I've been trying to get my GM D514A coils running on my 2004 Mazda NB MSM for a while now and finally managed to do it on the weekend but wonder how/why this worked:

I'm running the coils in wasted spark mode, just like the OEM coils and had previously created two harnesses to connect the CNPs to the OEM harness:





Coil plugs 1&4 connected to OEM harness' coil-1 socket, coil plugs 2&3 connected to OEM harness' coil 2 socket (this is how the spark plugs are connected with my OEM coils and the car runs).
Grounds for both DIY harnesses are connected to the engine block.

I changed the Nominal Dwell to 3.2ms (see Dwell Battery Correction comments below for why I used this value), Spark Duration to 2.0ms and Cranking dwell to 4.5ms (had it as high as 8ms at one time).
The car would crank but never start.
The spark plugs all fired when testing them one by one on the rocker cover so I knew that electrically everything was OK.

One of the things I'd never bothered with on the MS was running the "Output Test Mode - Inj/Spk" Coil Test - I knew that the injectors sparked so the coils MUST be working!
SInce I'd run out of ideas, I ran the Coil Test and was surprised that the "Coil A" test actually fired coils 2&3 and "Coil B" test fired coils 1&4 - opposite to what I expected.

I switched over the OEM plugs connected to my DIY harness and she fired up right away!!! WTF???

Additional details (initial tune attached below):
2004 MX5 NB (born as a MSM)
EFR6258
3" exhaust
ID1000 injectors
D514A CNPs
FM 36-2 timing wheel
Iridium BKR7EIX-11 spark plugs

I used the following information from LS2 coil / megasquert signal wiring question - Turbobricks Forums to define the Dwell Battery Correction table:
The dwell time for an LS2 coil at 13 volts is 3.15ms, taken directly from an LS2 ecu.
Your dwell battery correction table should look like this (again, values taken directly from an LS2 ecu):
6.0v 193%
8.0v 182%
10.0v 141%
12.0v 113%
14.0v 89%
16.0v 75%
Since I couldn't enter 3.15 as the Nominal Dwell, it got converted to 3.2, I applied a formula to the above values so that they're based on 3.2 instead of 3.15.

WTF DID THAT WORK AND WHY IS IT THE OTHER WAY ROUND ON THE OEM COILS???
If anyone can answer this, please let me know, it makes no sense, especially given that all I changed was spark dwell times.

 

WELL IT'S CONFIRMED, I'M A BLOODY IDIOT!

On several sites, including this one, nobody has answered my query on why swapping the OEM injector harness connectors worked to get my D514A CNPs running so I had a look at the ignition system manual for the car.

Here's the Ignition coil setup on a 2004 MSM:


The OEM Ignition Harness' longer Connector (the left Connector labelled "1" above) powers Coils 2&3 and the shorter Connector (the right Connector labelled "1" above) powers coils 1&4.

For some reason, I thought the longer connector was powering coils 1&4 and the shorter one coils 2&3 so had them connected the wrong way around for the D514A coils.

If I'd paid ANY attention when I re-installed the OEM coils after failing to install the CNPs the first time, I should have picked this up!