MoTec Scatter Plot Analysis for tune data

Hi everyone,

Looking for some thoughts and opinions on ways to further utilize scatter plots in MoTec i2 to analyze large data sets. This is for MoTec MX00's and M1 series ecu's.

Recently I have been using MAP X RPM against fuel injector duty and lambda, aswell as long term fuel trims as a colour plot. I use MAP X RPM to look for a straight line with no abnormal peaks or troughs and large changes in the colour plot. How are others doing this? From my testing so far I've found that all the tunes that have a nice correlated line for MAP X RPM have been 'good' tunes with respect to drivability, power output and client feedback.

For the long term trims with MX00's ecus , I haven't applied the following changes just yet but this is what I'm looking to do:

Identify the pockets within the colour plot were the trims are excessive, then apply the correction to the fuel table, then smooth out the neighbouring areas to maintain the shape of the fuel table that was built on dyno. Thoughts?

The Mx00's ecu's are more old school, compared to say a haltech, but I believe the long term trim principles are the same?

I've attached 2 photos. One is of a tune were the MAP X RPM shows a correlated line. This vehicle tuned up real nice and made good power, approx 410hp at the wheels (on a VERY old dyno) 355ci Holden V8, Procharger making 11psi and 4L60e transmission.

The other photo is a Subaru Sti were I was having various fueling issues and the tune wasn't right. Because of this I needed to find another way of analysing large data sets quickly and make the necessary changes on the road.

These are both MAP based engines, but what about N/A. The scatter plot will be shaped differently, but the the principles of looking for trends will still apply?

Appreciate any input!

Thanks,

Harry.

Anyone have thoughts on this?

How do you filter out transient data?

Can this be achieved on i2 standard? Otherwise using MLVD is the only way I know so far?

No you will probably require I2 Pro to build a filter function (Math channel using "Choose()" to make the channels you are plotting "invalid()". On a graph this will show up as gaps in the trace, in scatter plots this will just be missing data points (the ones you want filtered out).

My question was "What function have you developed to filter the data?"

Harry,

Get Tim to put his hand in his pocket and get the "Open i2 Standard file in Pro" enable for your laptop.

I like Stephen's suggestion. :)

Stephen,

I wish :) I'd have better chance winning 30 million lotto through a scratchy ticket. However, I'll definitely look at buying this myself so thanks!

David,

I haven't developed a function yet. So, let me rephrase my question. Where would be a good starting point to look up equations/functions/math laws to filter transients?

How about something like:

Choose(abs(Derivative('Manifold Pressure'[kPa])) < 5, 'Lambda 2'[La], Invalid()). So the data would only contain situations where the Manifold Pressure was changing less than 5 kPa / sec. Or you could use RPM rate of change, depends on what you want to examine.

Great thank you David, I can work with that. I will get this Pro laptop upgrade and see what I come up with