AFM (+DME) modifications revisited... OR tuning a 944 for free!

[1.2gees]

OK, here are a few findings:

Firstoff, the avarage motronic DME (with sequential injection) has the capability to alter timing, AND fuel mixture, between the firing of one cylinder to the next! It has the ability to pick up on rpm change from about 2.5degrees of crank rotation. Timing/mixture are also changed on the fly, keeping emissions/fuel consumption in check at all times (there are exceptions, more on that later). My point? Duplicating a performance chip at some point, maybe. Upgrading (extensively anyway) the parameters in the eprom chip myself? No way.

The most power limiting single thing on a 944 NA (as well as most other motronic, specially early motronic/L-jetronic equipped cars) motor is the AFM. The spring tension on the barn door flap is THE single most restrictive part of the intake, diameter wise alone, the end of the intake (at the fender), is the second most restrictive part of the intake system. The mass of the flap itself/spring tension both delay the intial response, and open past the point required by the engine air flow through the intake causing a mixture overly rich(very slightly, but I'm discussing every aspect of this I can think of). The former is the bigger problem of the two by far however.

My first idea of upgrading this system without spending any $ was to back off the spring tension off the AFM, and keep it at that. However there's a couple of problems with this, The velocity of the moving sensor flap(after the throttle is suddenly opened), will open the flap even more vs. extra movement at stock tension. Also, another initial downside, perpahs more important, looked like the flap would be open all the way, before it would with the stock setting on the spring. Which would make the car run much richer than required, since the AFM door movement would not be linear to the air flow, as it was with the stock setting.

Still with me?

------------------
It's all the driver...

[1.2gees]

So I was ready to dismiss the idea of playing with the tension of the AFM sensor, as I had with the idea of programming the EPROM chips myself.

However, I realized that the AFM sensor is fully opened, at rpm above 1700-2200 or so, during full throttle. So the input is most usefull for part throttle applications. But more importantly, the mixture is determined by FIRST the rpm signal from the flywheel position/speed sensors, and then a combination of other inputs. These are the AFM signal (which estimates engine load, through how much air is sucked in by the engine), engine temp sensor (giving a much richer mixture when cold, more on THAT later too ), throttle position sensor, etc. Actually, at full throttle, the AFM signal is ignored!!!

So keeping in mind the AFM signal is important for drivability, yet the spring tension hurts throttle response/max power, here's my new approach to modifying the AFM: drill holes on the AFM flap, backing off the tension also. This way, restriction is reduced, both through less tension on the flap, AND the flap itself having holes through it will let much more air move through it at lower engine loads too. Since the spring tension is much lower, the door will still be open. So far so good. But also less mass means, when the flap is on the move, it won't go past the point that is required by as much as it used to, since first there's less force moving it (arguable that the reduced spring tension vs. reduced air drag amount to anything apreciable) but also because it has less mass. Looks good on paper.

OK, more on the next post,

------------------
It's all the driver...

[1.2gees]

Now the second thing to modify for free/cheap on a 944 NA, for more power, that is the signal received from the engine temperature sensor.

The biggest reason emissions are worse with a cold engine (besides the fact that the cat isn't very effective), is because the engine is run a lot richer, with timing optimized towards more power as well. Ever notice how your car feels faster with a cold engine, part of the reason is this (the other part is lower intake temps, but I think the higher drag of a cold engine has a bigger effect than the low intake temps. However the power increase because of the modified air/mixture is even larger than that provided by the lower intake temperatures).

So in short, altering the signal from the engine temperature sensor, (making the DME think the engine is cold) will optimize both timing, and mixture for more power.

Keep in mind the most fuel efficient (besides producing theoretically the least amount of pollutants) air/fuel mixture, or the mixture for the most complete combustion, is also about 16% leaner than the theoretically ideal air/fuel mixture. This is on top of the extra cooling effect that running rich will cause! (I don't know about you, but by now I'm already thinking 'running rich' in the traditional sense is pretty much exactly what I want! Ofcourse not by too much, only about 10 to 15%)

So far the engine temperature sensor modification looks to be a very good thing to modify as well. Hope I don't need to mention it'll lower gas milage, and raise emissions! My idea is to put an adjustible resistance switch, with the ability to adjust for fuel economy/power as desired.

One thing however, if the signal is not within parameters, it will be ignored by the DME!

Also, the later cars (I believe starting with S2, but that's from memory) will start ignoring the signal, if say it looks like the engine hasn't warmed up in 600 hours of continious driving. They may even store a trouble code, but a trouble code for a broken engine temp sensor isn't exactly fatal.

I welcome comments, or ideas.

I'm surprised I could find the time to go through all the fuel injection/engine management books I recently purchased. I'll try to put some more time into this, and think it over/go through the comments before I actually proceed to perform either one of these mods on my own car.

PS: Finally got done with all the physics work I was missing from school, I've been up since 4:30 AM Sunday, my head is about to explode!
Ahmet

------------------
It's all the driver...

[This message has been edited by 1.2gees (edited 04-22-2001).]

[Ehab]

Sounds interesting Ahmet ... but I would think that making the holes in the flap door requires that the smaller force on the flap be counteracted by a larger motion due to the reduction in the stiffness in order to achieve the correct mixture.

I've been reading about throttle switch and MAF stuff to fix this annoying problem of the rpm overshooting the idle speed when the throttle closes when I slow down or come to a halt, and it seemed like the MAF tension would be the best thing to modify. However, that would change my whole mixture profile, so I dicided to let it be until I work out other problems in the car.

What I've found thought, was that on some Jaguars, some modification company actually does exactly what you were about to do ... loosen up the flap door spring tension to allow more air in ... However, they counter act for the reduced tension by modifying the DME to put out less gas at the same opening angle of the MAF ...

Other interesting things ... the throttle switch detects either idle, or full throttle, and at both extremes, the signal from the O2 sensor is ignored, and the signal from the engine temp sensor is 'tricked' to make the DME think that the engine is cold, thus richening the mixture ...

Do you know the small 5mm hex nut on the MAF .. I screwed that in all the way to get rid of that rpm overshoot problem, and I think it worked. Of course I had to compensate for the higher idle speed by closing the idle screw on the throttle body. Does that hex nut actually control the idle position of the flap door?

Good luck.

------------------
Lock Stock '84 944 and two smokin' tires ...

[Ehab]

So Ahmet, I have another question for you, if you unplug the temperature sensor from the engine while at operating temp ... is the idle speed supposed to go DOWN or UP?????

I thought that as the engine gets warmer, the resistance of the sensor is reduced, so unplugging the sensor is like making the DME think that the engine is really cold ... hence it should make it rich ... and the rpm should rise right?? But when I tried it, the rpm actually dropped!!!

[1.2gees]

Well, unplugging the temperature sensor will NOT make the DME think the engine is cold, this is why I need to make an adjustible switch varying the ressistance, instead of just an on/off switch. (As I noted above, if the values are NOT within reason, the signal will simply be ignored). You're right with the resistance going up, as the engine is cold, but there's a max resistance value set in the DME, I believe it's around 3700 ohms, anything larger, and signal is ignored. Ditto for the lower end of the scale, smaller than 200 or so (if I remember correctly) is ignored as well.

Now to the holes on the AFM door, I don't understand what you mean, here's my take on it: Holes=more air required to move the door the same amount, given spring tension is constant. Since more air can open the flap a given amount with the holes in it, you can back off the tension, and a given amount of air passing through will move the flap as much as the stock setting, but since air can flow through the flap, air has less resistance going THROUGH the AFM. This isn't a huge concern, as I also mentioned, the AFM door is open all the way at full throttle above around 2100rpm ANYWAY. This would only matter for part throttle, and to fine tune holes on the flap vs. spring tension, I have unlimited tries!

On a side note, I contemplated disconnecting the oxy. sensor which would also run the engine richer, but without modifying the timing. Also, the oxygen sensor output is NOT ignored at idle, nor full throttle, do you mean this is what the tuning company did with the DME? The 944 DME will ignore AFM reading at full throttle, not positive about idle, but I don't think so, it is considered in most motronic equipped cars to compensate for small vacuum leaks, engine wear, etc. (at idle)

I'm not sure I know which screw (on the AFM) you're talking about, if you're talking about the mixture adjustment screw, then you simply turned up the fuel.

PS: You shouldn't disconnect anything that sends/receives a signal from the DME, with the ignition on.

I've also noticed the idle drop!, the reason for this is because the 944 DME shuts off all fuel delivery, if the engine is turning slower compared to the transaxle, for better fuel economy. I believe this is a design flaw, all later motronic systems start delivering fuel, at a preset rpm above idle. I believe the 944 does it exactly at the idle rpm (around 950rpm), so it takes a while for the engine to react, and get back up to idle speed. I suspect this is also the same reason part throttle can be quite jerky with 944s.

Thanks for the post, please keep it coming!
Ahmet

[This message has been edited by 1.2gees (edited 04-23-2001).]

[marksportcts]

Does the DME send any engine codes, and if it does, how do we decipher them?

[mike944]

Ahmet,
you make some iteresting points about the AFM, and since i have a spare i was about to go home and try it. BUT.... as i thought about it more, modifying the AFM doesn't make any sense. Here's what i think. if the AFM is fully open at 1700-2200 rpm at full throttle, drilling holes in the door won't let any more air flow through, because nothing would need to flow through the door, because all the air is flowing around the door. Right? Therefore modifing the door, and spring tension would only reduce restriction and increase power at part throttle, or under 2200 rpm. But, at part throttle, if you need more power, just push the gas pedal farther. Under 2200 rpm, most of us won't load the engine at that rpm. there isn't much power there anyway, besides loading an engine at low rpm is not good for the bearings.

I think it would help throttle response though.

2. Altering the temp signal. I will have to preface this part by saying i do not specifically know how the Bosch Motronic DME is programmed to respond, but i will assume it responds the same as the majority of other engine control computers. Most computers take the signal from the O2 sensor as gospel, they will ignore all other sensors to please the O2 sensor. Therefore if you F_ with the temp sensor signal, and tell the engine to feed more fuel, the O2 sensor will tell the DME that it's running rich, and the DME should back off the fuel. Regardless of how cold the DME thinks the engine is, if the O2 sensor says the engine doesn't need more fuel, it won't give it any more. This is also why i think screwing with the 6-position fuel/spark adjustment switch on the DME won't do anything. And on most engine computers, the O2 sensor takes over fuel mapping as soon as the computer starts seeing valid readings.

Now maybe you could F_ with the O2 sensor signal. but this might not work either because of the non-linearity of the O2 sensor. I'm not sure exactly how far on either side of stoichiometric the sensor readings go to hell, but it's not far. I know some of the new cars have wide range sensors, but i know absolutely nothing about them.

I'm not trying to shoot your ideas down, i think they certainly need exploring, but i just wanted to bring up a few points you may not have thought of.

Mike

[justin]

Very interesting Ahmet,I'll think about this too,I'll have to agree with Mike I think the O2 sensor and the temp sensor would fight each other....hmmm

[1.2gees]

marksportcts, the early 944 motronic does NOT store trouble codes. The later ones do, codes can be read with a 'reader', as with any OBD engine management system.

Mike, the door posses a large restriction in the intake as I've noted. Just because the flap is open, doesn't mean it's not restrictive anymore. It still drops air pressure (by a lot), because the air has to push on it constantly, keeping it open. Drilling holes in the AFM flap will increase throttle response, I doubt it'll do much at full throttle, that's where the spring tension comes into play. The benefit(s) drilling holes would provide is two fold. First (again as noted before), it would remove a lot of the hesitation (in other words improve throttle response) off idle, but also it lets you back off the spring tension a lot more, as air flowing though the flap won't be pushing it. So less ressistance to movement by the flap, as well as less force applied by the air to move the flap. This would also have side benefits, as I've noted earlier in my original posts as well.

I did not suggest tampering with the 02 sensor at all, but since you're bringing it up, let me clear out a few things. First off, the variation in the mixture, BECAUSE of the 02 sensor output's extremely small, specially in comparison to how much the engine temp. sensor output mkes a difference. One more time, as I noted earlier, timing is also modified for more power, when the DME thinks the engine is cold. I never suggested playing with the 02 sensor, let alone both 02, and temp. sensors! [b]In short, keep in mind the 02 sensor output is only used for minor mods to the air/fuel mixture, but modified timing for more power would be an additional benefit EITHER WAY.

I welcome ALL input about this, don't see your comments as shutting anything down, I'm just trying to make sure I'm not overlooking anything. I'm quite confident both are good ways to improve power, and even more so throttle response for very cheap to free.
Ahmet

------------------
It's all the driver...

[mike944]

Ahmet,
good point about the spring tension. you've convinced me to the point that i'm about ready to hack up my spare air flow meter. I have an idea how to recalibrate it using a shopvac and an ohmmeter, both before and after. Any idea how to get the flap out so it can be modified without damaging anything?

I'm not convinced yet that messing with the temp sensor will affect the mixture. Although it might affect the timing, which would certainly help. I still think that the O2 sensor will overrule the temp sensor. Once then engine warms up to the point that the O2 sensor is giving good readings, the DME should rely on them. The O2 sensor is closer than any other sensor to monitoring actual combustion conditions. If the O2 sensor says the engine doesn't need more fuel, i don't think the DME will feed more. If the DME did feed more fuel, it would likley throw the O2 sensor out of it's very narrow linear range. Once it's out of that range, the DME will interpret this as a "bad sensor" I don't think the DME will ever ignore the O2 sensor. Unless, of course, it can't get ANY good readings out of the O2 within the allowable A/F adjustment range.

A way to tell would be to get access to an engine diagnostic computer that plugs in-line between the wiring harness and the DME. The snap-on one has a light that can be configured to read the O2 sensor output. Every time it goes rich (<14.7), the light comes on. Every time it goes lean (>14.7), the light goes off. If the engine computer is doing it's job correctly and the O2 sensor is good, the light should always be flickering. It should never be always on, or always off. I have a friend who owns a garage who has one of these snap-on computers, and i've played with it on some other cars. He doesn't have the Porsche adapter, and besides he lives too far away to borrow it.

I think i'm going to try the AFM mod, whenever i get a chance to do it.

Mike

[1.2gees]

I have one more book to read before I actually do anything, I may find something there. If you want you can let me do it to my own car, and report results before you mess with yours.

One thing that I THINK you're overlooking, the DME doesn't know that the oxy sensor is at operating temperature. One more thing, since the DME will think the engine isn't warm (yet), it will run the car richer until it thinks it's warm. This is why emissions are a lot worse when the engine is cold as well. Regardless of the O2 sensor output.

The temp sensor seems to be a more important input to the DME for the mixture, IF the sensor reads cold. Let me put it this way, the DME is programmed to run for max. power (or let's say richer, with timing altered for more power) when the engine is cold, so even if the oxygen sensor output was considered, it would still not make for much of a change. If anything, this would keep the mixture at the desired (but richer than stock) ratio. Anyway, there's again why I don't want to install an on/off switch, instead of an adjustible one. Firstoff, often changes that occur in a smaller amount of time, then is logical are ignored, also the ability to get back perfect emmissions and fuel economy is always there with an adjustible switch.

As for calibrating the surface area removed from the afm, vs. tension (reduced)? I'm thinking mark the flap position at given rpm, with given throttle opennings. Drill holes +back off the tension, and try again. Then keep playing with the spring tension, until all the positions match, or are close.

PS: I've got one more book left to read, then I'll be doing these mods to my car, specially the AFM, you can wait until I try it on my car, then I'll be able to help you if you run into anything as well. I don't plan to take the AFM apart, taking it off the air box, and the injection boot should be enough. Just get a drill, and drill it from the engine side, to the air filter side. There's nothing to hit that I can think of if you go this way, since the potentiometer is above the assy, on the outside. With the exception of the temp. sensor that's located inside the AFM.
Ahmet

------------------
It's all the driver...

[mike944]

I don't know if the DME considers the temp sensor to be "more important" only when cold. Maybe it does, maybe it ignores the O2 sensor until the engine reaches a certain temp. We would have to have the source code for the DME software to know that.

BUT it does know when the O2 sensor reaches operating temperature, because of the output. When an O2 sensor is cold (<575 Deg)
the output is zero volts. Once the output reaches something greater than zero, it's up to temperature. Newer computers will "test" the sensor by forcing a momentary lean condition, and waiting for a correct response. I'm sure our older computers don't do that though. So basically, if the O2 sensor is putting out voltage, it's up to temperature. Whether the DME ignores it until the coolant reaches a certain temp, who knows?

I was going to modify my spare AFM, only because i have a spare. It's sitting in my basement collecting dust, and probably providing a home for mice. (I still have to test it, it may not even be any good) I probably wouldn't be able to do it until next weekend, so if you are going to try yours before then, you'll beat me to it.

Do you have access to a dyno, or are you evaluating these mods using the "ass dyno" (i.e. it pushes your ass back into the seat farther)

Mike

[Gip87s]

Interesting ideas floating around here, far above my head but I do have a question. Why are Mass air style sensors so dam expensive for a Porsche? I also have a 95 Mustang GT and I can buy upgraded sensors for it all day long for Less than $300. almost makes me wonder if a sensor from another vehicle wouldnt work.

Gipper

[marksportcts]

So where would I find this "reader" to read the engine codes?

[lm6y]

Ahmet, I think your head is going to explode!
Also, don't forget about modifying the signal from the air temp. sensor in the MAF. All of your points are valid, and many are used on the latest generation of the Camaro/Firebirds. Moving the air inlet temp to a cooler location gets you a few HP. "Tricking" the computer into running in the open loop will make it run richer also making more power. This is done with a special engine temp sensor, and a 160 degree thermostat. Like you said, the computer never thinks the car warms up.

I think the biggest benefit would be from adapting a hot wire MAF to the system, instead of the more restrictive barn door AFM. If someone can do it for the 951, it CAN be done for the N/A cars as well. Unfortunately, there's not much market for a MAF system for the N/A 944.

Do you have the Bosch Fuel Injection & Engine Management book by Charles O. Probst, SAE ? It says that if a large adjustment is made to any one sensor (by an adjustable switch as you say), the DME will ignore any changes made to that sensor, and ignore it, running the engine in a "limp" mode with preset values to only keep the engine running, but not at it's optimum. A better idea would be to figure out the resistance of the temp sensor when cold, and wire a resistor in series to give the same value when hot.

[porsche1987_944s]

I have seen the engine temperature sensor modification that you are talking about done to Escort ZX2's. This works very well on the escort because the ECU is set conservative from the factory. Here is how they did it:

http://pages.prodigy.net/keman/tempr.html

[This message has been edited by porsche1987_944s (edited 04-23-2001).]

[1.2gees]

mike, the DME knowingly runs the car richer when the engine is cold. This is what it's programmed to do. Let's just say it starts getting a signal from the 02 sensor, this'll only help the engine run at exactly how rich the DME wants it to be. In any case, when the DME thinks the engine is cold, it'll run richer.

Gipper, the Porsche MAFs are expensive, cause there aren't many of them floating around.

marksportcts, where to get a trouble code decoder? Oh I dunno, dealer?

Im6y, I don't think I need to modify the air temp signal in the AFM. I'm going to make the car run richer only through making the DME think the engine is cold. This optimizes the air/fuel mixture, as well as the timing for more power. Also, keep in mind, with the holes offsetting how much the flap is pushed vs. the tension, the movement of the flap after the mods should be close to it's movement before. Only less ressistance to air movement, don't need to mess with air temp. signal. Or at least that's what I think. No doubt there'd be a bigger benefit from a hot wire MAF, but that'd cost money! (As you've noted, not much market for this upgrade on 944 NAs). Have the book, and read it through, I need an adjustible switch that'll alter the ressistance between the values that the DME would expect from the temperature sensor. Thinking about the resistor there, this doesn't let you adjust everything back to stock. I'd rather set my car up for more power, and leave it there, but still have the ability to switch to better milage (next time the car's started maybe?).

porsche1987_944s, thanks a bunch for the link, this is pretty much what I'm suggesting, except that an adjustible switch, in place of the ressistor for the ability to get fuel economy, emissions, etc. back as desired.
Ahmet

EDIT: I see what you're saying about the temp. sensor inside the AFM, so timing is advanced even more right? Perhaps that's doable, and another mod, however I'd like to just add a ressistor to the air intake temperature sensor's output, rather than have it totally out of service. The best thing perhaps is, I can try this, without spending more than $5, and it'd be reversible. I should take a timing light, and see if the timing is advanced even more, when I fill up the air box with ice!

------------------
It's all the driver...

[This message has been edited by 1.2gees (edited 04-23-2001).]

[3sonfather]

Too busy to read every word of this post but here's some input for what it's worth: if you start drilling holes in the AFM flap, I would think that would cause degraded performance since the flap will not open as quickly or as far during throttle since air wont be forced completely around the flap - but through it. Also, the low voltage circuit in the AFM is non linear, in fact it is somewhat parabolic meaning its resistance rises to some point then falls. It is fully relational with the other sensors.
Although I truely respect the creative spirit of this topic, I doubt whether there are any gainers here. I'm afraid Motronics has got the best of us. Good luck and have fun!
Scott

[1.2gees]

That's why you back off the spring tension 3 sonfather.

+The real gain from this mod is throttle response at lower rpm, as well as less restriction. But I'd look forward to running rich-er, as well as pushing more air in the engine, on top of advanced timing to get me(us?) some real power. That's how chips make up to an arguable 15hp.
Ahmet

------------------
It's all the driver...