View Single Post
Old 02-23-2023, 07:29 PM   #15 (permalink)
Ryan_z34
Base Member
 
Join Date: Mar 2021
Location: Sydney, Australia
Posts: 2
Drives: 09 Nissan 370z 7at
Rep Power: 4
Ryan_z34 is on a distinguished road
Default

Quote:
Originally Posted by ByThaBay View Post
It doesn't matter if I tune the car on a dyno or if I tune it remotely. I ALWAYS require lots of data collected over a long period of time to track trends with the fuel system , the knock correction system, and other variables. Here's how it works:

each driving scenario results in a different engine speed & engine load. So you have tables in within the ecu and most of them are indexed as such

X: engine load
Y: engine speed

As you drive, the ecu does a lookup in various tables to control fueling, vvel, cam timing, etc etc. Here is an example of a timing map:



So as you can see there are many cells. If you are only tuning full throttle at sea level, then generally your load will stay around 100% as your rpm increases, this is a very small portion of the "tune".

There are some scenarios where you will have a very hard time reaching them on the dyno and other scenarios that are difficult to reproduce on the dyno, so real world driving is required.

In addition, and this is probably the most important part, you can't just go off of a small sample. You need to drive the car as conditions change and repeatedly log the vehicle to determine the ideal values for each cell. One day your car might pull timing, the next day it might be fine, so the best value will lie somewhere in between. It's common to see knock sensors desensitized or turned off completely to avoid having to do this extra work. Analysis of why a car pulls timing can be complicated and made more difficult by lack of tooling, especially in a remote tuning situation.

Seeing as how the values may change somewhat significantly, some smoothing of the tables will be needed and this will again require additional adjustments until there is a convergence..

This process can take a while to complete depending on the vehicle, there is no other way.

Each tune I do comes with unlimited revisions and data processing/review, until no more practical changes need to be made. You can keep collecting data for the entire duration of the tune support. The more data you collect, the more accurate the results of the tune will be. It's only in your favor to collect as much data as you can over a longer period of time to get the best results.
Exactly right. Tuning on a dyno is great, but airflow characteristics change once you get on the road (air speed, angle, covered bumper, temp etc), so ultimately you end up needing to collect data in usual driving situations.

Having large data sets increases confidence when making tune changes because outliers don't skew the value. You also don't always catch knock in certain load/rpm cells in a single drive or weather condition. I don't understand why people are complaining about starting a data log for a drive they would have already done when it means a more accurate tune that covers wider drive conditions?

Plus, after your first couple of revisions with Eugene, you still probably have a better tune than most out there. The last few revisions is icing on the cake.
Ryan_z34 is offline   Reply With Quote