Maybe that's the typical timing lead for target crank angle considering other things like taper, rich fueling, emissions, and knock sensitivity. I doubt they would have given much thought to smoothing Dr. Yagishocksumamoto's ignition map back to what he started with after they spent millions fine tuning it to what it is now. Would be nice to see a stock car's datalog though to see the load and timing, as well as afr, run.

 

V2 does not smooth as well as v1. It does look better though

 

here is my hacked up timing map. I think it looks better now.

 

hmm what makes you say that? Honestly.

I ask because the actual smoothing calculations are the same. Did you increase the smoothing limits from where I had them set at? If they are too low then it doesn't smooth as well. But they give you the ability to limit the amount of smoothing in each direction which can help if you don't want it getting richer or leaner. And you can copy paste the smoothed values back into the middle table to achieve even more smoothing, again limited to only going so rich or so lean.

Though I will say what I could look to do, to provide more range is reduce the amount of center weighting which will increase the effect of the averaging. Make the limits even more useful too. In an ideal world I'd make the center weighting variable too but I can't think of a way to do that without macros (or at least a macro function) and I was trying to avoid that.

 

I didn't mess with any limits...just copied and pasted.

 

ahh well I had the smoothing limits set to arbitrary values that, in most cases, would result in a slightly limiting smoothing on both the rich and lean side. consider setting them both higher, and watch the graph to see if things shift. If they do, then that was it. But play around with it because you might find that some facets of the map change, after most things are smooth, that you don't want to change.

 

Ver 2 is very slick indeed. I wanted to smooth the fuel map w/o adding any enleanment, so I set the lean limit to 0 and forced all of the smoothing to be enrichment. After copying, pasting, and resmoothing several time I hit 100% smooth. Very nice indeed. Engine runs VERY smooth.

Jumperalex do you plan to add the +/- limit functionality to the timing map as well? If so, great. If not, I may attempt it my self. I just don't want to hijack your project.

 

Excellent glad to hear it.

As for timing, yes I do plan on doing it. But it is also not my project to hijack, and any additional effort is of course welcome. That being said, I will likely work timing v2 tonight/tomorrow and do it along the lines of the fuel. That means having a table of unsmoothable cells in addition to the limits and scaling.

This raises a general question for those in the know ... How are you all dealing with areas of the map that are clearly "unsmooth" for a reason. In this case I'm mostly referring to the "lean ledge" in the fuel curve at about 2500 rpm, and the "mountain ridge" in the timing curve at 60 (or was it 80) Load?

I ask because applying smoothing to these areas seem to dull some things that I would call "good": High timing at 60-80 load for all RPM seems good for efficiency. The lean ledge at 2500 rpm for all loads seems good for the same reason.

In the last case I know at high RPM's we're in open loop anyway but my point is, is everyone smoothing these areas? I'm seeing some smoothed maps that seem to dull these features of the map in a way that is perhaps not what we would want. This is why I added the ability to identify some cells as unsmoothable so that I could allow more smoothing in certain areas without impacting others. I've toyed with a weighting factor (vs an on/off switch) for those areas but at some point it just becomes silly

Finally I was thinking about adding another table as an intermediate to allow easier copy/paste/repeat while retaining the original modded map; though truth be told at some point the stock map table become pointless for us and just be used as a "original" table

I'm also planning to add a macro to do the copy|pasting for you ... ohhhh but I just thought of a way to do it without a macro that will allow the ability to "scroll" thru various levels of repeat smoothing ... [sigh] this is why I don't sleep at night damn it. Easy enough to do actually; I'd canabilize some formulas I use here at work.

 

A copy & paste function would be very sweet - I was afreaid to ask.

An iteration counter would also be nice. I lost count of how many times I copy, pasted, and resmoothed my fuel map to reach 100%.

Looking forward to Ver 3.

 

- the reason I am thinking of doing it w/o a macro. It also would allow you to "undo" more easily than I could program a macro to undo. To be more specific, it wouldn't even be an undo. In essence my plan is to generate several levels of smoothing in hidden tables and then use one of the clickable rollers to select which level of smoothing to display in the table and on the chart. Then you can click back and forth back and forth to see what is changing where.

It is actually really easy to do using the offset function. What gets crazier is using definable ranges in conjunction with graphs to create scrollable, scalabe charts. But it starts with the offset function. Very powerful along with v/hlookup

BTW: how many times did you smooth do you think ... I'm hoping no more than 10 but I could do a couple of thousand to be crazy hehe)

 

I haven't used this tool yet. So which map do you copy and paste to (High Octane Ignition)? And why smoothed, modded and stock?

I'm assuming that if you have an untouched stock map-use stock, modified map-use modified, previously smoothed map-used smoothed, etc.

Previously, I had done it all one cell at a time.

Thanks - Jason

 

The High Octane stock map is there for reference to your already modded map. It just lets you see how far you've deviated and where via the bottom difference table.

The modded map is for ... well your modded map

The top, smoothed, map you should not touch unless you know what you're doing with the formulas. It is there to show you what the results of the smoothing are and you can also see what has changed relative to the original modded map via the difference tables. This is what you would copy and paste back into ECUFlash.

So ... the only thing you HAVE to do, is copy and paste your current map into the Modded Map table (the middle one) and then start playing with the limits. Generally, unless your maps looks like brail for people with no nerves in their fingers, any limit above about a 6 means effectively no limit. I just don't allow THAT much smoothing on the first pass because I find it tends to smooth features of the table that, IMO, should not be smoothed away because they are there for a reason.

Also, as mentioned but not really documented in the spreadsheet, is the table towards the bottom with all the 1's in it. Anywhere there is a one, that cell is copied dorectly from the modded table to the smoothed table with no smoothing. This is another expression of my feeling that some cells/features should remain unsmoothed. The classic example here is the lean-island of 14.7 values. The edge of that isle will get richened up if you don't have smoothing restricted there. I don't feel that area needs to changed from 14.7 If you disagree you can uncheck the green box at the top and/or remove the 1's

At some point I'll improve the documentation I promise

 

I see, thanks.

One more quick question, and I know you already touched on it, but am not sure you explained it in explicit detail; I just want to be sure so I don't skew the values and do it exact: how do I add load columns 280 and 300 and retain the proper formula averageing? All my tables extend to 300.

Again, many thanks.

Jason

 

add them as in add additional columns? That would take a few minutes of work that I can do for you easy enough when I get home. Way faster than I could explain it. Again, this is only if you need to increase the total number of columns which I didn't think we could do.


If instead you mean relabel the columns so that the last one is now 300 that is easy ... copy and paste your column headings from ECUflash into excel and that is it. Same with the RPM axis. The formulas do not take Load or RPM into account so they are only there for reference. You could delete them entirely and it wouldn't change anything.

 

Yeah, the worksheet has 19 load rows and I have 21. So I would need to add and recalculate, I believe.

How have the others been using it? From 260 load back to 0?

Thanks again,

Jason