The key to the success or failure in replacing the stock AFM with any MAF or HFM sensor, is the quality of the specially written chip software, and integrated electronics of the system. Just wiring any MAF or HFM sensor into the Bosch AFM wiring will result in immediate disappointment, as the output voltage and flow curves in no way resemble the stock flow meter.

The system is a product of two years of research and development. A specially developed analog to digital circuit was designed to convert the HFM5 signal into one the Bosch DME can process. The circuitry is contained in a separate aluminum enclosure not pictured that resides inside the airbox housing. In addition a separate barometric sensor was implemented to compensate for adjustments due to changes in altitude and atmospheric pressure. Also important is the implementation of an intake air temp sensor which is used by the software to ensure maximum performance in cool ambient temperatures, yet variably trim back the ignition tables as temperatures rise, providing anti knock protection in hot weather, a key function that was missing in other MAF systems.

The specially developed and programmed chip that is required is included in the system. The stock chip, or any other performance chip will not work with the setup. The system performs flawlessly, cold start, idle, decal, part throttle, and full throttle performance is fantastic.

We intend on having detailed dyno results available soon. Cars will be dyno tuned and optimized for the maximum performance attainable with just a chip alone while retaining the stock AFM sensor. As Ralph indicated, his will be one of them. We will then swap out the AFM sensor with the HFM system and redyno. The results will ultimately determine how restrictive the stock AFM is, and how much more power is attainable by eliminating it. Eventually a 964 system will also be available, as the 964s use the exact AFM sensor as the 3.2s. Stay tuned.

"A specially developed analog to digital circuit was designed to convert the HFM5 signal into one the Bosch DME can process." - SteveW -

What??????? The original 911 3.2 DME already has an A/D converter for the AFM
and the other analog inputs, e.g. temp. Additionally, it's a simple process to
to modify the fuel maps for the MAF versus a AFM.

Check here ( www.systemsc.com/tests.htm ) for the output equations & voltages
for both the 3.2 AFM and the 996 MAF which are both linear. So, an easy soulution
is to use a simple 996 MAF unit ($250) and modify the fuel maps, if one wants to
waste time & money.

"Also important is the implementation of a temp sensor which is used by the software to ensure maximum performance in cool ambient temperatures, yet variably trim back the ignition tables as temperatures rise, providing anti knock protection in hot weather, " - SteveW -

Sounds like more marketing hype.

The temp sensor changes very very little once the 911 3.2 reaches normal temp.
That's why Porsche and others didn't waste efforts developing a knock system
based on using a temp sensor. Besides it wastes performance for load conditions
where pinging doesn't/won't occur. Remember, it's the combustion chamber temp,
engine load, fuel octane, chamber design, and spark location which all affect pinging
and not just the CHT. That's why knock sensor systems provide the best solution.

Bottom line: Very little to no performance increase results from the switch to a MAF
without engine mods, e.g. engine size. Porsche used the MAF because it provided
more precise air measurements for emissions and no need to use an altitude
sensor. Before and after dyno results have been posted on Rennlist indicating
insignificant performance increses for stock engines, thus indicating a waste of
money. For those SoCal ones that still want to make the change, Valley European
Porsche ( www.valleyeuropean.com ) can do the change very easily.

Just what the forum doesn't need, B.S.

Hate to dredge up "old news", but what happened to Steve's remarks above...

"We intend on having detailed dyno results available soon. Cars will be dyno tuned and optimized for the maximum performance attainable with just a chip alone while retaining the stock AFM sensor."

I see the MAF on Steve's site but only vague references to power increases.

Do we assume that the echoing silence about dyno results means that the MAF doesn't give much of an edge over AFM?

Steve - tell us it's not so....

I'm a technician (BMW). My two cents on the mass air conversions are: First go to pro-flow.com they have and offer mass-from-air meter conversions for early bosch motronic systems; as in BMW & Porsche for under $500.00 (last time I checked)
They are a supplier to many shops who in turn re-package as a "kit" with a healthy $$$ mark-up. One well known shop is ...........Autothority in Va. If you check pro-flows website carefully you'll find a list of their 'distributors'.

Regardless of the kit or manufacturer of the mass conversion you will still have to have it (the DME chip) dialed in on a dyno to obtain the best performance possible. Most guys with upgraded cams, induction, displacement should do this anyway even with the stock air meter.

As far as HP increase you'd most likely find 10-12 at the flywheel for a stock or near stock 3.2 -as displacement and mods increase (like Jeremy's) you'd probably see more gains in HP and TQ because the OE set-up is designed for bone stock every day conditions.

The REAL DIFFERENCE is in driveability. VERY responsive with a real increase in seat-of-the-pants feel performance. But Steve's kit @$2700?? I dunno.............. a little on the pricey side considering you'd still have to have it fine tuned. A kit under $1000.?? Now I'd say that is good value.

SteveW,

While you're at it, why not include some adaptive learning and on board diagnostics so we could plug in and see what's going on while we drive?

I had all that when I converted my 78 3.0 to a Chevy TPI 2.8 injection system complete with hot film MAF and OBD that I could access with my scan tool.

Joe

The expensive part is the carbon fiber box.

I already asked him how to do it on the cheap, being that I'm an "insider". But there really isn't a cheap way. Building my own box and buying the rest was still going to be fairly expensive. But the priciest part is that box, which is super light, and very well made. A big part of the reason I wanted one was the weight savings. The stock flapper is HEAVY!

Like anything else, we try to undercut someone elses leg-work, only to find we spent ridiculous hours, lots of running around, and in the end, it costs almost as much.

Nowadays, I just either save my pennies and buy it, or I don't bother.

My only interest was to do away with the mechanical failings of the AFM and it's wiper-arm potentiometer which seems to be a source of many rough running problems for these 3.2's. The MAF setup seems to be cleaner, and would likely last longer. Any additional HP or throttle response would be bonus to me, not the goal.

Hi guys - just very busy on this end and haven't had much of a chance to do full dyno comparisons on the HFM system. It does produce a stronger mid and top end, along with a crisper more responsive response, but thats just IMHO. I promise it will come soon though.

We spent some time doing several dyno runs on Ralph's twin plug 3.5 motor with the system, but his motor is not stock, and there really isn't a 'stock' dyno run available as a basis of comparison, as I'd prefer to present the data as on a 3.2. The day we were dynoing the 3.5, the car was down 27 rwhp from the runs a year ago, which I suspected to be a non firing coil. Subsequently we found the Andial twin plug splitter defective, only powering up one of the coils, which with the reduced timing curves used on the twin plug program, was effectively like running the motor with 7-9 degrees reduced ignition timing. It was interesting that even in this state, the HFM system produced something like 10-12 more hp than I could extract from the stock flow meter with the chip.

It should be noted that it is not possible to take a standard MAF sensor, such as the one from pro-flow and just adapt it to the Motronic system, without completely reworking the entire program. It is not just a tweak of the fuel and ignition maps, but to get it to idle, start, warm up, and function and properly, the many trim variables need to be changed, reworked, or zeroed out. The AFM output voltage curve has a completely different output curve than all the MAF or HFM sensors. It would take many many hours of reverse engineering the software, and fuel remapping to get it to run acceptably, and even then, you wouldn't know if your part throttle ignition tables vs. load were coming in properly.

Our way around this was to design a microprocessor that interfaces in between the HFM5 sensor, and the AFM harness plug, so that effectively the output seen by the Motronic unit is the same as the AFM curve, and the entire system is transparent to the DME. One other advantage found is that with the stock AFM meter, the barn door bottoms out at around 5000 rpm, so the output voltage maxes out at Uv/Up of 4.3 volts. If your motor produced increased flow such as from engine modifications, it will will only deliver as much fuel as has been preprogrammed in the full throttle map of the chip. With the HFM5 setup, you no longer have this voltage limitation, so it can continue to measure air flow past the engine's capability, and the chip software will continue to meter to the rev limit. Also built into the HFM5 sensor is a temp sensor, which is used by the Motronic software to trim back overall ignition timing in high temperature ambient conditions to prevent predetonation in hot weather. Your stock flow meter also has one. One fault with the few aftermarket MAF systems that were on the market was the lack of support for the intake air temp sensor, which without reduces to eliminates the margin of protection from predetonation when it gets hot.

The HFM5 sensor used is very fast and accurate as we use the sensor from the BMW E46 M3 which is a 3.2 liter 333 hp ITB motor. Measuring voltage response from a Pro-M sensor vs. the HFM5 sensor is like night and day. The Pro-M sensor has a voltage output latency that can be seen as much as a half second. A more immediate sensor means more immediate fuel delivery and access to the proper load cells in your ignition and fuel table.

I have also found in the years of tuning many 911s, there is a large variability in the accuracy and output signal of the Bosch barn door AFM sensor. Internal spring tension variations/wear affect the load calculation of the ignition and fuel cells, wear on the resistor track causes voltage dropouts that affect drivability and emissions, and increased track resistance from wear causes the Motronic to meter fuel to the injectors leaner than originally intended. I've seen swapping in one flow meter out for another show 1 full AFR difference in the full throttle fuel curve. Renewing the stock flow meter by moving the internal wiper arm up or down the rotation axis helps a lot to restore original performance in such instances.

The highest expense of the system is the carbon fiber housing. It is fabricated by one of BMW's carbon fiber subcontractors. It was originally designed in the OEM profile to work with the factor air filter and airbox cover to 'fly' under strict TUV regulations and inspection in Germany, as just inserting a HFM5 MAF tube with a cone filter would never pass approval and inspection. A 'lite' version may be fabricated for the 'non-TUV' such a the U.S. in the future, that is probably in the distant future. If so, the system will also fit and operate on the 964 motors, which use the same AFM sensor as a 3.2, and show to increase engine output with improved intake flow here.

another way to save weight and reduce the intake restrictions is to rid of the AFM/MAF altogether and use a standalone ECU that does TPS/MAP. I am sure you can do it with an Autronic SMC for less than US$2700.

Good thread. Free bump. This loren guy always makes me laugh.

I'm almost afraid to ask, is it Calif. smog legal? Do you plan to get a CARB number?

I dissagree. Map/tps or what is commonly referred to as "speed-density" systems are still an approximation of fuel needed. There is direct way to monitor volume. Those sytems are the cheapest and simplest to do .....provided you have a fuel/ignition map that is close to start with for your app.

As tech by trade MO is nothing beats exact air flow meaurement into the engine as a mass-air unit. As you increase compression, cam timing and even displacement mass-air is the only way to go to achieve the maximum output from the engine..

Years back a lot of the detroit pony cars were OE speed-density systems. As long as the engine remained stock all was, well -just ok performance. Most people wanted more tuning ability. IE: inatke, exhaust, cam mods.

There were several companies that offered mass-air conversion kits retaining the stock DME. Ford motorsports was one of the first manufacturers that offered a retro-fit kit.

Cars with MAFs need a whole lot of calibration as well. Hotwire sensors measure the flow of air over the sensor in _both_ directions which means modifications do not necessarily fall within the range of adaptation for the system.

Steve,
Have you seen the setup BMPD has for the same upgrade for the BMW? It also comes with a control box to mount in the cabin for adjustments on the fly.
I have seen a couple, and have seen where the money throwing at people, have gotten it all wrong.