Ford Focus Forum banner
Cancel
Create thread New Garage Forums

Aftermarket MAFs versus OEM Ford MAFs

Jump to Latest
21 - 40 of 40 Posts
#21 ·
Interesting read on MAF's I stumbled on the other day: Link

And Rich...sixth order just means: ...+X^6 like in y = aX + bX^2 + cX^3 + dX^4 + eX^5 + fX^6 I'm sure you've seen that before.

What it means though, is that if you know x, you can calculate y. Or vice versa. So if x is the voltage, then y would be the coresponding flow.
Click to expand...
step in......listen, look around uneasily.....and quietly excuse myself after saying "o..k....that sounds about right guys"
Image
 
Save
Reply Quote Like Like
#22 ·
Interesting read on MAF's I stumbled on the other day: Link

And Rich...sixth order just means: ...+X^6 like in y = aX + bX^2 + cX^3 + dX^4 + eX^5 + fX^6 I'm sure you've seen that before.

What it means though, is that if you know x, you can calculate y. Or vice versa. So if x is the voltage, then y would be the coresponding flow.
Click to expand...
step in......listen, look around uneasily.....and quietly excuse myself after saying "o..k....that sounds about right guys"
Image
Click to expand...
Come on. Its not that complicated.
Image

The 9 points supplied with the flow sheet are simply (x,y) coordinates that create a curve (just like the curve on a dyno). You plot them on a graph, and connect the dots with a smooth curve. For any curve, there is a mathamatical function that can describe that curve. That mathmatical function is pretty hard to come up with by hand so its usually done using computer software. You input in the 9 data points and it spits out a function like I described above. Except there would be actual values for the a, b, c, d, e, & f. Once you have that function, you can enter one of the variables to determine the other.

If x = MAF voltage
and y = flow

Then if you enter in a voltage number and it will tell you what your flow is.

Now you take the function that you just determined to describe your MAF flow characteristics, and enter in the voltage values that you want coresponding flow numbers for in your 30 point MAF transfer function table.

I'm not the best at describing math and I am in no way a math major so feel free to correct me if I'm wrong or unclear...

But if you can get a 30 point flow sheet from Pro-M like Tom says, that would make life much easier.
 
Save
Reply Quote Like Like
#23 ·
How do we go about getting these 30 Point prom flow sheets? I bought mine from CFM, would contacting them get what I need? Or is this something I should be going through directly Pro-M?

Image
 
#24 ·
I've re-read through the thread, and I think there is some statements that could be misleading. from what some people were saying, it's like you don't even need the 30 points of the MTF for the car to run perfectly.. just enter the 9 points, and somehow the ECU has the wherewithall to determine the function that describes the MTF.

For example, Tom's following statement is a bit confusing:
The ECU looks at the MAF at every .000 voltage NOT at the 9 or the 30 points that they are calibrated at like some would have you to think.
Click to expand...
this leads me to believe that the MAF only looks at 5 points along the 5v curve at 0.000, 1.000, 2.000, 3.000, 4.000 and 5.000, but it should be interpretted as the EEC seing everything from 0.000 to 5v in .001 increments... and even then, the EEC software (or SCT as it were) sees everything in analog-digital counts (take voltage, and multiply by 205), and has 30 elements in the MTF table to describe the function. It doesn't matter where the points lie on the A/D axis, so long as you cover the 5 volt range of the MAF to your satisfaction.

Couple that with the statement:
There seems to be some confusion on this point. It doesn't matter how many sample points are used when calibration is done on the MAF. That does NOT define the Transfer Function table in the PCM. The Transfer Function in the PCM ALWAYS contains 30 mapped points. These values in the PCM need to be matched to the output of the MAF (Aftermarket or OEM).

The Aftermarket MAF does this in the opposite direction. By "calibrating" it's output to the STOCK transfer function with a shift in values to compensate for different sized injectors. It is then CHECKED at 9 points. That doesn't alter the 30 points in the PCM. If you ordered an Aftermarket MAF calibrated to STOCK injectors it should closely match the OEM MAF output.
Click to expand...
and people are led to believe that the 30 points in the Mass Transfer Function don't seem to matter... and that's sort of misleading...

couple the two arguments together, and one gets the impression that the EEC is able to calculate the MAF function from whatever points are available in the mass transfer function table.. wether it be from 9 points, or 30.

Which I'm pretty sure is not the case. Though I wouldn't mind being proven wrong... I'm still just learning.

So regardless of how many points pro-m gives you, you still need to create a MAF transfer function to make it work. You could get away with just calibrating the engine displacement for the injectors that the MAF is calibrated for, but you still have to go back and tweak the MAF transfer function.. .


I think that it really boils down to the following:
Regardless of how you slice this pickle, you still need to enter in 30 points worth of data into the MAF Transfer Function with your MAF. with an aftermarket MAF, you can get them from a 30 point flow sheet you request from pro-m, or you figure it out from the 9 point sheet they give by default--draw your own graph, figure out your own equation and make it work. If you go ford oem, the values are easily available in files and are 99.99% correct.

p.s.
if you have a hard time drawing graphs and coming up with your 30 point maf transfer function from the pro-m flow sheet, check out this excel spreadsheet ... it's the closest 'automated tool' I could find, but it's not perfect. You'll need to subscribe to the yahoo newsgroup to make it work.

I found this link in this article .. both good reads if a bit advanced.
 
#25 ·
Man you guys are gonna kill some brain cells in a noob
Image


I had a Pro Flow MAFS when i first got my car after reading Dennis W , back in hte day, posted about getting 14hp or so from one.

I do know when you order one it is calibrated to your cars year, and your ecu code, and what type of filter you are using. Pretty car specific like a chip.

The first one i had seemed to work great. but i got rid of it for some odd reason thru all my changes.
Image


I then ordered another and they asked all the information i listed, its seemed ok but no great difference. I then read the tuning book and they said it really did nothing, so i swapped back my stock one and the car ran better.

I will say the readouts they send you for both of the ones i had were different with a different COB code and my car hasnt changed.

I was at New EDGE when TOny got a shipment of 3 in a few years back and each one had a different COB code and didnt mention what year car they were for either.

I personally think its a Hit or Miss on them.

But Talking to Ray at McNews, he has a Mafter burner type unit so you can program the MAFS and fine tune it. Id say thats the best way to go.

I believe Toms test all his so he knows, they have the correct calibration needed. Cause I think Tom used to get 10 in at a time from ProFlow and have to send a few back.

But like tom said unless you are runing out of injector or make alot of HP you really dont need it.

Just my Big .02
 
#26 ·
Come on. Its not that complicated.
Image

The 9 points supplied with the flow sheet are simply (x,y) coordinates that create a curve (just like the curve on a dyno). You plot them on a graph, and connect the dots with a smooth curve. For any curve, there is a mathamatical function that can describe that curve. That mathmatical function is pretty hard to come up with by hand so its usually done using computer software. You input in the 9 data points and it spits out a function like I described above. Except there would be actual values for the a, b, c, d, e, & f. Once you have that function, you can enter one of the variables to determine the other.

If x = MAF voltage
and y = flow

Then if you enter in a voltage number and it will tell you what your flow is.

Now you take the function that you just determined to describe your MAF flow characteristics, and enter in the voltage values that you want coresponding flow numbers for in your 30 point MAF transfer function table.

I'm not the best at describing math and I am in no way a math major so feel free to correct me if I'm wrong or unclear...

But if you can get a 30 point flow sheet from Pro-M like Tom says, that would make life much easier.
Click to expand...

thanks, that was perfect.............for future reference sock puppets help me learn as well....
Image
 
Save
Reply Quote Like Like
#27 ·
I've been researching Mustang MAFs for when the SCT flasher comes out, and something people brought up was the OEM work best in a suck-through operation, and for a blow-through, you are better off with something like a Pro-M. Opinions?
Click to expand...
that IS true...but i'm not a fan of blow through MAFs, turbulence is gonna be higher downstream of a supercharger, and as we all know, MAFs are HIGHLY sensitive to turbulence
 
Save
Reply Quote Like Like
#28 ·
1. What led you to make said purchase?

My philosphy that if you are going to modify a system you need to modify all parts of it.

2. What brand and size?

Pro-M 70mm calibrated for stock injectors

3. What mods do you have?

Volant, Pro-M, FC 65mm TB, 2K manifold, FR Stage 2 cams at 0,0, Screamin Demon coil, FS street header, RT Cat, FC Flex, FS Stealth cat-back,

4. How did it work?

Very Well. Fitting took 20 min to openup hole in Volant. I think that at 135TQ / 132HP I am at the point where the benefits are just begining to show. If I were to go to a raceheader, chip, and some cam tuning I think that I would be maxing out the stock MAF. The one time that I took it off and put on the stock MAF(trying to solve a problem that turned out to be unrelated) I noticed a difference in performance. Not in the peak output but in the time required to reach that peak. In other words, it revved more quickly with the Pro-M.

5. Was the MAF worth the money?

IMO, Yes

6. Would you do it again? Please explain in detail.

Yes, See #4 for reasons.
 
#29 ·
Don't make this Transfer Function table any harder than it really is. It's just a lookup table with 30 values that represent A/D counts for a particular MAF. The final value always represents the same voltage, 15.9998 volts. This is used as an infinity value so that the table is never outranged. The first value represents a voltage in the range of 0 to .5 volts and is a variable defined as the minimum MAF count. That leaves 28 other entries that are a 6th order polynomial curve fit to the output of that particular MAF.

When the PCM reads the MAF voltage it converts it to a digital count. That count is compared to the values in the Transfer Function table. When the actual count falls between 2 entries in the table (most of the time it will) the PCM will do a linear interplotation of the count to derive the actual flow data. It's not doing any complex calculations.

Check this out for a closer look at the calculations....
MAF Questions and Answers
 
#30 ·
Dyno'd before and after Pro-M install and there was NO change in HP, TQ, or curves. I have had repeated dyno's since, without any gains. I don't think it's needed unless you F.I. for run some major N/A HP.
 
Save
Reply Quote Like Like
#33 ·
Dyno'd before and after Pro-M install and there was NO change in HP, TQ, or curves.
Click to expand...
Thats what speaks VOLUMES there. The Dyno results.
Click to expand...
EXACTLY I made power with mine. You need to make enough power that the engine needs a higher volume of air. I think the right combination of parts before and after the MAF makes a difference.
 
#34 ·
I should have posted some of the mods:
FS head (decked)
Kent cams V.3
cool flo
a host of other parts, but the important ones are above. I still have the stock intake manifold and TB, tan an aftermarket TB for a while, no gains there and drivability went in the toilet in traffic. I think if I ran the FR intake mani, it may make a difference here with both the TB and MAFS, but the stock mani is quite a bottleneck. I'm currently at 160 whp 142.6 tq.
 
Save
Reply Quote Like Like
#35 ·
Here's a few answers and opinions.

Pro-M MAFs vary from unit to unit, even with the same calibration. Each unit itself if very consistent. What this means for the end customer is if you are dyno tuning, the Pro-M is fine. If you want a mail order chip a Ford MAF is better because they are consistent from unit to unit.

To convert from a 9 point sheet to a 30 point function we use a 6th order curve fit program. The 9 point sheet is fine as input. i think Pro Flow charges $75 for a 36 point flow sheet.

Once the transfer function is entered in the EEC it does not use a 6th order polynomial for voltage (a/d counts) between points. It does a linear interpolation instead.
 
#36 ·
Wow I'm late to the party as usual, just wanted to confirm what I know "from the inside".

Darkpuppet, it's pretty cool you're here. You're kinda picking up where I left off... learning stuff, the right stuff, and breaking it down for others. You're right on track with this one:

The EEC does NOT do a 6th order polynomial on the MAF transfer (Xfer) curve. It does a linear interpolation between points. It looks at the voltage from the MAF, and finds where it falls between two points on the X axis of the MAF Xfer. Then does a linear interpolation between those two points to find the airflow.

An interpolation requires what, a couple lookups, and 4 calculations? That's easy easy stuff for a computer. It does this, however many times per second.

The EEC does not do a 6th order polynomial, or even a second order polynomial...
Image
on the input values. That would be a LOT harder for it to do, and I believe running a 6th order polynomial would actually require MORE calculations than a linear interpolation.

When you're talking about a table with 30 values, that's PLENTY of points to make a decently smooth V=airflow graph.

Now, here's where we get back to the 9 point thing.

I'm guessing here, but I would bet you could enter that 9 point table into your SCT, and I think the EEC can work off of it. It would do th same linear interpolation, but it would do it between 9 points instead of 30. So the points are further apart. The interpolation lines which "bridge" the points are bigger.

Here, you'd really see the lines. If you overlaid the 9 points and 8 bridging lines, over the ACTUAL curve, you'd see the error between all the points. If you overlaid the 30 points with 29 bridging lines, you'd see the lines are MUCH closer to the real curve.

THAT is why it's better to use 30 points. Even if you yourself do a 6th order polynomial on the 9 points, get your 30 points, and input those, that is still going to be more accurate than using the 9 points.

Remember, computers are fast, but we're still smarter.

In theory, you could even try to run your car off two points. 1V and 5V. It would interpolate between those points, but the error would be huge.

As for the question of OEM vs. aftermarket MAF's... it's like just about everything else. The quality of OEM parts is usually better than anything in the aftermarket, so use OEM parts if you can find ones that have the performance you're looking for.
 
#39 ·
Just one more point on the 9 pont vs. 36 point vs. 6th order polynomial following on to Rob and sct's posts.

When you have only 9 data points, the ECU wants 30. So as Rob mentioned you COULD just connect the dots (linear interpolation) but the error is large. Instead, you can FIT a 6th order polynomial to those nine points, and then use the resulting curve to PREDICT where the remaining 21 points should be. This will help reduce your error, but even a 6th order polynomial curve fit is not perfect esp. with only 9 points to start with.

If you want a "perfect" curve fit, you need to use an equation that accurately describes the heating and cooling of the wire element in a wind tunnel. So you would really want to fit a Stokes equation but that would be way too complex for most people so the 6th order polynomial plus some tuning is usually "good enough".

The best way to get 30 data points is to MEASURE it rather than trying to PREDICT it, on the actual car. Pro-M will do this is you are near their location.
 
Save
Reply Quote Like Like
#40 ·
Jeff, I agree 100% with what you just said... except I think you misunderstood something I said.

I never suggested a person use linear interpolation to connect the dots. The ECU would do that if you only gave it 9 points. I suggested what you say, do a polynomial to predict your 21 missing points.

Given 30 points, the ECU will still interpolate between points anyway.
 
21 - 40 of 40 Posts
About this Discussion
39 Replies
19 Participants
Last post:
P-51
Ford Focus Forum
posts
3.6M
members
67K
Since
1989
Forum community dedicated to Ford Focus owners and enthusiasts. Come join the discussion about RS/ST performance, modifications, troubleshooting, maintenance, and more!
Full Forum Listing
Explore Our Forums
TEAM Pit Stop Ford SVT Focus Headquarters Forced Induction & Nitrous MK I Tech New Member Introductions