I took the plunge and bought the Split Second MAF conversion kit with their PSC1-4, laptop programmable signal calibrator, along with their optional intake air temperature sensor. Split Second does not have a specific E28 M5 fuel map available. Unfortunately, no one on any of the E28 M5/S38 forums responded to my posts for a copy of their fuel map. So, I didn't have a good starting point but the system was easy enough to adjust to get a drivable base fuel map that got me to the dyno shop. Mark has also been extremely patinet and helpful in answering all of my questions. Following are the major points of interest:

Dyno Numbers- (Peak Rear Wheel Power & Torque)
2004-July: 244 HP, 237 ft-lb (w/ MAF, Conforti chip & gear)
2004-July: 230 HP, 220 ft-lb (w/ Conforti chip & gear)
2001 Dyno: 221 HP, 216 ft-lb (w/ Dinan chip & gear)
1999 Dyno: 223 HP, 213 ft-lb (w/ Dinan chip)

Engine Modifications-
Besides the new MAF conversion, Conforti chip, and exhaust cam gear, my engine is stock with 180k miles and no renewed parts. I recently removed the injectors for cleaning by a local Dr. Injector franchise but did not notice any difference in drivability afterwards. The flow test revealed a couple of injectors were about 6.5% low in flow compared to the highest flowing injector. After the cleaning all injectors were within 1.5% of each other.

Overall, my car runs fairly smoothly, as if it still had the AFM installed but revs more quickly with better torque from idle to redline. I have some very minor hesitation when I accelerate aggressively off idle. After re-installing the AFM, I noticed that the hesitation was always there but is a little more perceptible with the MAF. Otherwise, acceleration is smooth with power coming on seemingly quicker. The engine will rev very quickly to about 3.8k, and then rev similarly to when the AFM is installed to redline. I am hopeful that additional fine tuning of the fuel map will eliminate most of the hesitation. Intake noise is slightly higher but probably not noticeable to anyone else. I'm using the factory air box modified with a larger opening as the factory air filter has a lot of surface area compared to some of the cone filters included in cold air intake kits. With the optional intake air temperature sensor, I get pretty close to the same fast idle characteristic as it had with the AFM, though I think fast idle isn't quite as high but will have to wait until winter to know for sure. Idle quality is typical M5, a little lopey with no improvement over the AFM, though it was never too rough (for me). Though the MAF doesn't make my M5 feel like a new car, it is more fun to drive around town with the noticeable torque gain and quicker revving engine. Switching back and forth a couple of times between the MAF and AFM, for comparison, reveals enough of an improvement that I prefer the MAF.

MAF Setup-
First of all, if you decide to get a Split Seconds PSC set up, I would highly recommend that you spend the extra $50 to get the auxiliary input option which allows you to record up to 4 data points, like O2 sensor voltage, manifold pressure (for turbos), etc. I was not aware of this option until after my purchase. I was informed that the option cannot be retrofitted. The PSC1 does, however, have the capability to record engine rpm, MAF voltage and the fuel map value that is used to modify the output voltage to the DME. With this in mind, I cut into the PSC1's wiring to add spade connectors to the MAF signal and airflow to Motronic wires so that I can disconnect the MAF signal and connect the AFM signal to the MAF input into the PSC. This allows me to record the AFM's signal when I put things back to stock. To be able to reinstall the AFM, I fabricated an adapter plate that allows me to reinstall the AFM to the factory air filter box, as I cut a larger opening for the MAF. The rest of the MAF installation is per Split Second's instructions. One thing that isn't clear is where all of the ground wires for the new equipment should be connected. I think I noticed a little better performance with all grounds connected to the AFM sensor ground compared to connecting to one of the large DME ground wires.

MAF Tuning-
Split Seconds pre-installed fuel map was extremely rich and could not be used. By monitoring the air/fuel meter, it was relatively easy to get a base fuel map that would run acceptably but I knew I needed some dyno time to be sure I wasn't running lean. Once I figured out that I could record the AFM's signal, I made the wiring modifications, reinstalled the AFM and recorded an idle voltage of 1.0V and a WOT voltage of 4.2V. I was pleasantly surprised to find out that the MAF has the exact same output range of voltage from idle to WOT. After my dyno session, I had the 2,500 rpm to redline dialed-in with an air/fuel ration of about 12.5:1. Since then, I have made 8 different fuel maps with several iterations in each trying to optimize off idle acceleration and cold start drivability. After several weeks of tuning, I am pretty close but there is an infinite amount of adjustability with the PSC.

Other things-
I noticed that the AFM has a rounded rectangular entry nozzle which I assume reduces turbulence. I plan to do a little searching to see if I can find something similar for the round MAF. Is the MAF conversion worth it?...The power and torque increases are significant but not astounding. I was hoping for at least 20 rwHP. But because of the tunability, the knowledge I have gained and the fun I have been having with experimenting and fine tuning, I would do it again even if I was guaranteed only 14 rwHP. The entire MAF conversion kit cost me $770 (including a 10% discount) delivered plus another $315 for dyno tuning. So far, it appears that my "normal city" and highway mileage have not changed, 14 and 22 MPG, respectively, though I'll have to stop with the fine tuning for another tank or two to be sure.

I have read that at WOT, the DME ignores airflow. This may be a simple mistake as I would agree that the O2 signal is ignored but not airflow. The stock AFM is sized to measure the full airflow that the engine can pump as I have recorded the voltage ranges produced by the MAF and AFM during dyno runs and they are identical. RPM by itself is not a very good indication of airflow or engine load. Airflow will differ greatly depending upon the conditions on the engine. For example, you could be traveling up hill in a lower gear, versus highway cruising in a medium gear and be at the same rpm but very different airflow/engine loads. At the instant you go WOT, the fuel requirements are different. Even if the DME were to take this into account and is conservative across the rpm range, it seems like there would be a scenario where too much fuel is added hurting performance. I acknowledge that Iím no expert but it doesn't make sense for the DME to base fuel control solely on rpm at WOT when it can get an airflow signal to determine how much fuel is needed. Also, the real world reason why I believe this is that I was able to change the air/fuel ratio at WOT via the PSC1 during my dyno runs. My recorded dyno runs show this. Mark at Split Second has also told me that O2 is ignored, not airflow. I donít mean to offend anyone but want to share my experience.