When you click on links to various merchants on this site and make a purchase, this can result in this site earning a commission. Affiliate programs and affiliations include, but are not limited to, the eBay Partner Network.
Curly has literally been on these forums fixing people's tunes for years. He's helped hundreds of people in threads like this when their car won't start or is running like ****, and on different types of ECU's. He's probably one of the most reputable people on here when it comes to tuning. There are plenty of people who don't know what they're talking about (probably myself included), but I wouldn't put Curly in that camp.
^^^
This!
I was basically saying the same thing to another member here. I think my exact words were: "Curly's advice is literally worth it's weight in gold."
I will get to this after work today or sat morning, and answer the questions as well. Long day at work yesterday (14 hrs) and it's likely gonna be the same today.
@curly If there are any differences in our changes I am happy to explain my reasoning, i am solid on this been looking at this type of data daily for over a year straight....
Out of curiosity, does your tool (or megasquirt for that matter) factor in a lag between the AFR reading getting to the ECU and EGO responding to it? I know the G4x has a table that allows you to adjust that delay/update rate based on RPM so you can factor in how much/fast the exhaust gasses are flowing. Similarly, does it factor in things like MAT corrections or does it assume that those are all set up properly? Generally the advice I see is to dial in the entire VE table with stable conditions and corrections off and then dial in corrections (IE MAT) as conditions change, but I realize that isn't as applicable in a back and forth like you've got here.
The tool assumes corrections are set up reasonably -> it does not back out MAT/CLT/Baro/AE to give you a "naked" VE correction. It corrects the VE table that the engine is actually running with all the existing corrections active. That's a deliberate choice for forum tuning.
The way I deal with it is filtering, not isolation:
Cold CLT filter (drops data below threshold)
AE filter and post-AE cooldown (drops accel enrich frames and the recovery window)
Fuel cut filter (6 methods)
Battery window filter
Wideband delay alignment
Off-throttle non-cruise gate (v15.8.1: pedal<1 AND DC<5 AND RPM>1200, MAP-agnostic - kills overrun/coast garbage)
Neutral-AFR gate
AFR ROC filter
RPMdot/MAPdot transient filters
Warmup enrichment filter (
Code:
Fuel: Warmup cor > 105%
)
Wideband sanity bounds (AFR < 9.5 or > 22.5 dropped as sensor fault)
Cold-wideband sensor warmup detector (v15.16, drops the rail-pegged window at log start)
Also, I'd be curious to hear what your tuning experience is. Mostly grassroots tuning your own car or have you done full tunes (IE not just ve tables) on other people's cars before?
Mainly Grass Roots, and I only do full tunes on families cars
"Just because this dude says he has been doing the Dyno tunes for 10 years does not mean he is any good" Sorry, this one was too good not to acknowledge, might have to throw that in my signature. Curly has literally been on these forums fixing people's tunes for years. He's helped hundreds of people in threads like this when their car won't start or is running like ****, and on different types of ECU's. He's probably one of the most reputable people on here when it comes to tuning. There are plenty of people who don't know what they're talking about (probably myself included), but I wouldn't put Curly in that camp.
Well I just felt like he was taking a shot at me and I kinda did that same...
The point of the filter chain is to only accept data when corrections are stable, not when they're zero. If MAT cor is doing its job within its normal range, the EGO correction in a stable cell reflects VE error, not MAT error. If MAT cor is misconfigured wrong table, wrong sensor scaling, way off the VE table will absolutely get corrected to compensate for it, and that's a real limitation worth being honest about. And I am about to get to work on factoring exactly this into the tool. Thanks!
I mainly asked about corrections because it's usually something that comes into play when dialing in a rough VE table. I've also seen some base tunes (I want to say some of the speeduino ones) that have really aggressive corrections that could throw some things off.
Also, on the dyno tune corrections, did the tool just add the AFR difference to the cell? IE 0.7 points lean so it added 0.7 to the box? I glanced at the tune/log last week and guessed the WOT values would mostly be in the 100's. I would've expected the tool to be doing much larger adjustments.
Having said that, I'm pretty sure I used that trick on Speeduino with good success, but it didn't work on the G4x as well, presumably because of their fuel model. I'm going to guess that your table up their is still pretty lean in the WOT cells, but I also would expect your tool to want more data to crunch than that log provided.
Yeah that log was lacking, The data was thin the car was not fully warm although that could have been due to the huge fans that Dynos use, so I mainly did that log manual. My Tool is EGO primary built for street tuning, this is the first time i had a log under a controlled session...
Edit: To be honest, I wanted to throw some higher corrections, but was worried I'd get toasted by Curly in a way since the data was so thin on this car, I just assumed this was a test, and just kinda went lighter on the corrections, just didn't wanna catch any flack. I am genuinely trying to develop this tool and if a mod who's word....
This!
I was basically saying the same thing to another member here. I think my exact words were: "Curly's advice is literally worth it's weight in gold."
This made me go conservative cause I am just some random mfer on the internet new here...
@SimBa Could you possibly grab me a data log from your car? Obviously you do not have to apply the corrections that my tool comes up with, I'd like to have your thoughts on what it does, tho Perrhaps Give Ego authority over your entire table, You seem very sharp on this subject and getting your feed back on the corrections that my tool comes up with would be highly valuable, just if you get the time take a ride and datalog if you could that would be great...
Edit: I do not know everything I am here to learn, just as much as I am to help
I talk a lot, but I'm also far too self aware to think that I'm above average when it comes to tuning knowledge (or that my tune is spot on).
I can give you a log, but I'm assuming your tool doesn't work with Link tunes/logs? CLL (Closed loop lambda is Link's EGO) has authority over my whole table, but the amount it can add/remove depends on RPM and MAP. At higher boost levels it can only add fuel, most other places it can add/remove 15%. Also, as I alluded to yesterday, I'm using their modeled multi-fuel mode, so an addition/subtraction to the VE table doesn't align very well with the same values on a MS map in my experience.
I can give you a log, but I'm assuming your tool doesn't work with Link tunes/logs?
No but I'd like a crack at it anyhow, I was thinking the math behind the tool is pretty solid so in theory I could transfer over to any tuning program it would be be a matter of re mapping the column names etc...
Honestly I might have to go collect a new log with some different settings. All the ECU log on my laptop are way too big to upload (20 MB limit) and when I convert them to CSV files they get bigger somehow.
this is all very interesting, i assume having speeduino is no problem as im using tuner studio? my tune is dog ****, but going turbo soon so was planning to put the effort in when i created that base tune.
Speeduino runs through TunerStudio so the UI is similar but the .msq file format is a different signature than MS3. Drop me an .msq and a CSV log from your current setup and I'll check whether my parser handles it. The CSV column resolver already reads MS variants, I will just extend the lookup table with Speeduino's column names. Send the files and I'll figure it out. Drive a good bit at least 10 minutes of data....
Edit: A lot of things are similar...
Code:
MS3 Speeduino
VE table: veTable1 veTable (16x16, both)
RPM axis: rpmBins rpmBins (same!)
Load axis: mapBins fuelLoadBins (already an MS variant)
CSV columns:
RPM, MAP, TPS, CLT, IAT: all same
AFR raw: may be 'O2' in Speeduino
VE1, egoCorrection: same
0
Sorry, this one was too good not to acknowledge, might have to throw that in my signature. Curly has literally been on these forums fixing people's tunes for years. He's helped hundreds of people in threads like this when their car won't start or is running like ****, and on different types of ECU's. He's probably one of the most reputable people on here when it comes to tuning. There are plenty of people who don't know what they're talking about (probably myself included), but I wouldn't put Curly in that camp.
