provided a great opportunity to log the most recent changes I made to the ECU bin file. After many long discussions with @MotorMouth93, we determined that the "original" tune was not very good. It was a mashup of the original Prospeed RDX tune, some of John's changes to get Power Enrichment to engage properly, and @mskrotzki's changes to some of the VTEC fuel cells. Crazy enough, this is the tune that still (barely) passed emissions. After looking at the bin file, it was clear the car was in open loop for most of the emissions test (!)

To rectify the problem, I changed the throttle position back to OEM for open loop (PE), but kept the minimum pulsewidth at the "RDX" adjusted number.




Next, I scrapped the dodgy Prospeed base map 1 for an OEM 1994 fuel map. I then simply re-scaled that OEM map to 60%, which is roughly the difference between the OEM 240cc and RDX 410cc injectors, or 40% leaner, as shown below.



Finally, I set the VVIS and VTEC limits back to normal from the messed up Prospeed values, and set the rev limit back to 8,000 from 8,300. Apparently Prospeed didn't care about oil pump gears...




These changes worked very well, as the car stayed in closed loop for the whole drive and ran much more smoothly. Although it might be too good, since I could not get the car to go into PE even with proper throttle and PW. Have to do some more work there.



Also, I was able to log STFT in a history table that matches the base fuel map. While the history table contains all of the "noise" from transient throttle conditions (i.e. press gas, lift-off, etc.), the average STFT values do provide a rough guidepost as to where the base map is rich or lean on a cell-by cell basis. Basically, as shown below, the low-rpm and idle regions of the map are a little too rich and the rest of the re-scaled OEM map is a little too lean. This makes sense since, as John points out, fuel injectors do not have a linear flow rate. I can use this STFT logging method to get the map "close" and keep the safety margin of the OEM closed loop system to protect the engine. Once it's within safe limits, I can shut off closed loop and use the wideband to dial in the rest of it with John's python AFR scripts. My goal is to have a very smooth and powerful closed loop range before I start messing with PE and VTEC maps and the Comptech cam lobes. In truth, those areas of the fuel maps will have to be tuned on a dyno to gain the most torque and safety margin. Doing pulls on the frontage road isn't the best way to really do it...

 

[Honcho]

Definitely one of my top 5 NA1/2's in existence. My only issue I always had with my Zanardi was the lack of color options, shame we never got lime or orange and Type S trims here. Your interior is simply unreal and the whole build is nothing short of amazing. I have the same engine cover and used the no drill kit too. Just curious, what did you do for audio? I must have missed that portion.

Did you PPF the car? I'm hoping to see it in Georgia.

Thanks! I deleted audio, since it's a S-Zero spec. I do have a bluetooth speaker that I can use for long trips though, and an integrated USB charger in the center box to keep it charged lol. The car will be going in for full PPF, but first it needs a full buff-out, detail and touch up paint for all the nicks and bare fender bolts. Next stop after alignment is the body shop for those items, and then PPF.

That's not too bad of a drip to deal with. If the shop is going to have the car up for alignment maybe they could tighten it?

Always nice to find a good shop that does good work. I've been taking all my D/B series builds to an old school muscle head shop for a while and I remember when I brought them my first D series block and head and they asked what hp I was going to make with it, they were all shocked when I said 400.

Since then they've actually taken on a lot of work they wouldn't have and developed a way to drill out the oem valve guides and press in a bronze insert as they found pulling out OEM guides and putting in good aftermarket ones (supertech, etc) could lead to some play.

Also, that paint is amazing.

I agree- there are still good mechanics out there and their skill level is more valuable than any perceived "Honda" expertise. I still am mystified when a NSX owner tells me they only have the car serviced at Acura. None of the techs there have even seen a Gen1 NSX at this point, so they just follow the FSM step by step- like any other basic mechanic would. I'd rather have a guy who has been wrenching for 20 years and knows all the tricks of the trade. Sometimes what we see on the car doesn't exactly match what the FSM says and I would trust a skilled person more when this inevitably happens.

The most expensive part of my head work was the ST guides. It took many heating cycles to safely remove the old guides without galling and then install the new guides.

 

[Old Guy]

Your printed paint code stickers are a nice OCD detail.

Your comments about non linearity in the injectors surprises me. I was using a very old Bosch series injector on another car. Because the internet data on the injector was all over the place (including sources that claimed to be based upon Bosch catalog info), I actually did my own flow tests, measuring the fuel delivery for a fixed set of pulses from pulse widths starting at 0.25 msec right up to 10 msec at 0.5 msec intervals. I wanted to confirm the flow rate, determine the injector offset and offset voltage sensitivity and determine the linear region. To get an accurate assessment of where the injectors went linear I did 0.25 msec spacing for the PW between 0.25 and 1.5 msec. When you plot this data and take the slope the slope turns out to be the injector flow rate.

These injectors turned out to be 551 cc/min @ 43.5 psi. As expected, the injectors were highly non linear at very low PW values, actually flowing nothing at 0.25 msec PW. However, once above 1.0 msec the injector flow versus PW was an absolutely straight line out to 10 msec PW. I did a validity check by opening the injector for 30 seconds for a bulk flow measurement, measuring a delivery of approximately 277 cc which indicated that the injector appeared to be linear from 1 msec to 30,000 msec (approximately because I didn't have a graduated cylinder large enough for 277 cc so I had to use a measuring cup).

The normal practise is to size the injectors so that the idle pulse widths are just out of the low pulse width non linear region. If you have a flow versus PW curve for your injectors and you find that your idle PW are in the non linear region of operation you are kind of in terra incognito. As I recall you are using OEM injectors and I would be surprised if Honda specified injectors that were operating anywhere close to the low flow non linear region. That operating uncertainty is not good for passing emission compliance tests. The other variable is the injector offset value (where the injector slope line crosses the X axis). Some people refer to this as the opening time; but, it isn't. Errors in the injector offset value can cause the flow delivery to skew rich or skew lean depending on the sign of the error. At low pulse width values this error can be a problem because the offset may be significant compared to the injector PW. If the injector is running at high PW, errors in the offset are less significant. My offsets are around 0.21 msec at 12 volts. If you are idling with a PW of 1.5 msec, the offset is around 15% of the PW value and errors can be material to your AFR. If the PW is up around 9 msec my offset is about 2.3% and errors are less material. If you have unintentionally biased your fuel mix to correct an offset error at low PW values, depending on how you determined the bias you may have biased the AFR in the other direction at high PW values. The added complication is that injector offset is very voltage sensitive. My injector offset of + 0.21 msec at 12 volts actually slides across the axis and becomes -0.08 msec if the injectors are operating at 16 volts. Most ECUs have a compensation curve that corrects the injector offset for measured voltage to allow good idle performance.

I remember that when the Prospeed RDX injector kit came out, there were people complaining about poor idle performance. My initial reaction was that you put a larger injector on the same engine, chances are you have pushed that injector into its non linear low flow region which is going to result in unstable operation. I recall that there were some software tweaks to try and fix this and thinking to myself that they may have fiddled the offsets to try and make the injector better behaving at low PW. Since you said you had started with a Prospeed tune, you might want to check to see what is in there for injector offset values and voltage sensitivities. Unless somebody else has done this work, it might be useful to set up a small flow bench to confirm the linear region of the injectors and actually measure the injector offsets and voltage sensitivity. That would actually seem to fit right in with your OCD attention to detail .

 

[Honcho]

Your printed paint code stickers are a nice OCD detail.

Your comments about non linearity in the injectors surprises me. I was using a very old Bosch series injector on another car. Because the internet data on the injector was all over the place (including sources that claimed to be based upon Bosch catalog info), I actually did my own flow tests, measuring the fuel delivery for a fixed set of pulses from pulse widths starting at 0.25 msec right up to 10 msec at 0.5 msec intervals. I wanted to confirm the flow rate, determine the injector offset and offset voltage sensitivity and determine the linear region. To get an accurate assessment of where the injectors went linear I did 0.25 msec spacing for the PW between 0.25 and 1.5 msec. When you plot this data and take the slope the slope turns out to be the injector flow rate.

These injectors turned out to be 551 cc/min @ 43.5 psi. As expected, the injectors were highly non linear at very low PW values, actually flowing nothing at 0.25 msec PW. However, once above 1.0 msec the injector flow versus PW was an absolutely straight line out to 10 msec PW. I did a validity check by opening the injector for 30 seconds for a bulk flow measurement, measuring a delivery of approximately 277 cc which indicated that the injector appeared to be linear from 1 msec to 30,000 msec (approximately because I didn't have a graduated cylinder large enough for 277 cc so I had to use a measuring cup).

The normal practise is to size the injectors so that the idle pulse widths are just out of the low pulse width non linear region. If you have a flow versus PW curve for your injectors and you find that your idle PW are in the non linear region of operation you are kind of in terra incognito. As I recall you are using OEM injectors and I would be surprised if Honda specified injectors that were operating anywhere close to the low flow non linear region. That operating uncertainty is not good for passing emission compliance tests. The other variable is the injector offset value (where the injector slope line crosses the X axis). Some people refer to this as the opening time; but, it isn't. Errors in the injector offset value can cause the flow delivery to skew rich or skew lean depending on the sign of the error. At low pulse width values this error can be a problem because the offset may be significant compared to the injector PW. If the injector is running at high PW, errors in the offset are less significant. My offsets are around 0.21 msec at 12 volts. If you are idling with a PW of 1.5 msec, the offset is around 15% of the PW value and errors can be material to your AFR. If the PW is up around 9 msec my offset is about 2.3% and errors are less material. If you have unintentionally biased your fuel mix to correct an offset error at low PW values, depending on how you determined the bias you may have biased the AFR in the other direction at high PW values. The added complication is that injector offset is very voltage sensitive. My injector offset of + 0.21 msec at 12 volts actually slides across the axis and becomes -0.08 msec if the injectors are operating at 16 volts. Most ECUs have a compensation curve that corrects the injector offset for measured voltage to allow good idle performance.

I remember that when the Prospeed RDX injector kit came out, there were people complaining about poor idle performance. My initial reaction was that you put a larger injector on the same engine, chances are you have pushed that injector into its non linear low flow region which is going to result in unstable operation. I recall that there were some software tweaks to try and fix this and thinking to myself that they may have fiddled the offsets to try and make the injector better behaving at low PW. Since you said you had started with a Prospeed tune, you might want to check to see what is in there for injector offset values and voltage sensitivities. Unless somebody else has done this work, it might be useful to set up a small flow bench to confirm the linear region of the injectors and actually measure the injector offsets and voltage sensitivity. That would actually seem to fit right in with your OCD attention to detail .

@MotorMouth93 has a much better understanding of the NSX fueling cycle then me, but I know a little bit. The base fuel tables are just that- base values. When the ECU sees a certain load/rpm, it will select the base fuel pw and then apply all of the modifiers on top of it, including injector latency. The "final pw" that I log is the final commanded pulsewidth for the next injection. This value is useful for figuring out where/whether the car exits closed loop, since it uses minimum final pw (along with throttle position) as one of the key thresholds, as noted above. It's a bit of a mystery here, because the car should be going into open loop, but it's not.

The original Prospeed RDX mod had idle issues because of the "injector latency", as it is called in the NSX ECU, for exactly the reasons you mention. By adjusting those values (specifically the 12V and 14V cells) to ones measured on the actual RDX injector, we were able to solve those problems. The car will idle smoothly all the way down to the 650 rpm base idle setting, so even the 410cc RDX injector appears to remain above the non-linear threshold at very low pw. The other big problem with the Prospeed kit was that the injector seating was wrong, which caused the injector spray pattern to be aimed poorly. Brad's fitment kit fixed that.

What is interesting/fascinating to me is that simply shifting the curve down 40% does not preserve proper fueling, notwithstanding the linearity of the injectors. You can see in the STFT chart that the ECU is pulling or adding more than the standard-ish 3% +/- at many points on the base map. Part of this is likely because the car is using different modifiers in these cells at different times, but because they are running averages of sometimes 500 or more samples (I only adjusted where n>30), it's a pretty good indicator of how the NSX feels about its fueling at each load/RPM cell. It's quite variable. During this phase of the tuning, I'm shooting for about 3% too rich as my closed loop baseline. I'd like to see - 2 to -3% in every closed loop cell in a perfect world.

 

[Old Guy]

The algorithm of reading a base PW, applying modifiers based upon the engine's operating conditions and then add on the latency value is a pretty standard fuel control algorithm even in aftermarket ECUs. I expect that latency is just the term that is used for the offset. If you have flow tested your injectors and are confident that you are out of the low flow non linear region and have correct latency / offset value then my comments about offset errors and non linearity contributing to your AFR error bias are probably not relevant.

Using fuel PW to determine closed versus open loop is interesting, but, not what I am familiar with which is typically using RPM and MAP to establish the transition threshold. Working backwards, PW is probably a reasonable proxy for engine load. On aftermarket ECUs you can typically set an AFR target in closed loop for each fuel cell so forcing the ECU out of closed loop is less of an issue (if you trust your wideband sensor at high loads). Since you are using the OEM ECU with narrow bands I am guessing that you are stuck at 14.7 until you can force it to exit closed loop? Not being familiar with the ECU code, I can't really comment on what might be going on other than 'interesting'.

Your comments on short term trim are interesting. Based upon what my OBD II ECU reports through a scan tool I had assumed that the short term trim as reported by the ECU was a global correction applied on a per cylinder bank basis, not on a per fuel cell basis. I think that was pretty standard for OEM ECUs. You are showing short term trim values for individual fuel cells. Is that something that your logging software is calculating based upon the instantaneous fuel correction from the O2 sensor or does the OEM ECU actually calculate and store short term and long term trims on a per cell basis? The global long term trim is calculated from short term trim values and is applied to try and make short term trim 0. The long term trim value is important in terms of assessing emission compliance because it is an indicator that something has gone bad and triggers a CEL when it passes a threshold. It is not a particularly helpful performance feature. If you still have long term trim active and it is being calculated and applied on a global basis, then it can generate a permanent skew for individual cells. If the global fuel trim is being calculated on a time spent basis which is what is implied in the service manual, it will become biased toward correcting those cells where you spend most of your time. This correction may not be correct for all cells which will result in high fuel corrections for those cells. If you have disabled the OEM global short term and long term fuel trim adders, or the ECU is actually calculating short term and long term on a per cell basis then my comments are probably not relevant. As a matter of interest, how are you dealing with the front bank versus back bank O2 error signals?

Final note, how does the ECU manage the base pulse width values from the fuel table? If you have a cell at some MAP and RPM value and an adjacent cell, as you transition from the first cell to the next cell, is the algorithm interpolating between the stored base PW values to calculate transition PW values or does it just change from the first stored value to the next stored value as it crosses the cell boundaries? For those areas of the fuel map where there are transitions, if there is no interpolation then it is probably unrealistic to expect that the you will ever get a fuel trim to small numbers for that cell. Even if there is software interpolation, if the fuel map is in a sharp transition area it may be impossible to get the trims low for those transition cells. However, that would not contribute to that rich at low load, lean at high load bias that your log data shows.