Using Excel to dial in your maf curve (meth users, perfectionists, etc.)

[tavio025]

I'm using the template that was uploaded. I can't get the chart to display the MAF Voltage on the bottom of the chart that is created. Any suggestions are greatly appreciated. I also uploaded the excel sheet that I am using with all my values filled in.

[TiGraySpeed6]

No sweat-

Delete row 1 - DOn't need the "From" bit, and if that's not there it's easier to do the chart by just selecting the column

Select Column E & Column G

Insert > Scatter Chart - Scatter with only markers

Then back to following the steps as outlined

[tavio025]

That worked out great! Now I have the chart up, but now I cant get the chart to show the power curve. I can pick the other trends except that one and exponential.

[TiGraySpeed6]

There are two extra rows at the end/bottom hosing things up-

Select the chart, so the data in the columns have the purple & blue box around the data selected for the chart, then scroll to the bottom of the data and you'll see em. Grab the bar and move it up two rows to good data- both data boxes will move at once.

Then with the good data selected you can go back & do the trendline again and power will work

[tavio025]

You sir are a excel guru!!! Thank you!

[mrmonk7663]

Ok guys and gals, I made an updated template for everyone. It is essentially a complete walkthrough on how to perform what rfinkle has laid out. This should make it easy for even a novice to grasp, understand, and implement the calibration of the MAF based on the line of best fit. I hope this helps some of you!!! @rfinkle2;

Mike

[RE-ACTIONENGSPEED3]

Very nice...great help for us the knobs....

[shane5747]

I have been testing this to see if can get my AFRs more on target, but I had a question from what I've been seeing.

It seems every revision I use increases the corrected curve g/s and in turn just keeps increasing the g/s in ATR in each cell. An example is my 5.0v cell was 347 OTS, but after 3 revisions it was sitting at like 408.

Is there a point that this will eventually stop increasing the corrected curve? I'm worried this is going to make my OL AFRs go all over the place.

If it helps, I have a gen2 which maybe this really isn't effective for with OL maf cals.

[rfinkle2]

Originally Posted by shane5747 I have been testing this to see if can get my AFRs more on target, but I had a question from what I've been seeing. It seems every revision I use increases the corrected curve g/s and in turn just keeps increasing the g/s in ATR in each cell. An example is my 5.0v cell was 347 OTS, but after 3 revisions it was sitting at like 408. Is there a point that this will eventually stop increasing the corrected curve? I'm worried this is going to make my OL AFRs go all over the place. If it helps, I have a gen2 which maybe this really isn't effective for with OL maf cals.

The idea is to smooth out meth injection curves and some possibly hard to dial in intakes @ wot (open loop).

In some cases, it can help if you are using the same intake, but have upgraded the turbo and may see voltages much greater than your k04.

It is a predicted value / line of best fit given previous logs so if your wot logs are close to your targets, running the equation over and over again will sometimes exaggerate the end value.

You should be well below 5v @ wot on the k04 so no worries.

[ms3blackmica]

Agreed with above statement ^^^^^

If the curve is off by a good amount, using this method will help you out a lot.

If you are already on target and are just trying to do it to "smoothen" things out, it actually will exagerate your MAF curve by a lot.

[shane5747]

Originally Posted by rfinkle2 The idea is to smooth out meth injection curves and some possibly hard to dial in intakes @ wot (open loop). In some cases, it can help if you are using the same intake, but have upgraded the turbo and may see voltages much greater than your k04. It is a predicted value / line of best fit given previous logs so if your wot logs are close to your targets, running the equation over and over again will sometimes exaggerate the end value. You should be well below 5v @ wot on the k04 so no worries.

I guess I'm being too picky trying to get a straight line on my AFRs at WOT like I see a lot of the gen1s and ms6s. Currently I am only seeing a +/- .2 in my AFR curve at WOT which is pretty acceptable. I was only seeing how dialed in I could get the curve at WOT and if this method could do it better than traditional methods.

[CorkSport]

Anyone ever think about making a macro in Excel to do this? Would save time if it would auto sort the values and then make a MAF curve for you. At the end you could just copy your original MAF values in and it would auto give you and output.

[phate]

Enki was working on one. Takes quite a bit of work to interpolate between dynamic values (from data log) and generate the values for the mafv curve points.

[ms3blackmica]

Its a neat idea Brydon.

I prefer to manually set these things up though.

P.s. I still use this method now and then and it works every single time.

Thanks again Rob.

[phate]

Thinking about this again. I'm not sure VBA has the capability to grab an equation from a chart. I use vb to control a lot of things, but charts and pivots were a bitch in 2007. Maybe they've added some functionality since I last looked. That would be sweet.

But, if it doesn't, that brings us back to generating either an equation or a piecewise correction curve. I do the latter, manually, and it is quick and works really well. But, taking that logic that is in my head and putting it into code is rather difficult. This is where Enki was going with his program, it became extremely complex, lol.

[Outlaw4shrt]

@rfinkle2;

ok so I have a question, and maybe this has already been covered and is just over my head, but how do you calculate Commanded AFR's between rpm breakpoints from your Fuel OL/WOT Commanded EQ table? I'm assuming you interpolate correct? So between 3500 and 4000 rpm, for my commanded AFR as I transition from 11.8 to 11.4 AFR, I just interpolated based on the number of cells I had in my log the commanded values for each point. Is this correct? Does that make sense?

Anyone willing to check my work to make sure I didn't f' this up?

[mrmonk7663]

It would be much easier to command a baseline of 11.6 if you want extra safeguards by running richer, than going from 11.8 down to 11. 11 is much too rich. 11.6 will be fine....and more than likely 11.8 would be fine as well. The graph looks correct, but I would not use the graph until I picked the AFR I wanted for sure, maf cal manually as close as you can, then use the graph as needed. The graph will make your g/s higher, so you might have to multiply it by a reduced percentage...which will give you a tight curve with a smooth overall g/s transition.

[ms3blackmica]

Originally Posted by Outlaw4shrt @rfinkle2; ok so I have a question, and maybe this has already been covered and is just over my head, but how do you calculate Commanded AFR's between rpm breakpoints from your Fuel OL/WOT Commanded EQ table? I'm assuming you interpolate correct? So between 3500 and 4000 rpm, for my commanded AFR as I transition from 11.8 to 11.4 AFR, I just interpolated based on the number of cells I had in my log the commanded values for each point. Is this correct? Does that make sense? Anyone willing to check my work to make sure I didn't f' this up?

Typically what I do is target a static AFR first all the way to redline usually around 11.3-11.4. I will correct the MAF curve at this AFR until I am done.

Then all I have to do is just change my target as I like. You can taper the AFR richer as you please or whatever it is you are trying to achieve. Once dialed in, the car should hit AFR you command.

This will save you ALL of the hassle of trying to interpolate in excel and what not..

1) Target static rich AFR
2) Adjust MAF
3) Set AFR's the way you want.

[mymps11]

hey guys, wow!!..... firstly id like to thank rfinkle2 for starting such a helpful thread & secondly mrmonk7663 for his walkthrew.

Now i have 2 questions, firstly regarding the formula that mrmonk7663 speaks of in his spreadsheet, i got the formula F(x)=2.8082863306x^3.0526544764, he says "Take the "Y" Value of the formula and type It into the column "BASE Y". Copy the value down to the last ATR Voltage you have listed."

Then

"Now take the "X" Value of the formula , copy it, and paste the value in the column "RAISE TO X". Again, paste down to the last ATR Voltage Value. Your new Maf Curve will now be autopopulated."

Based on my formula is the first numerical the y value and second the x value?

Also with my newly posted maf curve from 3.13 volts & up it is smooth and predicatable as expected however the between the closed loop portion and 3.13 volts there is an imediate steep raise, this isnt ryte, have i dont somthing wrong or do u then horizontal interpolate from 3.13 volts backwards until smooth, however if i do that im worried my closed loop area will be too lean

Cheers Mitch

[Mfinlay04]

Originally Posted by mymps11 hey guys, wow!!..... firstly id like to thank rfinkle2 for starting such a helpful thread & secondly mrmonk7663 for his walkthrew. Now i have 2 questions, firstly regarding the formula that mrmonk7663 speaks of in his spreadsheet, i got the formula F(x)=2.8082863306x^3.0526544764, he says "Take the "Y" Value of the formula and type It into the column "BASE Y". Copy the value down to the last ATR Voltage you have listed." Then "Now take the "X" Value of the formula , copy it, and paste the value in the column "RAISE TO X". Again, paste down to the last ATR Voltage Value. Your new Maf Curve will now be autopopulated." Based on my formula is the first numerical the y value and second the x value? Also with my newly posted maf curve from 3.13 volts & up it is smooth and predicatable as expected however the between the closed loop portion and 3.13 volts there is an imediate steep raise, this isnt ryte, have i dont somthing wrong or do u then horizontal interpolate from 3.13 volts backwards until smooth, however if i do that im worried my closed loop area will be too lean Cheers Mitch

Sounds like your sample size is too small. If your basing your calculations off just WOT logs don't use that function to calculate closed loop.

I normally make a long log of all driving conditions and multiple WOT pulls and use all the data collectively for the first revision or two and then go back and tweak just the WOT part of your MAF curve. Shouldn't need much tweaking after doing the prior.

Sent from my SCH-I605 using Tapatalk 4

[Deldran]

So I am attempting to do this in excel 2013 and cannot get it to paste vertically anyone else have this issue. Or know how to fix it. Been banging my head on it for an hour

Sent from my XT1080 using Tapatalk 4

[phate]

Originally Posted by Deldran So I am attempting to do this in excel 2013 and cannot get it to paste vertically anyone else have this issue. Or know how to fix it. Been banging my head on it for an hour Sent from my XT1080 using Tapatalk 4

If it will past horizontally, do that first. Then highlight, copy, then paste special using the 'transpose' check box (cannot overlap any portion of the horizontal data).

[Deldran]

Thanks phate

Sent from my XT1080 using Tapatalk 4

[TiGraySpeed6]

If needed there are a couple of other cases earlier in this thread that go into detail with screenshots & such too

If it continues to be a problem post the log & we'll get you sorted

[mr.speedy3]

Originally Posted by phate Thinking about this again. I'm not sure VBA has the capability to grab an equation from a chart. I use vb to control a lot of things, but charts and pivots were a bitch in 2007. Maybe they've added some functionality since I last looked. That would be sweet. But, if it doesn't, that brings us back to generating either an equation or a piecewise correction curve. I do the latter, manually, and it is quick and works really well. But, taking that logic that is in my head and putting it into code is rather difficult. This is where Enki was going with his program, it became extremely complex, lol.

I'm a C# developer, sometimes, and would be willing to attempt to help. I haven't done any VBA in Excel for a while, but something like this may work:

Public Sub getEquation()
Dim trendData As Chart
Dim strEq As String
Set trendData = Charts(2) '<-- would be the chart name
trendData .SeriesCollection(1).Trendlines(4).DisplayEquation = True
With trendData 
    strEq = .SeriesCollection(1).Trendlines(4).DataLabel.Text
End With
End Sub

If working, it would generate a y = someNumber * x ^ anotherNumber

you would then parse that and get your x,y values.

Completely untested and randomly generated, it may even blow up your computer. I don't have excel on this box, but will test it out later on if no one else gets the gull to.

[phate]

Originally Posted by mr.speedy3 I'm a C# developer, sometimes, and would be willing to attempt to help. I haven't done any VBA in Excel for a while, but something like this may work: Public Sub getEquation() Dim trendData As Chart Dim strEq As String Set trendData = Charts(2) '<-- would be the chart name trendData .SeriesCollection(1).Trendlines(4).DisplayEquation = True With trendData strEq = .SeriesCollection(1).Trendlines(4).DataLabel.Text End With End Sub If working, it would generate a y = someNumber * x ^ anotherNumber you would then parse that and get your x,y values. Completely untested and randomly generated, it may even blow up your computer. I don't have excel on this box, but will test it out later on if no one else gets the gull to.

That's pretty close to what we need, but I'm not sure why it's not working. I created a pivot chart and added the trendline/equation. From there, I recorded a macro of activating the chart, then activating the equation text box (so I knew their names):



Altered your code a bit, but get a subscript out of range error:



I tried altering the chart name to reflect the .activate procedure's chart name and it still doesn't work. Not really sure why at this point (might be a pivot chart thing, haven't tried it with a regular chart).

[mr.speedy3]

Originally Posted by phate That's pretty close to what we need, but I'm not sure why it's not working. I created a pivot chart and added the trendline/equation. From there, I recorded a macro of activating the chart, then activating the equation text box (so I knew their names): Altered your code a bit, but get a subscript out of range error: I tried altering the chart name to reflect the .activate procedure's chart name and it still doesn't work. Not really sure why at this point (might be a pivot chart thing, haven't tried it with a regular chart).

Can you send me the template you have? Maybe I can tinker with it.

[phate]

Originally Posted by mr.speedy3 Can you send me the template you have? Maybe I can tinker with it.

No template, I just made a quick pivot chart from a data log I had. Threw in RPM as the x-axis and used MAF V for the data (not the correction since this was just a test). Added the trend line and data label from there.

[mr.speedy3]

Originally Posted by phate No template, I just made a quick pivot chart from a data log I had. Threw in RPM as the x-axis and used MAF V for the data (not the correction since this was just a test). Added the trend line and data label from there.

Som bitch! Alright, my brain probably isn't working right at the moment, but I'll take a shot at it. Worst case scenario, I'll do some tinkering at work when I get bored.

Minor update, @phate;

I recorded a macro, that did this:
create a scatter plot
select maf voltag (random) as x
selecy g/s (random) as y
add equation
copy equation
paste equation
end macro

 ActiveChart.SeriesCollection.NewSeries
    ActiveChart.SeriesCollection(1).XValues = "=datalog2!$K$2:$K$55"
    ActiveChart.SeriesCollection(1).Values = "=datalog2!$L$2:$L$42"
    ActiveChart.SeriesCollection(1).Select
    ActiveChart.SeriesCollection(1).Trendlines.Add
    ActiveChart.SeriesCollection(1).Trendlines(1).Select
    Selection.Type = xlPower
    Selection.DisplayEquation = True
    ActiveChart.SeriesCollection(1).Trendlines(1).DataLabel.Select
    Range("U23").Select
    ActiveSheet.Paste

Wife just got home, feel free to tinker.

[mr.speedy3]

So I am almost positive this is what we need:

    ActiveChart.SeriesCollection(x).Trendlines(1).DataLabel.Select
    Range("y").Select
    ActiveSheet.Paste
'would vary, most likely 1 though
'y would be the desired location of the cell.

After that we can take the equation, format it as normal text (you can use "paste special" and past as values only). Then use a vairant and split the string. You may have to split it twice, 1 time between = and x, the other everything after x. This way you get x and y. Seems simple. But I have to go back to work. So I'll use this as a record of theory.

I'll give it more though later in the evening if that doesn't work!

[phate]

Originally Posted by mr.speedy3 So I am almost positive this is what we need: ActiveChart.SeriesCollection(x).Trendlines(1).DataLabel.Select Range("y").Select ActiveSheet.Paste 'would vary, most likely 1 though 'y would be the desired location of the cell. After that we can take the equation, format it as normal text (you can use "paste special" and past as values only). Then use a vairant and split the string. You may have to split it twice, 1 time between = and x, the other everything after x. This way you get x and y. Seems simple. But I have to go back to work. So I'll use this as a record of theory. I'll give it more though later in the evening if that doesn't work!

I'll be able to play with it later tonight, also. I did try using activechart, but I was attempting to set a chart variable to it... looking at that code make me think it's possible and way simpler lol.

[mr.speedy3]

Originally Posted by phate I'll be able to play with it later tonight, also. I did try using activechart, but I was attempting to set a chart variable to it... looking at that code make me think it's possible and way simpler lol.

Any luck there? I am trying to find time this week to look at this, but my boss keeps trying to send me to Taiwan . So I haven't had much time in between this and tuning lol.

[rfinkle2]

Thanks to you guys mr. Speedy and Phate for taking this to a whole other level.

BTW, this does have its place when trying to use the aux fueling kits available now.

It is much easier (and more accurate) having excel predict the shape of a fuel curve in voltages not previously seen. LOL.

[phate]

Originally Posted by mr.speedy3 Any luck there? I am trying to find time this week to look at this, but my boss keeps trying to send me to Taiwan . So I haven't had much time in between this and tuning lol.

No, Thursday and Friday were getting ready for an autocross, then Saturday and Sunday were busy autocrossing/fixing car/autocrossing, lol. Haven't had much time for anything since we got back. This is the new project that I'm part of. Slightly faster than my MS6, haha.

YouTube Short

YouTube Short

Everything I've learned tuning the speeds is helping tune this car; including the MAF correction stuff we're discussing here, lol. But, having a completely open ECU (megasquirt III) is a real eye opener in terms of what we can and cannot do with the speeds. Some things I greatly wish we had, and others I wish we could port over to the MSIII, haha.

[rfinkle2]

So jealous of those build programs and competitions... I had a friend who built some of one of those cars and it seems like such a great experience.

What engine is in that car Clint? A motorcycle engine?

[phate]

Originally Posted by rfinkle2 So jealous of those build programs and competitions... I had a friend who built some of one of those cars and it seems like such a great experience. What engine is in that car Clint? A motorcycle engine?

It's a Kawasaki ZX-6R engine. 600cc 4 cylinder (makes ~80hp with a 20mm intake restrictor). It doesn't really wake up until ~8500 rpm, and we're revving to 13k. Hoping to get some solid tuning time in the next two weeks to build power in the mid range and get more throttle response out of it.

[udntknw]

Edit: I'll post this in my thread... Not to sidetrack this one.

[mr.speedy3]

Originally Posted by phate No, Thursday and Friday were getting ready for an autocross, then Saturday and Sunday were busy autocrossing/fixing car/autocrossing, lol. Haven't had much time for anything since we got back. This is the new project that I'm part of. Slightly faster than my MS6, haha. Clint SIUC FSAE Autocross - 10/20/13 - Run6 - Fastest & FTD (45.516) - YouTube Everything I've learned tuning the speeds is helping tune this car; including the MAF correction stuff we're discussing here, lol. But, having a completely open ECU (megasquirt III) is a real eye opener in terms of what we can and cannot do with the speeds. Some things I greatly wish we had, and others I wish we could port over to the MSIII, haha.

Hey now... you COULD go megasquirt on the MS3.... it'd just take a shit ton of work lol
'
Not sure where the "Your mother"emoji came from... but I'll keep it .

[m4tic]

mmm dat curve

i just discovered this thread, and tried it out using the template... wow

[udntknw]

I've used this on my cp-e cold air intake and was pretty well spot on. Now that I'm using the JBR 3.5, I'm getting consistent lean spots. Is that normal for the MAF to not be linear, or could something be throwing it off?

Sent from my ADR6300 using Tapatalk 2