After trawling the web I finally bit the bullet and decided to do my own testing on NB Speed Senders.
But first the back ground story on the fault I was experiencing.
Mine has been playing up since about mid last year.
It first started at a Supersprint where my speedo just stopped for all sessions and then came good in the last one.
The speedo was pretty much flawless with maybe 4 or 5 quick drops and recoveries until coming back from Fan Fest.
When driving home from Fan Fest I started getting more drop offs. I suspected I was developing the soldering issue that plagues these senders as I did drive on a dirt road.
Once home it then developed into the speedo would not operate until I hit about 40km/h.
At this point I was doing the bushings and decided to do the soldering repair while the car was up off the ground. The soldering did look rubbish so it was redone. Put it back in and it seemed ok.
A week later I started getting the same random drops in the speedo. It then developed the same not working until I hit 40km/h.
Another Superspint and on the way home the speedo died altogether.
Since then I have borrowed 2 senders. One with a Blue cog for 3.636 diff like mine and one with a white cog for a 3.909 diff from a nice yellow coupe.
My speedo operates perfectly with the replacement blue cog unit.
This is the table they have worked out on miataturbo forum which is not 100% guaranteed as it is for the USA and based on user findings not a Mazda book.
P/N: - gear teeth, gear colour >> trans+rear+tire (year)
M527-17-400A - 23-tooth, Green >> 5MT+4.3+15" (99-05)
M526-17-400A - 22-tooth, Red/Burgundy >> 5MT+4.3+16" (01-05)
M526-17-400A - 22-tooth, Red/Burgundy >> 5MT+4.1+15" (99-00 Euro market?) I guess this would be the Australian spec NB8A
M528-17-400A - 21-tooth, Yellow >> 6MT+4.1+16" (04-05, and 14" on Euro NB1s)
M528-17-400A - 21-tooth, Yellow >> 6MT+3.909+15" (99-00 Euro market?)
M534-17-400 - 20-tooth, White >> 6MT+3.909+16" (01-03)
M533-17-400A - 19-tooth, Blue >> 6MT+3.636+16"/17" (Australia market only)
Testing procedure.
I tested for resistance in ohms. I then used a drill at maximum speed for the AC voltage test and also hooked them up to a CRO.
The faulty unit had Open Circuit, 13V AC and 48V Peak to peak Triangle wave at 201Hz.
The good units had 233ohms, 18V AC and 60V Peak to Peak Triangle Wave at 201Hz.
I suspect my faulty unit needed at least 40km/h due to the lower voltage.
Based on all this testing I would be confident to say that the electronic sender units are all identical. Just fit the correct colour cog for your Diff/Gearbox/Wheel setup.
See below for pictures of the results.
My faulty unit with blue cog.
Good unit white cog.
I removed the circlip and cog so I could attach the units to a drill.
Resistance of my faulty unit. OPEN CIRCUIT!!!
Resistance of good white cog and blue cog unit where identical at 233 ohms.
AC Voltage of my faulty unit was 13V AC.
AC Voltage of working units was 18V AC.
CRO measurement of my faulty unit. Triangle wave at 201Hz with 48V Peak to Peak.
CRO measurement of good unit. Same triangle wave at 201Hz but higher 61V Peak to peak.
Many Thanks Luke 
