Rogue_Ant

+1
You are correct about the car running lean. Without chaning the transfer function values, the 944DME will not understand the 951 AFM.


-Rogue

Dare

Actually, I’ve been thinking about this a bit and I find it hard to believe that the output of the AFM follows a perfect logarithmic curve with respect to air flow. While the overall shape is clearly logarithmic, my guess is that various physical effects tend to steepen/flatten the curve at different places (likely at the extreme ends). If this is true, it means that the output of the AFM code inherently contains a certain amount of algorithmic error relative to the true air flow, because the algorithm is incapable of representing the deviations from a perfect curve. Therefore it seems reasonable to assume that the DME engineers must have compensated for this error elsewhere in the system.

Now a 20g/s error at the bottom of the curve may be too much to compensate for. I don’t know. It would be interesting to know the typical idle air flow because maybe parts of the curve are below the normal engine RPM range. Also it may be possible to make the lower end more accurate at the expense of the upper end, if compensating for errors up top is easier. Another option, of course, would be to try a different MAF, perhaps one with a range closer to that of the stock AFM.

Agreed.

V2Rocket

is there any way that timing can be adjusted with the fqs scaling like it is stock?

like i want a map for e85 and a map for gas. obviously the gas map would be with less timing than the e85, but as far as i can see i can only tell the computer to pull fuel when running gas. would i have to kind of settle for a lower timing advance?

Rogue_Ant

The FQS scaling is stock in the .bin I posted. As such, the timing changes are in effect as well (depending on FQS position).
I could add another constant for FQS timing adjustment... Let me double-check this, and I'll update the bundle.


-Rogue

V2Rocket

thatd be marvelous thanks

want a little more timing adjustability than 2.8 degrees when switching from corn juice to gas

mikey_audiogeek

Hi y'all,

Using Rogue's spreadsheet and data for the 951 AFM, I also mapped an equal-percentage (logarithmic) characteristic to the output and found it matched to within +/-2% from 0 to 5 volts. The formula is flow=2.65^(volts+1).

I wonder if the Bosch engineers were focussed on achieve an equal percentage characteristic for the AFM, and incorporated physical geometry to achieve this characteristic throughout its working range? If you look closely at an AFM, you can see some subtle contouring of the airflow channel that suggests this may be the case.

One of the advantages of an equal percentage element is that frictional/mechanical/hysteretal errors create an equal percentage error at all positions. Linear elements tend to have greater error at small values.

Industrial control valves can commonly display equal percentage behaviour consistently over a flow ratio of 1:100. About what we're looking at here.

Edman951

Hey Rogue,nice work you did on this.

Got a few questions.
1-The max rpm on most of the maps is 6240.
If my rev limit it set at 6600rmp, will the adjustment set at the 6240 rpm zone apply above it?
2- Can we adjust the oveboost above 20.85psi?
I cant seem to get it over that number in the program.

Also, any sugjestions for a MAP setup using this?
I'm still runing the Guru racing AFMLink.
Should set the afmlink values at 0, and tune the chip instead?

Thanks in advanced

Rogue_Ant

1, The 6240 rpm point will effect revs above it.
2, the overboost numbers are just ballpark... Setting it max value, basically eliminates overboost. So if you want to run more then ~21psi, simply max the overboost map.

How much hp and what size injectors? The stock transfer function is limited to ~330rwhp, safely. If you are running more power than this, you will need to do a global adjustment on your PB, to bring reported airflow down...
Otherwise, I would try to keep most adjustments 'on-chip'. Remember to flatten the AIT and baro compensations; they will help keep your tune from drifting with the weather.


-Rogue

Edman951

Do you mean adjust my PB to correctly tune my AFR since im over the factory range?
Yes i did set the AIT and baro at 0, Thanks for the heads up.

I am running 72lb inj on a GT30R turbo with the AFMLink map sensor kit.
Being running it for many years.
Last times i drove it above 20 psi 3yrs ago, it was hitting the overboost that was set on my chip at 20.71psi. So if the max setting is 20.85, then i will hit it again.

Havent driving my 951 in almost 3 years. because i was playing with my other cars, and I will get it back on the road next summer.

Rogue_Ant

The AFM was very accurate...


-Rogue

Rogue_Ant

Setting overboost to the max eliminates it... So don't worry so much about the actual number.

I would do a global scale with your PB to get the AFR close, then fine-tune via TunerPro.


-Rogue

blown 944

one thing to remember when using the afmlink is that the map sensor will only register to ~21 psi. It then goes back to the start which means it leans out and completely cuts out. The only way I was able to get around this was to fool the sensor by relieving some pressure with a grainger valve.

I have tried a 3 bar sensor and it would not run worth a darn. Maybe you can get it to do better, I gave up on it. I did get a 2 bar to work pretty well, however it would max at 17 psi requiring more relief.

Just making sure that you don't waste time when there may be something else causing the problem.

Dare

Actually, I wasn't referring to the accuracy of the AFM. Rather I was commenting on how likely it is that the output of the AFM follows a perfect logarithmic curve. I'm sure they put a lot of work into making it accurate. But I really wonder whether the output of the AFM matches a curve described by y=a*b^x to 5 digits behind the decimal.

Again, this is all just "recreational engineering" on my part.

Rogue_Ant

Hmm, well fluid dynamics is not my specialty...
I would think that even if the AFM did follow that basic function, simple manufacturing tolerances would dictate some inaccuracy. I would assume that number would be within a percent or two.


-Rogue

Edman951

Thanks Sid,
Now i remember, the limitation was the sensor,
Like i said havent played with the car in 3 years, so forgot some of the special thing my setup has.
I will limit to 20 max then, anyways it will save the motor alittle LOL
Cant wait for the summer...