sjsj

* IDEA **
Provide tunable throttle response for any Porsche 928 using:

- L-let with AFM: swap out flapper based spring loaded AFM with free flowing MAF

- LH 2.2 with MAF, or LH 2.3 with MAF: retune your MAF or complete replace with a different one altogether.

A background on MAFs: Better flowing MAFS to that provided with a 928 are available from Pro-Flow and Pro-M. Any arbitrary, inexpensive voltage based MAF from ebay (eg: Ford, JECS/Nissans & Subarus, Fords, or even the aftermarket ones that provide very clean airflow passage) can be used IN PLACE OF the existing AFM/MAF, as long as it supports the same airflow level or higher to the existing car (ie: aim for MAF output voltage to be 5V or less at maximum airflow).


** BENEFIT **
Crisper and tunable throttle response for via better flowing MAF. Way of inexpensively catering for modifications which alter intake, combustion or exhaust of the vehicle. Eg: headers, different plugs, forced induction, better sparkplugs, etc. Not confined to using circa 1990 technology MAF which might not be the best flowing, or worse yet a spring loaded AFM with a noticable pressure drop across it's airflow input and output.

* COST **
1/ DFA + Handheld controller + time to build it:
[US/Canada]
http://www.jaycarelectronics.com/pro...Max=&SUBCATID=
$51US
http://www.jaycarelectronics.com/pro...Max=&SUBCATID=
$38US

[AUS/NZ]
http://www.jaycar.com.au/productView...Max=&SUBCATID=
$80AUS
http://www.jaycar.com.au/productView...Max=&SUBCATID=
$60AUS

2/ An Ebay voltage based MAF with maximum airflow that supports the car. I aim for MAFs supporting engines with equivalent cyclinders and HP. MAFs runs 0->5V. Can be had for as little as 99c. Avoid GM frequency-based MAFs such as on the LS1.

3/ Technical details:
The DFA can be considered a general purpose voltage remapping device. It works within 0-1V (so can modify an O2 sensor signal), 0->5V and 0->12V range as per article http://autospeed.drive.com.au/cms/A_2418/article.html . From my experience it can also be set to intermediate ranges such as 0->6V by changing the adjusting the onboard adjustable potentiometers.

So then, the idea is to remap the signal from the new MAF to closely follow that provided by the original device. The DFA is able to provide fine adjustment along 128 points from 0->5V before the signal enters the ECU to as best approximate that volume of air from a spring loaded AFM (or existing MAF).

The only thing the DFA doesn't do is invert the signal if you are for instance mapping 0-5V to 5-0V, as is the case for my example when changing from a AFM to a MAF, yet the ECU still requires an AFM type signal. I'm playing with my DFA to do this via an opamp and a couple of resistors. (i'm also having a go at simply hacking the DFA devices onboard 2xdual opamps to do this).

*** How to tune? ****
1/ You may wish to map out the RPM to o2 signal voltage at every 500rpms, or even every 250 rpms. This will serve as a map later when you use the DFA to simulate this signal from a new MAF.

2/ Then connect the new MAF in series with the existing MAF to gather it's output voltage at different RPMs and correct it via the DFA device. Or if it cannot be connected in series, remove existing MAF and use a potentiometer to simulate the input voltage into the ECU at corresponding RPM.

Connect the new MAF via the DFA, monitor it's voltage output at the corresponding RPM values. Adjust the DFA with handheld controller until you match the factor specs from the table generated in (1). Easy!

If the previous MAF or AFM also had a temperature sensor output, ensure the new MAF has one too, or obtain an equivalent thermistor from your electronics shop to wire into the corresponding ECU input wires (probably requires a little research to ensure the new temp sensor is the same as the original. The temp sensor input allows the ECU to enrichen the mixture when the air is denser and cooler).

3/ Now retune outside of the factory specs. Unplug the O2 sensor to go open loop. Wideband sensors would be the one of the more accurate ways of tuning. However,the narrowband sensors I've evaluated are accurate to within 5% of wideband sensors from 12.0:1 right up to 16.0:1, as long as the engine is at normal running temperature. Meaning a simple digital voltmeter and the chart is http://wbo2.com/sw/20-tables.xls with a stock narrowband sensor is sufficient to monitor the AFR.

Tuning is not my speciality. All I can say is refer to internet docs on this. From what I've read, tune for rich in the upper RPMS (eg: 12.0:1, and for stock to lean off the mark , eg: 14.7:1)


** Comments **
Probably a weekend project because of the details required. But I consider it to be quite a lot of fun and certainly helps one understand more of how the car works. Best part is the end result means gone are the limitations of crappy AFMs/MAFs from yesteryear.

Always wondered how come we can upgrade computers, yet not the ECU and brains in our cars. We just buy new ones? Shame, coz the 928 is look-wise, my favorite and WAY ahead of it's time. Truly a rocketship with very sexy lines.

Inspired by the discussion as per thread at https://rennlist.com/forums/showthre...=238574&page=5
Many thank yous to everyone for the technical questions raised.

sjsj

drnick

what about combining the signals from 2x MAF units if you were using a dual throttle body intake?

sjsj

If using two signals, you may consider using adding the two signals, each with a gain of 0.5 (ie: half the signal).

So consider taking the average of the addition of both signals. This can be performed using the opamp circuit at http://homepages.which.net/~paul.hil...der/Adder.html

Consider:

Rf=470k [high impedance so it doesn't load down the circuit]
V1=MAF1 input [say up to 6V]
Rf/M1=470/0.5=940k [ie: gain=0.5, we half the input voltage]
V2=MAF2 input [say up to 6V]
Rf/M2=470/0.5=940k [ie: gain=0.5, we half the input voltage]
Va=GND
Rf/Ma=470/1=470k.
S=1+Ma+Mb+Mc-[M1+M2]
S=1+1-[0.5+0.5]=1
Rf/S=470/1=470k.

Vout=(V1.M1+V2.M2)-(Va.Ma)=[V1*0.5+V2*0.5-(0*1)]

That is Vout is then (MAF1+MAF2)/2. The average voltage of their combination which is what you want.

If there are any electronic engineers on this board, feel free to check my calcs and present critique to your hearts content. This is where I'm at with my DFA project to invert the signal. Just trying to get some time to do it. I am simply using an opamp cheatsheet linked above.

NOTE: The supply rails to the opamp need to be greater than the voltage output. So good idea to use a 7808 voltage regulator to supply +8V from a +12V battery supply to the opamp.

If anyone here does purchase the Jaycar DFA, the schematic is included with the kit. It gives a really nice explanation of the supply to the opamp. Bugger, wish I could scan the schematic and put it somewhere. Anyone volunteer to hostimage it if I can do that?

sjsj

MrLexse

SJSJ - Welcome to the board. New ideas are always appreciated and their value (or lack thereof) will be borne out by time. Be advised, however, that there some very astute (IMHO) members out there who will question everything. This serves to enhance the level of discussion, a benefit for everyone.

sjsj

Thank you. You mean there are lots of devil's advocates who worship a horned goat like this ? That's OK. Being challenged simply raises the levels higher and higher. Got an extinguisher at hand in case the horned goat flamess my butt..

One slight correction given to step 1.

1/ You may wish to map out the RPM to !!AFM/MAF!! signal voltage at every 500rpms, or even every 250 rpms. This will serve as a map later when you use the DFA to simulate this signal from a new MAF

sjsj

Herr-Kuhn

You could do the MAF conversion for about $800.00 The largest benefit is the increase in throttle response. I don't believe you gain all that much true power. I've done 2 conversions on L-jet cars...there are some tricks.

toofast928

sjsj,
The AFM voltage sweep is 3.2 V (idle speed) to 7.5 v (full throttle). Typical MAF sensor is .8v to 5V. So why would you have to invert the signal? MAF voltage just needs to be boosted.
The trick is programming the MAF to signal 5 volts around 3000 RPM. That's were the AFM maxes it's full sweep. After 3000 RPM the analog ECM adjust fuel pulse by RPM.
Is the DFA voltage adjustable in steps vs, RPM? If so then it's the ticket!

sjsj

On one car the AFM had a signal from 5V with no airflow down to 0V at maximum airflow. A MAF does the reverse, 0V at no airflow up to 5V at maximum airlow. So before the signal goes into the DFA it needs to be inverted. Why? Because the DFA provided +50% and -50% correction, so by it's very nature it cannot turn a 5V signal into 0V, however it can change it to be between 2.5V and 7.5V at maximum load. So I needed to invert the MAF signal to be betweeen 0 to 5V prior to allowing the DFA to provide load point adjustments.

The DFA provides a 128x128 grid within which to work. Eg: 0 to 5V of a MAF will map out to points 0..127. When the MAF is at 2.5V it is point 63. At that point you can issue a delta value of upto -127 or +128 which added or subtracts voltage.

So the way you'd do it for 3000rpm is identify the MAF voltage at that point and add/subtract. If it is 5V then you'd be in the 128 value range if optioning to use the 0 to 5V range.


How could you modify fuelling if the AFM signal is not relevant at greater than 3000rpm?

There is another Jaycar device that you may be interested in. It's a Digiital Pulse Adjuster (DPA). Using the same handheld controller as for the DFA, it allows the the duty cycle (eg: 0->100%) to have delta correction values added or subtracted again along 128 load poits. (not sure on the limits of the correction though, coz never used one). Details at http://autospeed.drive.com.au/cms/A_2471/article.html . This is directly modifying the fuel injector duty cycles!! Presumably at say 4000rpm, the fuel injector duty cycle may be 65%. So at the load point that corresponds you could simply add duty cycle such as 10% to richen the mix or say subtract 10% to lean out the mix.

Available at Jaycar once again:
[US/Canada]
http://www.jaycarelectronics.com/pro...Max=&SUBCATID=
$51US

http://www.jaycarelectronics.com/pro...Max=&SUBCATID=
$38US

[AUS/NZ]
http://www.jaycar.com.au/productView...Max=&SUBCATID=
$80AUD

http://www.jaycar.com.au/productView...Max=&SUBCATID=
$60AUD

* Further reading perhaps? *

Well worth getting is Silicon Chips's "Performance Electronics For Cars" , reviewed at http://autospeed.drive.com.au/cms/A_2457/article.html . It explains the basics of car electronics then gives and explains a stack of electronic circuits kits to modify the car operations to bring up to your needs and the 21st century. Jaycar stock the kits mentioned in the book and the book iitself. This is worth it's weight in gold IMHO, but what makes it extra nice is its pretty inexpensive. See below:

[US/Canada]
http://www.jaycarelectronics.com/pro...Max=&SUBCATID=
$15US

[AUS/NZ]
http://www.jaycar.com.au/productView...Max=&SUBCATID=
$20AUS

sjsj

Jack '84 928s

Its just a bandaid and not worth it, why spend 800$ for a halfway fix when you can spend around 1500$ for a full engine managment system. Do you even own a 928?

sublimate

Jack, he's saying it could be done for less than $100, which if true would be nice.

Herr-Kuhn

Jack, it is worth it. My car is MAF converted and makes over 400 HP with such a system. The best part is you don't have to rewire the whole injection system or swap the sensors, etc. It is a very viable solution. Of course more can be done on stand alone, but it costs a lot more $ and takes time. Your $1500 example doesn't begin to touch all of the extra things you would have to do to make it all work...the wiring time would be huge.

sublimate

Based on what you say it doesn't sound like it'd work.
A maximum of +50% correction would adjust the .8v of the MAF at idle up to only 1.2v, which is nowhere near the 3.2v that the L-Jet is expecting to see from the AFM at idle.

sjsj

Correct, that is why it is important to know the transfer function of the AFM and the MAF you are adapting. In my experience

- Vane flapper AFMs provide a logarithmic function starting at 5V for no airflow and tapering to 0V at maximum airfow.

- MAFs provide an exponential function starting at 0V for no aifrlow and rising to 5V at maximum airflow.

So they are virtually the complete opposite. How to correct this? If you wish to substitute a MAF in place of an AFM, it's signal will need to be inverted prior to processing by the DFA. Ie: have two inputs into a opamp. One being the MAF input signal. The other being a +5V.

Apply a gain of -1 to say the MAF signal, then add 5V to the signal. Cheatsheet of circuit to do this is at http://homepages.which.net/~paul.hil...der/Adder.html . Also a voltage divider could be used if needing to scale the signal as well.

So you'd see the inversion works as per the table below:
V_MAF | V_INVERTED_MAF [ 5V-V_MAF ]
5 | 0
4 | 1
3 | 2
2 | 3
1 | 4
0 | 5

Just ensure the rails to the opamp are higher by using say a 7808 voltage regulator to give a steady 8V signal from the car's 12V.

sjsj

sjsj

For those interested, the attached pic is a Porsche 944 Bosch vane flapper AFM from http://frwilk.com/944dme/afm.htm . The same website explains in detail the components inside it and how it works.

toofast928

sjsj,
Cool response I been looking to do this for some time. I'll look into what Jaycar has to offer.

I haven't delt with every fuel system out their but any Bosch, Siemens, Ford, Delphi, MAF works on a 5V voltage supply. ECM signal voltage at idle speed .8 mv. Near WOT 4.8 volts. No need to invert the signal voltage.

Anyway for L Injection all that's needed is to boost the signal voltage to roughly 2.5 volts across the engine demand sweep (not RPM). Maxing out at 3000 RPM.

Mapping out the AFM signal voltage per 500 RPM increments and match it to the conditioned MAF signal is a base. After that we need a dyno to fine tune.