9pt to 30 pt MAF conversions

[TT331FB]

What math model is used to perform these conversions? My guess would be a polynom curve fit?

Also, I noticed that Tweecers conversion produced seven points below .4 MAFv. Given that my car idles at >.5 MAFV, these lower points are totally useless and result in less curve accuracy where you need it. My hunch is that the EEC interpolates between points and the distance beween these curve points increase error as it reduces resolution in the normal operating area of the MAF curve. Why have points below the min MAFv?

Anyway, my plan is to do a curve fit and check it against what Tweecer does. Once I have the curve solution, I can pick my own MAFv points and they will start a little below my exspected min MAFv.

Input?

Thanks, Eric

[Dale McPeters]

Eric,

If you use Jack Hidleys spread sheet from the Yahoo EEC-Tuner list you can make it 5th, 6th or whatever order fit that you require for your application. Clint has also added a maf fitting tool to the EA if you have it.

[TT331FB]

Hi Dale,

I was motovated to come up with my own curve fit spreadsheet just to learn.

I used a sixth order with maximum error of 1.3% between calibrated and predicted data. I customized my own MAFv points to get more resolution in the ranges I operate in. Happy to send it to anyone. Just need a e-mail address.

Thanks!

[86GT]

Eric
You are correct, the EEC does a linear interpolation between two point on the curve. You said that your car idles at .5 volts. You will need one or two points below that and here is why. When you snap the throttle body closed the velosity of the air moving into the engine hits a brick wall (TB plate) and bounces back. This reverse air flow causes the MAF to read lower values. This may only happen in a split second, but if it happens at the point at which the tweecer loggs then you will see it.

When I first saw this, I thought it was data garbage data being droped by the tweecer. It was sparatic and non I could not duplicate it. When I finally started snapping the throttle, it was repeatable.

The lower MAF voltages may also come into play during decel. In the A9L there is a table set aside for MAF backflow. This table is what got the gears turning in my head.

The EA can also do the conversion for you. It can also do the custom voltages. Send me the flow sheet and I will post what I get.

[QWKSNKE]

Hi Dale,

I customized my own MAFv points to get more resolution in the ranges I operate in. Happy to send it to anyone. Just need a e-mail address.

Thanks!

If you would like, you can upload the file in our EEC Archive.

[Italian LX]

The EA can also do the conversion for you. It can also do the custom voltages.
How come when I save a MAF curve in CalEdit, the curve comes up in EA with some weird numbers? Is there any way around this, or am I doing something wrong?

[86GT]

The export feature of the Tweecer is FUBAR. The voltages actually get exported in about half the real values.

[86GT]

Eric
Here is what i come up with for you MAF that you posted in the other thread.

[Dale McPeters]

The export feature of the Tweecer is FUBAR. The voltages actually get exported in about half the real values.

:D.......:wave2:

[TT331FB]

Eric
Here is what i come up with for you MAF that you posted in the other thread.

Clint, Thanks!

I plotted the EECA results against my 6th order fit and got a match at all points except at 5.0 MAFv . My fit says 1802.252, yours says 1914.198. Big difference. I will send you my excell sheet for your review.

[86GT]

I will take a look. The EA uses a third oder poly not sixth, so it may have a little difference.

[Dale McPeters]

Big MAF's 6th order on bottom of curve, 3rd order on top of curve. Gives more resolution down low?.......... :popcorn:

[TT331FB]

Big MAF's 6th order on bottom of curve, 3rd order on top of curve. Gives more resolution down low?.......... :popcorn:

I suppose it can be done, splitting the curve up. Generally, the stranger the curve, the more coefficients needed to get a solution. MAF curves are fairly simple but my software does a 6th order as easy as a second order solution.

It could be a simple data entry mistake.

[QWKSNKE]

my migraine is coming back after reading this thread :jester:

[Dale McPeters]

my migraine is coming back after reading this thread :jester:
:rofl:

[Dale McPeters]

I suppose it can be done, splitting the curve up. Generally, the stranger the curve, the more coefficients needed to get a solution. MAF curves are fairly simple but my software does a 6th order as easy as a second order solution.

It could be a simple data entry mistake.

Yep it can be done. I am running one now that I did this way.
In your case it may just be an entry mistake though...

[Dale McPeters]

Clint,

I know you have comments on this so let's hear them..... :jester:

[TT331FB]

Clint's working on it. I'm sure he will have something soon.

[Dale McPeters]

Thanks Eric.

[86GT]

I personally did not have any problems with the third order, but I do understand that the sixth order will give better resolutions. I did not implement it on the current versionbecause it is a night mare to implement. The difference that Eric is seeing is related to the order of the polnomial. I have found a way to add the sixth order to the EA and will do so on the next revision.

Spliting the orders based on the location in a MAF may not be the best approach. Im not to sure of the advantages of doing this.

[TT331FB]

I personally did not have any problems with the third order, but I do understand that the sixth order will give better resolutions. I did not implement it on the current versionbecause it is a night mare to implement. The difference that Eric is seeing is related to the order of the polnomial. I have found a way to add the sixth order to the EA and will do so on the next revision.

Spliting the orders based on the location in a MAF may not be the best approach. Im not to sure of the advantages of doing this.

Thanks Clint, I don't see this as a big deal. I agree that splitting the curve could be problematic. Assuming that a sixth order is a good fit should not be jumped to without actual calibration info at 5 MAFv to compare to.

[TT331FB]

Anyone see any problem with using MAF pt 30 MAFv as 4.99 instead of 15.999 typicaly seen? Why waste the point with a voltage not readable by the EEC.

[86GT]

As long as the 'MAF Voltage Maximum' value in the scalar is at 4.999 it should not be a problem.

[Dale McPeters]

I personally did not have any problems with the third order, but I do understand that the sixth order will give better resolutions. I did not implement it on the current versionbecause it is a night mare to implement. The difference that Eric is seeing is related to the order of the polnomial. I have found a way to add the sixth order to the EA and will do so on the next revision.

Spliting the orders based on the location in a MAF may not be the best approach. Im not to sure of the advantages of doing this.

Clint,

How big of a deal would it be to have the 3rd & 6th order options available in the EA? Would it be to much of a nightmare to implement?

On the item of spliting the orders based on the location etc. of the MAF. It may not be the best approach but so far it is working fine.....at least until I can get the AEM setup and get everything changed over..... :jester:

Of course I would not recommend this to anyone else as my situation is a little different than others would be.

[86GT]

I am currently woriking on implementing the sixth order as an option.

[Dale McPeters]

I am currently woriking on implementing the sixth order as an option.

Thanks Clint.

[86GT]

I have it completed now, but it need to be tested.

[86GT]

The new version is up on the web site. It now supports 3rd, 4th, 5th and 6th order polynomials.

[Dale McPeters]

The new version is up on the web site. It now supports 3rd, 4th, 5th and 6th order polynomials.

Thanks Clint, You the man..... :metal:
I will get it downloaded this weekend and try it out on a few
flowsheets that I have and on the GT....... :banana:

[Dale McPeters]

Clint,

I tried to download the latest rev of EA 2.6.6.6 last night and
it kept telling me that the zip file is corrupt? Got any suggestions?

[Dale McPeters]

Got it Clint it was the new anti-virus and firewall that I had installed...

[86GT]

Cool