Hey all, I'm tuning accel / tip-in fuel on my NSX with a Series 1 (probably similar to the Series 2). I know, not an S2k, but the concepts are the same and the community is larger here...

I originally started messing with it because there was a slight lean condition with a quick TPS change. Generally, it'd go lean for a sec, then go rich; it felt late to correct, then when it did, it'd overcorrect. Seemed like an easy fix: increase the TPS sensitivity, pull some fuel from the tables.

The original settings that came in the car:

Accel dTPS Trigger: 10
Accel Limit: 80
Accel Pump Sustain: 35.16
Accel TP Sensitivity: 69.92

I did some experimentation with these settings + adjusted the "TPS Accel Factor" table. I was able to get the AFR close to what I wanted on a large TPS change, but still cant get it to chill out when cruising.

Accel dTPS Trigger: 10
Accel Limit: 70
Accel Pump Sustain: 24.22
Accel TP Sensitivity: 78.91

I feel like I am missing something or dont understand how all the things fit together. Here are my assumptions on a couple options & params.

Accel TP Sensitivity option

The lower the value, the more it smooths the TPS signal. So the lower the value, the slower to react to TPS changes the trim will be. The goal is to get this as high as possible without triggering it all the time.

The Throttle param is smoothed by the Accel TP Sensitivity option to get the TPS Filtered param.

This all correct?

Accel dTPS parameter

On an AEM thread (series 2 tho) it says that Accel dTPS "is the difference between Throttle and TPS Filtered". Is that right? It's diffing the noisy raw TPS signal with the smoothed signal?

Accel dTPS Trigger option

The user guide says this

> Accel dTPS Trigger: When the change in TPS (dTP) exceeds this value, Fuel Trim (Accel) can be added. This acts as a minimum TPS change threshold for Fuel Trim (Accel).

So I assume when the "Accel dTPS" param is over "Accel dTPS Trigger" option, it will add fuel. And when it's adding fuel, the Accel Fuel param will have a non-zero value.

Is this correct? If not, how does Accel dTPS Trigger relate to the Accel dTPS param?

My assumption is not what I'm seeing. Here Accel dTPS Trigger is 10, Accel dTPS is like 3 or 4, but Accel Fuel (also "Fuel Trim (Accel)") is non-zero and affecting the AFR.



In this case I want it to correct at the actual throttle change (see big arrow toward the right), but not when cruising.

The cruising dTPS vals are less than 5, and the actual change is > 5. It _seems_ like I could set the Accel dTPS Trigger to 5 and it would work.



Other questions

* Is my TPS noisier than most? It fluctuates like 3%, is that normal? I have a new one coming as an experiment.
* How smooth should the TPS Filtered param be? If it's fluctuating 1% or so, is that ok?

 

Not familiar with older ECU wall-wetting/acceleration-enrichment/tip-in settings and parameters, sorry.

What I’ve learnt is stop trying to chase perfect AFR in tip-in and go by feel. It’ll always spike lean and then go rich, just aim for that AFR up-down trend to be as smooth as you can but it’ll almost never immediately settle at target AFR on a throttle stab.

Above 4000-5000rpm I almost never use any enrichment. When tuning for throttle response I do what you did, watch the logs and I also listen for tip-in knock with my audio knock detection. If majority of the fuel table is crispy clean and on target I’ve gotten a car to run fantastic with ZERO accel enrichment yet still chose to add some just to run well when car is stone-cold obviously with a different compensation table.

The biggest thing is if it’s doughy/lazy on throttle stab, it’s too rich. If it bucks/skips and then goes, it’s too lean.

I would also filter TPS a bit more to calm the enrichment engagement if V1 doesn’t give you another way of preventing it from engaging. I know I didn’t provide any technical answers specific to your parameters but I thought I’d ramble a bit about what works for me.

 

No prob, still helpful! Yeah, I think you're right that this will require some more experimentation, and be more by feel than I originally hoped. I was cocky that I could sort it all out in one session, lol. Noted that It'll probably always go lean then rich and on the doughy / bucking vibes. Good to know I should probably filter the TPS more. Hopefully the new TPS is less noisy... fluctuating 3% seems like a ton.

I have been able to get the rest of the fuel map in a great smooth place with no weird holes. It runs great now, feels pretty stock-like in cruising areas. But I wasn't paying a whole lot of attention to accel fuel for most of the main fuel map tuning, I was just trying to make no large TPS changes and tuning off those areas. Turns out accel fuel was triggering a bit even when cruising, so it's possible the cruising bits of the main map are actually a little leaner than I thought. I should probably set the accel trigger threshold super high and do a little logging to see how it goes...

Some random thoughts I had after staring at the logs a little more: I feel like maybe the TPS voltage dips are bigger than the logs are letting on, just happening between samples, which is causing a delta TPS above 10.

That, and / or some jankness with the series 1 and the logged Accel dTPS isn't exactly what they say it is. The latched dTPS value is derived from a table: input is Accel dTPS, output is a percent fuel change. Backing out the Accel dTPS from the logged Latched Accel dTPS value, the Accel dTPS that's creating the logged Latched dTPS value is like 8-12 while cruising, which is within the trigger range, despite the logged Accel dTPS vals being in the 3-4 range.

Seems like I probably just need to play with the trigger and sensitivity by feel even if I dont 100% understand exactly WTF is going on with it. Guess and check is slow tho...

 

Yep! Good plan. I eliminate enrichment/wall-wetting and I also turn off decel-cut to avoid more false data points/hits. Enrichment gives you falsely rich and lean conditions and decel-cut shows crazy lean. Garbage in, garbage out.

Does V1 have ThrottleRate PID? That's what the Infinity uses for enrichment so I zero that out completely.

Also, if you're using histograms, consider creating these filters to clean up collected data:
1. TPS 5% and above only
2. TPS Throttle Rate no more than -50/+50 (eliminates data hits gathered with enrichment engaged (IF I didn't turn enrichment off))
3. Low/High Cam selection (isolate data to specific cam lobe)
4. Wideband feedback ON/OFF

P.S. I use MegalLog Viewer HD instead of AEM Data for log reviews and it allows for me to weed-out any polluted data and dial in AFR accuracy.

 

There's no ThrottleRate PID on the series 1. The accel dTPS is supposed to be that I guess, but it's all over the place! I'm sure all this stuff is a lot better in the infinity. I can probably set the trigger to 100 and get the same result. I'll mess with the trigger and sensitivity this weekend and report back.

Noted on MegaLog viewer. I'm only using AEMLog to see what the Accel fuel is doing, or generally look at trends over time (e.g. how much does the IAT spike during a pull?). I actually wrote my own little log analyzer to help tuning the main fuel map: https://github.com/benogle/fuelie (it's open source, has downloads for windows & mac). This stuff is fun and I wanted total control to experiment . I have it setup to ignore samples with accel fuel over a certain percentage of inj PW. Though the AFR response is like 100-300ms behind the trim, so I've been meaning to build in a way to filter out subsequent samples.

Other than that, O2 feedback is def off. I haven't touched it in high cam other than logging. The tune was in good shape aside from < 5k cruising, it's been on the dyno at a reputable NSX tuner a few times with the previous owner. Right now, this is operation: cruise like stock.

 

Neat! Software dev is way over my head.

 

Ok, I have a better understanding of these options now after some experimentation and much staring at the logs.

Accel dTPS Trigger is the RAW value (0 - 255) of the Accel dTPS. Formula is `raw / 2.55 = real%`. I set mine to 14. 14 / 2.55 = 5.5% delta TPS

The X axis in the Accel dTPS table is 1/2 raw for some reason. So `xAxisVal / 1.28 = real%`.

I was getting 3 - 4% Accel dTPS fluctuations no matter what sensitivity I picked, so I figured I'd go for a high one and limit cruising trim with the trigger.

Higher sensitivity will do more smoothing. Here's 90



And at sensitivity 70, the filtered TPS has more noise:



I haven't had a chance to mess get cruising logs without accel trim yet, but I will do soon.