Transaxle

That is true again. The best way is to implement the MAF handling directly in the DME. It is a very bad way to use any adjustable piggy back solutions to fool the DME to generate correct ignition, dwell and injection values. That is a never ending adjustment story with to many variables.

But we found that the 951 DME is to slow to do a real time FFT as we do it in the MAF kit to smooth the MAF signal, that reacts very sensitive to the pulsing air stream . It works fine with the S2 and 964 DME but with the 951 DME and its 132 theeths, there is not enough CPU time to do it properly. And I belive that anything less is not good enough.

I do not think that it is a wrong way to build a MAF based AFM if you want to stay with the DME/KLR system. All the calculation is done by a resonable fast CPU and the new sensor is a very precise instrument that delivers accurate m3/h values. The DME can still work in the way it is was designed for and you can easily adjust ignition, dwell and injection directly in it.

fast951

951 DME too slow? It is slower than late style DMEs, but it's fast enough to handle all the tasks at hand.

Skip Wolfe

Well the truck is paid for by a car allowance from work so I get a free pass on that one, and the other stuff - well I pay the bills, so I am the budget committee, and if get it into the garage without her seeing it I am good to go. I have so much car crap in the garage she has no idea what is new and what is old.

Problem is that even though I am the sole member of the budget committee I still having a tough time getting this one past the committee.

Look who's talking - how many cars do you have now for only two drivers?

BTW - you up for meeting Mike maybe at Quaker Steak if the weather clears to check out his setup?

Skip Wolfe

Excellent thread guys. Thanks for all the info.

jimbo1111

Excuse my ignorance TT but I still have a mental block that has to be appeased. No disrespect intended. If you are running a 700 hp maf on a N/A with 72 lb injectors you can't possible be very accurate at getting the correct timing. If your maf is putting out a 3. volt signal at the max hp production of the motor. How is it possible to get the correct timing at that voltage? If you are using a maf that is sized to a specific output. That is believable. But with a 944 t that has different levels of tune. 200 rwhp on one and 600rwhp on the other. How is a one for all maf going to handle the task? If I have your 600 hp output model and I intend on just running 300 hp 95% of the time I just don't see how the timing will be perfect for 300 hp to 600 hp and anything in between. Can you elaborate?

944Fest (aka Dan P)

I'd be up for a meeting, but I won't have the Turbo. She's put away till March.
I know exactly what you mean about the committee.. Remeber, it is easier to ask forgiveness than permission!
My trick for the extra cars and the wife is to put her in one of them for several DE's. All of a sudden, I need track support x2 and the track expenses are doubled without the bat of an eye! (Just like Rhen with his pair of 951s!)
I feel bad having this personal chit chat deep in the middle of this great technical discussion, then again, it is your thread! Thanks guys, I'm learning a lot!

Mike B

John, is the smt6 required with your MAF? Thanks.

fast951

SMT6 is not needed for MAF. I recommend it to provide minor fine tuning.

fast951

Jimbo, we know the MAF in use, we know how to translate the MAF signal to actual flow/load value, the translation is what is called the transfer function. The load value is what's important. Load value is used for downstream calculations.
We can take the MAF you have now, get a larger or smaller (calibration wise), change the software to use a new transfer function, and you AFR, Timing etc.. will be the same as what you have now.
Timing same as fuel, it's calculated based on many variables some of which load and rpm. I'm sure TT will jump in here soon and will add if needed.

Mike B

Thanks John...As I already have your SMT6 I was thinking I might be able to offset the cost of the MAF.

I appreciate the unbelievably quick reply.

TT

You're overengineering the problem. The Vitesse MAF kit does no signal processing, the signal from the Hitachi sensor is dead on. On the 968, Bosch had a noise problem with the hot wire MAF, used a low pass filter on the front end and then oversampled the A/D input to get rid of what noise was left. They did no extra signal processing on the MAF input after that. The HFM5 already incorporates reverse flow detection and suppression in the resulting output signal so you should not be seeing much pulsation with that meter.

I've tested with the Bosch hot wire MAF and the only problem I had with it was the small dynamic range and the noise. The Hitachi MAF sensors have been wonderful and the latest style we use with the Vitesse MAF kit is the smoothest MAF I have ever used with the 944/951. You can go full throttle at low rpms and it will not pulsate.

Bri Bro

This has to be one of the more interesting and informative technical treads I have enjoyed reading in a long time.

Thanks for all the informative and polite input from TT, fast951 and Transaxle. There is always more then one solution to a complex problem.

TT

You're excused.

You are thinking in terms of voltage instead of air flow.

MAF 1
500kG/Hr airflow = 5V output

MAF 2
500kG/Hr airflow = 3V output

a) The DME is coded using the transfer function for MAF 1. When the MAF input is 5V then the DME will use the value of 500kG/Hr when doing the load calculations.

b) The DME is coded using the transfer function for MAF 2. When the MAF input is 3V then the DME will use the value of 500kG/Hr when doing the load calculations.

In each case the resulting load value will be the same and the same timing and fuel will be used for both.

Tom M'Guinn

Does the reverse flow detection eliminate the stumble that my old Huntley MAF gets when it returns to idle after boosting? Huntley used to swear that was not the MAF's fault but my datalog and a long intake snorkel say otherwise.

If you are using only a portion of the MAF's output range, do you have enough voltage range in use to produce a smooth throttle response at light load? I had an old "stage 4" huntley MAF that acted like a light switch in stop and go traffic.

Random thought: If the programming code resides on the chip, it's curious to me why porsche allows aftermarket companies to sell chips -- assuming those chips contain a copy of Porsche's (copyrighted) code with different data maps?

Bri Bro

TT
How much under the rated capacity of the MAF can you go before you get into a problems with resolution of the MAF voltage vs the timing/fuel? If you never go below 2 volts with MAF 2 above, wouldn't you see issues with idle and mid range performance?