MAF Curve for Roush Sealed Airbox

[CCS86]

True the difference is not big but when trying to get to get exact it makes a difference. Luckily these ECUs are fast and have widebands so it is not that big of a deal

Is that a yes or a no? Have you datalogged Coyote rail pressure?

First you laugh at the idea of people wanting to use accurate injector data. Now you say "the difference is not big". What are you basing these statements on?

Quit speaking so definitively, if you have no data to back it up. Join us in a discussion. Ask questions. Theorize.

[murfie]

I just want to point out the inferred MAP is just as important as the inferred rail pressure. If your ECU is thinking there is 5 psi more pressure in the manifold than there actually is, that could end up in your MAF curve with out a known good curve to start from. You start ending up over 2.0 load at less than 15 pounds of boost, your engine is just not likely to be that efficient.

[caniggia]

Correct, they are important. When I was doing first supercharged car I had to log rail pressure with external gauge because I had problem to setup good maf curve, sometimes I was getting lean spots, sometimes rich and missfire. Injectors and inferred line pressure must be correct.

[tbrtuning]

Is that a yes or a no? Have you datalogged Coyote rail pressure?

First you laugh at the idea of people wanting to use accurate injector data. Now you say "the difference is not big". What are you basing these statements on?

Quit speaking so definitively, if you have no data to back it up. Join us in a discussion. Ask questions. Theorize.

Yes. I have wired in analog pressure sensors for MAP and rail pressure on different forms of boost and experimented quite a bit with it.

If you run one to the point if dropping fuel pressure on E85 things start becoming noticeable. Depending on boost pressure boosted with just a voltage booster will start dropping fuel pressure and fuel trims will climb on 93 tune with only stoich changed.

[CCS86]

I just want to point out the inferred MAP is just as important as the inferred rail pressure. If your ECU is thinking there is 5 psi more pressure in the manifold than there actually is, that could end up in your MAF curve with out a known good curve to start from. You start ending up over 2.0 load at less than 15 pounds of boost, your engine is just not likely to be that efficient.

Absolutely. I think people downplay the importance of the Speed Density tables, which seem to govern the Calculated MAP. I'm shocked that even Roush supercharged tunes use stock GT Speed Density tables. That makes me question the validity of their MAF curves honestly.

Here is what my actual MAP curve looks like compared to Calculated MAP with stock speed density tables and tuned speed density tables. This had a definite impact on fueling, so it appears that Calculated MAP is used directly in the PW calculation. This means there is no reason to "lie" in the MAF table to combat boost related changes to effective injector pressure.

Calc MAP Tuning2.png

Correct, they are important. When I was doing first supercharged car I had to log rail pressure with external gauge because I had problem to setup good maf curve, sometimes I was getting lean spots, sometimes rich and missfire. Injectors and inferred line pressure must be correct.

I would really love to see some solid data on rail pressure during mixed conditions; along with some background data on the engine combination for context.

Yes. I have wired in analog pressure sensors for MAP and rail pressure on different forms of boost and experimented quite a bit with it.

If you run one to the point if dropping fuel pressure on E85 things start becoming noticeable. Depending on boost pressure boosted with just a voltage booster will start dropping fuel pressure and fuel trims will climb on 93 tune with only stoich changed.

Can you share any of your observations? Logs would be awesome to look at. Even just minimum rail pressure during WOT, for a given boost pressure would be insightful.

[murfie]

The SD formula is MAP=A*airmass^2+B*airmass+C.

You can get airmass by multiplying your air load by engine size. 0.00167 for coyotes. This gives you the airmass used in the formula, and the airmass that is giving the correct amount of fuel fo rthe target lambda. using excel you can scatter plot actual MAP vs air mass. add a polynomial 2 regression line. Display the equation and R^2 on the plot and that gives you the QUAD(A), SLOPE(B), OFFSET(C) of the best fit line. When blow through is disabled, it keeps it simple.

Id be more interested in seeing your MAP and CALC map plotted as Y and load instead of RPM as X. RPM should be a near constant or data filtered to be, while MAP and airmass vary. It would come out like a scatter plot A vs B. It would show how far off your SD really is, comparing the regression lines.

[CCS86]

Very cool!

Unfortunately, I don't have a digital MAP sensor, just a mechanical gauge. I started by tuning WOT, video recording my gauge cluster to associate RPM with MAP. I used that curve to make corrections to the SD tables and was able to get it closely matched.

Since I had success with that I moved on to the other MPs. I made tunes to lock my into each of my used MPs. Then I recorded a specific drive for each MP. With my mechanical MAP gauge and nGauge in view, I could manually plot RPM, MAP, and air load together. I picked RPMs for every other column in the SD tables to reduce my workload. Then I would hold that RPM steady by riding the brakes (thank you brake cooling ducts), as I rolled into the pedal and paused at each MAP row value. I took that hand collected data and interpolated/extrapolated to fill out the table, then ran it through the HPT SD calculator. I had to pass the data through a few times, reasserting portions of my data it was changing too much. The curve fitting they do in the calculator is annoying. This took quite a while, and built up a big spreadsheet to track results.

I'm sure it is not perfect, but it is surely so much better than the stock GT values, with my PD blower. The changes to the tables were very significant. Every time I look at the computer to compare against my mechanical gauge, it is within 1 or 2 in-hg now.

[mudar]

Building MAF curve form scratch takes 2 hours tops.

Can you give me the basic idea of it.

[veeefour]

For example take a look at the data canigga posted, horrible MAF curve TBH.

You wan to cover the entire range and you want to hit MAF period values that you are actually making: You need to avoid huge leaps and holes in your curve - best to increment gradually.

Not good:
Clipboard01.jpg

Better:
Clipboard02.jpg

Sometimes it's actually better to change the axis/period value to keep everything smooth.

[veeefour]

And this is a good example of beautiful MAF curve, it's what PMAS provides for their Cobra Jet CAI. Look how smooth it is.

Clipboard03.jpg

[caniggia]

Yes, roush curve isn't smooth. Idea is to make fuel trims as close to zero plus curve should be smooth.

[mstang_man]

Caniggia is the Roush stage 1 or Stage 2? Stage 1 suppose to be plug and play no tuning required per Roush instructions, stage 2 suppose to require calibration. so is the MAF Posted here Stage 1?

[caniggia]

Caniggia is the Roush stage 1 or Stage 2? Stage 1 suppose to be plug and play no tuning required per Roush instructions, stage 2 suppose to require calibration. so is the MAF Posted here Stage 1?

I described it there. Any kits with supercharger requires calibration, no other way. I think I posted there from stage 1 and stage 2