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.
So.. after some rounds logging and tunning, the car feels like is is running better, but my VE table looks odd. I am not sure if this is normal or not: Should I try to pull the peaks down?
Spikes are normal. The engine wants what it wants. Sorry if you already know this but to tune the VE table, go for a drive and log VE- AFR error in VCM scanner (you'll need a wideband). After you've driven at whatever combo of rpm and load is required to populate the cells, highlight, right click, and copy them. Then in VCM editor highlight the same cells in the VE table and right click, select "paste special" and select "multiply by %". Now you've adjusted those cells to give the engine what it wants. Procedure is the same for MAF table. Note that you should disable the MAF if you're tuning VE, and you should run MAF only if tuning the MAF. In both cases you'll want to disable closed loop so the O2 sensors aren't making corrections.
Spike are from bad data in your log. If you hit enough cells and have a high cell count with filters to help weed out bad data your VE will end up a lot smoother. The other option is to hand smooth it.
You can set the amount of times a data point has to be hit in order for the cell to populate. I can't remember specifically but I think default is 5. You might as well double that in your settings. If you hold your cruise in a manner so one cell or small cluster of cells can learn, eventually you'll see their values start to change by only a few decimals as opposed to whole numbers. That's how my VE table got its shape. When I first started using hpt I was real ocd about making everything perfectly smooth and pretty but then I learned it's not necessary.
Spikes are normal. The engine wants what it wants. Sorry if you already know this but to tune the VE table, go for a drive and log VE- AFR error in VCM scanner (you'll need a wideband). After you've driven at whatever combo of rpm and load is required to populate the cells, highlight, right click, and copy them. Then in VCM editor highlight the same cells in the VE table and right click, select "paste special" and select "multiply by %". Now you've adjusted those cells to give the engine what it wants. Procedure is the same for MAF table. Note that you should disable the MAF if you're tuning VE, and you should run MAF only if tuning the MAF. In both cases you'll want to disable closed loop so the O2 sensors aren't making corrections.
Here's my VE table
Um spikes aren’t normal. The engine will want smooth transitions cell to cell. Set up filters better or learn how to smooth out areas of the ve table. I clean this issue up a lot. To properly speed density tune takes a bit of patience and steady state throttle input. Hope this helps
Um spikes aren’t normal. The engine will want smooth transitions cell to cell. Set up filters better or learn how to smooth out areas of the ve table. I clean this issue up a lot. To properly speed density tune takes a bit of patience and steady state throttle input. Hope this helps
I mean... this has been discussed to death over the last 20 years. If you look up threads on this topic you'll find people generally agree that a super smooth VE table is good for aesthetics but that's it. True you dont want wild peaks and valleys but if your fuel trims stay within ~2% and the car feels and drives fine why bother spending the time to smooth it?
Data logs are always going to be noisy.
If you gather enough data, the noise averages out and you get smoother data.
If you don't get enough data, you can see the noise in the data.
If you use noisy data to edit your tables, the shape of the table reflects the noise in the data.
Every single Terminator fuel table I've seen from some of the tuner videos (Joe Simpson, Devin Vanderhoof,) all talk about having smooth tables not "jagged peaks", I think it can be just a professional quirk thing, like they want them to look nice and neat, just as they want the car to perform perfectly as well.
You can set the amount of times a data point has to be hit in order for the cell to populate. I can't remember specifically but I think default is 5. You might as well double that in your settings. If you hold your cruise in a manner so one cell or small cluster of cells can learn, eventually you'll see their values start to change by only a few decimals as opposed to whole numbers. That's how my VE table got its shape. When I first started using hpt I was real ocd about making everything perfectly smooth and pretty but then I learned it's not necessary.
The spots I pointed to are just from bad data, like NSFW said. The first arrow on the left you see 91,77,81 @45kPa which is impossible for the engine to want 91%VE at 2400rpm then at 2800rpm want 77% then want 81% at 3600rpm all at 45kPa. Its a alot easier on a dyno with load cell to get smooth data but it can be on the street with smoother throttle, filters and a higher cell count. Just by adding couple basic filters like 0 tps, below coolant temp 160, paste and multiply by 50%, and minimum cell count 100 it would automatically start smoothing out your ve table. Those filters where examples just to give you an idea. A picture perfect VE where someone clicks the smooth button five time isnt right either IMO.
The VE table "should" be smoothed. Now if it looks kinda off, and it's only off say cell to cell by say 3% it's fine. But if your running closed loop, that's what the stft's are for, to trim or "clean up" when there are minor changes needed for a variety of reasons.
Now I will say, if your new or newish to tuning, in order to get those cells filled with good data, you need to learn to street tune. You need to be extremely sensitive with the throttle both in and out of the throttle, use good filtering to remove bad data from say, deceleration, or changes in tps greater than "x"%, or big map swings, and also set your minimum amount of cell hits required way up.
It takes a great amount of time, to get a really good, smooth, correct data set. This is why most guys really don't find "good" tuners. People think a $400 or $500 tune is outrageous, but on a modified combo, to get it right, and I mean cold start, tip in, shift points, ECT, it can take days of adjusting, logging, changing, chasing issues.
Now there are tuners who can whack a good number out of a tune, but if you want drivability, good cold starts, crisp tip in ECT, it's time and experience that gets results.
I just started over. I got the MAF within 2% in just about every cell that does not require spinning to banshee scream. I will tinker with VE next weekend. This weekend is NHRA!
Yep, in AZ. Last time I was home to go was 7 years ago. Changed jobs in Dec so I am not traveling all of the time.
Last edited by FCar2000TA; Feb 26, 2022 at 09:14 AM.
The VE table "should" be smoothed. Now if it looks kinda off, and it's only off say cell to cell by say 3% it's fine. But if your running closed loop, that's what the stft's are for, to trim or "clean up" when there are minor changes needed for a variety of reasons.
Now I will say, if your new or newish to tuning, in order to get those cells filled with good data, you need to learn to street tune. You need to be extremely sensitive with the throttle both in and out of the throttle, use good filtering to remove bad data from say, deceleration, or changes in tps greater than "x"%, or big map swings, and also set your minimum amount of cell hits required way up.
It takes a great amount of time, to get a really good, smooth, correct data set. This is why most guys really don't find "good" tuners. People think a $400 or $500 tune is outrageous, but on a modified combo, to get it right, and I mean cold start, tip in, shift points, ECT, it can take days of adjusting, logging, changing, chasing issues.
Now there are tuners who can whack a good number out of a tune, but if you want drivability, good cold starts, crisp tip in ECT, it's time and experience that gets results.
I normally have the car professionally tuned, but this time all I did was change MAF and tube size. I want to learn how to adjust the tune for minor changes.
A VE table should be pretty damn smooth...the physics says it should. In practice however, does it make a huge difference...not particularly. The algorithm that calculates air mass for the fuel control to reference is filtered. If you're just passing by one of those peaks briefly, it will have little impact. Sit exactly on one of them though and you might pick up some errors in the fuel trims. FWIW, I smooth tables fairly well because peaks and valleys tend to be from measurement error...not engine airflow characteristics.
Is this normal for a VE table? 
