Preface:
I have a 99 Miata with the infamous voltage regulation issues that commonly appear in the DiyAutoTune 99-00 MSPNP2. The ECU has the tendency to slowly overcharge the battery to 16V+ over the course of a couple hours of driving, which causes the engine to lean out and stall at idle and over-run. I found that the reason this occurs is because of a setting that appears in the Injector Dead-Time/PWM settings on TunerStudio, called 'Battery Voltage Correction (ms/v)', which tries to correct the increase & decrease in fuel flow seen at the injector when the battery voltage increases & decreases, respectively. It is also a fact that when an alternator is outputting too much current , it induces a greater mechanical load on the engine, further contributing to the loping idle. I run a set of ID1050x injectors on my car, which is very, very sensitive to little changes in pulse width. The original setting from a 99-00 basemap is set to 0.2 ms/V, which will add 0.2 milliseconds of pulse width for every volt below 13.2V the ECU sees. This is a really good approximation for both stock injectors and the ID1050x for when you encounter voltages below 13.2V. However, if you know your electrical system will be subject to voltages well past 14.2V, this setting will actually take away 0.2ms of pulse width for every volt above 13.2V, which will cause your engine to lean out, especially when running hyper-sensitive injectors. It's not such a bit deal when you're running high duty cycle on your injectors(relatively large pulse widths), such as WOT and high boost conditions, but this effect becomes apparent at light loads, such as idle and over-run.

A Solution to overcharging:
I have been told by one of the kind engineers at DiyAutoTune that the reason the 99-00 MSPNP2 has such a hard time regulating voltage is because of "a ground voltage offset between the alternator and the battery negative which will cause the alternator field to 'latch' on, in spite of the ECU *trying* to turn it off." The solution to this problem is as follows: "you just need to add a good ground strap between the body of the alternator to a good ground point on the chassis (these alternators seem to ground poorly through their mounts, especially as they age). This eliminates any voltage offset between the two, allowing the ECU control to again work correctly." I did this, and it really did help quite a bit, but after a few hours of driving, the system was still getting up to 15V, albeit far slower than it did before. They recommended that if the ground strap didn't fix the problem, the ECUs VR circuit may have been damaged because: "it's possible the voltage climbed high enough to hurt the control circuit (as it kept trying to turn the field off in the face of the latched field getting stronger and stronger the whole time)." Instead of messing with the internal regulators, another solution is to replace the NB alternator with an NA internally regulated alternator, which is relatively straight-forward and in similar price to having the ECU repaired.

Anyways, I kind of went off on a tangent about an issue that won't even matter for most ECUs, but I felt like this setting is overlooked and should be taken into consideration when changing to a drastically larger injector. It can probably be overlooked with a good tune and an alternator putting out a safe voltage, but if you ever notice your car leaning out from a known AFR to a strange AFR at idle, this very well could be the reason for this. I included a photo below that I created that shows you how to decide what value to set your Battery Voltage Correction factor on the MS 2nd gen ECUs, and just for reference, I used Dynamic Injector Flow data for the common ID1050x that I found on this website.

Anyways, I hope this helps, have a nice day.


A collection of data, MATLAB plotting, MSPNP2 Injector Dead-Time settings and calculations.




This picture shows where I attached my O'Riley Auto Parts, Doorman Help! brand ground strap to the body of the alternator




This photo displays where I attached the end of the ground strap to an unused threaded hole on the chassis. Make sure you brush all the paint off of the chassis before you connect them.

 

Great write up! I applaud your diligence in working this out with DIautotune.

I solved a similar issue years ago by ditching the NB1 alternator for an RX-7 alternator (which is self-regulating and supplies more power than the NB one does). There was an additional issue caused by this switch, which was the exciter for the field winding wouldn't kick in until the alternator saw greater that 1500 RPM or so. That meant that the alternator wouldn't start charging on first start-up. I solved that by putting in a manual switch that sent 12v to the exciter - flip the switch at key-on and the alternator starts charging immediately. I've been running like this for more than 10 years.

One day, I'll join the 21st century and buy a MS3 Pro. But why do it "right" when you can continue to patch an outdated system?!