What's the difference? And when did they switch?

MAF: use mass air flow sensor. directly measures the amount of air coming in.

Speed Density: Uses a manifold absolute pressure sensor, and then calculates what the air flow should be, based on that, and other sensors (Temp being one....)

Changed in 89, I think. Maybe 90... (on tpi at least) Throttle body systems are (i think....) exlusively speed density.

I think the main reason for the switch is the fact that the computers werent sophisticated enough to handle the MAF properly, not to mention that the bosch MAF was crap. The stock MAF computers suffered from a serious lack of resolution. GM basically relied on closed loop and the O2 to get the proper AFRs. The stock SD systems have much better resolution.

The year was 1990 when Speed Density was introduced. My MAF car runs pretty good. See signature. I have no issues with it.

IIRC Gm went MAF for the first few years of F-body, then to SD. If you mod a MAF car, it senses more air incoming, adds more fuel, you go faster.. With SD you have to burn a new chip, giant hassle etc. etc...
Rustang went the opposite, SD in '86, then to MAF after. (again, this is an IIRC from my cousin with an '83 stang)...

IMHO MAF is clearly superior in every way. (enough acronyms for ya?)

At high loads the MAF definatly works well. The issues with the tpi MAF system where that the MAF cant actually get to an AFR at low loads due to the resolution of the system. I found that with the 8 bit setup the AFR on the WB would toggle back and fourth around the target AFR at idle because with an 8 bit A/D the resolution was **** poor at low voltages. This wont cause any problems if your target AFR is, say, 14.7:1 because there is some wiggle room on either side of this and you wont notice the variences. In closed loop, the computer will adjust for the lack of resolution and get the AFRs around 14.7:1. Oh, there was also the 255 g/s limit with those as well but thats a whole 'nother issue in itself.

The place where the lack of resolution becomes apparent is when you try to do lean cruise. I tried this with the LS1 MAF feeding through a converter and with that thing and only 8 or 9 grams of airflow while cruising, the voltage is on the order of millivolts. Basically the car was a bucking bronco.

With 10/16/32 bit votage/frequency systems the resolution will be better but even with that theres still the problem of low flows with the big MAFs. The airflow isnt real stable. What they do now is use both SD and MAF together. Provides the best of both worlds, and provides redundancy as well in case one of the sensors fails. Im going to try to do this sort of approach with one of the later truck PCMs. Should be interesting.

90 was the switch over year. The main difference is that M.A.F. systems can respond much better to changes in the air, such as quality and so forth. The Speed Density system is more useful for a tuned racing application. It has programmed fuel/ignition maps that it accesses based on the information it is fed by the other sensors. It doesn't respond directly to air quality changes. For example: if you were driving two cars from the base of a mountain to the top of it, the M.A.F. sensor would automatically adjust for maximum performance as you climbed higher. The Speed Density system's performance would degrade as you climbed higher since it was not tuned for those conditions. This view may seem controversial to some people on here, but it is based on test case data from Tuned Port Induction Specialties. In otherwords, I didn't acquire this data personally, but I feel the source is reliable.

As for the example, SD will also sense a change in altitude, and has a barometric pressure term that can be used for this. This is used to adjust the VE accordingly to account for the change in efficiency as the absolute pressure on the exhaust side of the engine changes. If the baro based calculations are good, then itll do just as well, if not better since things like timing will be based off of absolute pressure and not that LV8 term. With the stock tbi cars at least, the pseudo-baro/alt. compensation calibration seemed to be off some and at least one person has reported noticable changes in fueling when going to high altitude. Recalibrate the VE compensation and this wouldnt be issue for the rest of the cars life.

Both systems can work equally well. If the temp compensation and such is good with the SD system, then performance with both systems will be equal.

Equal? Well, maybe in some respects.

Just my opinion, but:

The MAF system is a fairly reliable system that automatically compensates for changes in altitude, air density, and temperature since the sensor is supposed to measure the actual mass of the air entering the engine. SD has to calculate that. The MAF makes it a little easier for the casual modifier to experiment with intake modifications, valve train modifications, exhaust work, chamber and piston design, and still maintain respectable control over closed loop conditions. The caveat is that this occurs in increments, and has the aformentioned limitations. This is due to the nature of the ECM used with the MAF and not the theory of MAF itself.

MAP based systems used after 1989 have the advantage of a more powerful ECM that can address more memory locations faster, and therefore get by with a MAP sensor calculating a projected intake flow and checking results against the Oē more frequently. The greater number of flow increments (memory addresses) make finer control possible, but still rely upon a projected calculation of intake air flow to administer fuel correctly. It can allow for good control over a wider range of operating conditions, but relies heavily on volumetric efficiency calculations that are changed whenever a modification occurs. That makes experimentation a little more cumbersome, since the VE needs to be considered whenever a different cam, heads, intake, or exhaust is bolted on (reprogramming). This is another step that some modifiers are not prepared to deal with on their own.

The later systems incorporate both, operating primarily on MAF at lower loads and flows for finer control, and MAP at larger throttle angles, and using both sensor to compare and check the other under all conditions. The flow limitations are a lot more realistic, allowing a lot higher intake air while maintaing control. A lot more modification can be done while still operating within a respectable control range.

Of course, that's a relatively simplfied description, but should be an accurate summary. Which one you would call "better" is pretty subjective.

All of the M6811HC based ECMs/PCMs all had the same processing power if Im not mistaken. The crystals where 2^23 with those and 2^22 with the old C3 ECMs. A quick look shows the 7730 as having 16K of additional prom space over the 165, which gives it a leg up over the previous MAF ecm. I think they both have around 512 bytes of RAM. I think GM could have done better with the MAF ecms, but maybe with only an 8 bit A/D, better coding might have been like perfect imperfection.

One of the only 8 bit based GM computers that got the nice non-multiplexed MPU was the later PCMs. In those they have over 2x the ram, 2x the addressable prom space, and a better MPU. Still the same speed as the ecms, but the hardware was much improved, and much of it appears to be the same as the off the shelf MPUs, making reversing a little easier.

Ive only been doing the coding thing for like five minuts, though, so I could be FOS.

About the dual MAP/MAF handling, has that stratagy that was stated above been confirmed with actual info on the late model PCMs? Just curious.

TPI MAF 85 & 86-89 MAP 90-92
Ford MAP 85/6-88 MAF 89-93

Dimented,

You are correct. I described the later P4 ECMs as "more powerful" as a simplification. The additiopnal memory addresses are the main reason. The tables can be expanded further for more finite control, while the MAF systems had larger "steps" in the tables. This was a necessity due to the limited ROM space and resultant I/O addresing capability of the firmware.

There is little doubt that the TPI MAF systems could have been more sophisticated, but Delco did what it needed to do to maintain control over emissions and costs.

That last question can best be answered on the LT1 edit sites.

Huh, I heard it was the other way around. Interesting... I take a look.