Necro post, but some of us are too cheap to part with our old 1st gen MSPNPs. And I'll bet nearly everyone that has an old MSPNP with it's original MAP sensor has low MAP readings. In my case, I was seeing ~86kPa on the red car and 88kPa on the silver car (key on prior to engine start -- atmospheric pressure 99kPa).
Believe it or not, this is normal behavior for an absolute pressure sensor. Modern, MEMS-based absolute pressure sensors typically rely upon a sealed vacuum reference on one side of a sensing diaphragm, with the other side open to the pressure of interest (MAP in our case -- see attached MPX4250A datasheet). Alas, nature abhors a vacuum. As air molecules leak into the vacuum reference over time, the output signal from the sensor diminishes for the same applied pressure. And the issue isn't limited to low-cost sensors, even aerospace-grade sensors for air data computers certified to 55,000 feet have this issue.
The best way to handle the issue is to provide for periodic re-calibration. Thankfully, this feature has been added to later MS generations. That doesn't help those of us nursing our older MSPNPs along though.
Some have stated that this is a harmless flaw, that all you need is consistency and you can ignore the lack of accuracy. Relative to fuel tuning, I would agree. The change in the sensor happens gradually. Assuming you periodically autotune your VE table, this will never be an issue for your AFRs. However, this can become a real problem for spark advance, boost control and overboost protection. For each of these functions, the lower-than-real pressure sensor reading cuts safety margins.
So, what to do for an early MS? One solution is to replace the sensor. I recently bought some new MPX4250s from Digikey and swapped out the sensor in the silver car's MSPNP. Not hard but tedious.
When it came time to do the red car, I figured there had to be a way to do this in software. My first attempt was to manipulate the calibration files like Matt suggested (these are barofactor*.inc and kpafactor*.inc files in your project \src directory). However, I discovered this didn't work. On the early MSPNPs, the converted MAP value is not transmitted on the serial port, rather raw MAP sensor ADC is sent. That means the TunerStudio MAP reading doesn't change -- so you'll end up with a mismatch between what is inside the MSPNP and TunerStudio -- not good.
My next attempt was simply to put an offset into the raw MAP sensor ADC at the point of acquisition. I did this with a series of inca statements. This is a screen capture showing the code diff on 029y4a software.
This worked perfectly and was a whole lot easier than replacing the sensor. With the 4-bar sensor in the MSPNP9495, each inca statement increased the reading by ~3.5kPa. The 3 inca statements corrected my pre-start MAP from 86kPa to 96kPa. You'll need to experiment with how many inca statements you need -- and also keep an eye on things in the future as the sensor continues to age.