My 1.01 board let the smoke out a few months ago. I had a 3.0 box that someone else had built but never ran. I verified that the ignition input was compatible with my chip (B&G MS1 code), plugged the old chip in the new box, and I was truckin'.
Well, sort of. It was lean and stumbly. Found out that the driver circuits are a bit different. My O2 sensor was killed by an oil ingestion incident, and I didn't have a laptop for tuning anyway, so I borrowed a laptop and tweaked the injector dead time settings until it ran okay. 1.0ms was too lean, 1.2ms was too rich, 1.1 was both. It was a hot week when I did this, and I thought it odd that the MAT was reading 133F after sitting for eight hours, but hey it was hot out.
Bought a new laptop, was setting up TunerStudio, and noticed that the MAT still read 133F. All the time. Even when you unplugged the sensor or shorted the connector. Well, that explains why it runs like crap when cold, and rich when really heat-soaked. So I gingerly picked and poked and discovered a bad solder joint at the board/CPU socket (pin 24). Resoldered that and...
...now the MAT is stuck at 233F. Shorting the MAT input to the computer (pin 20) makes it read down to 170F. Same with directly shorting pin 24 at the CPU socket. This is true for the B&G chip that ran just fine in the 1.01 board, and the MS1/Extra chip that came with the box.
I bet the car is going to run REALLY GOOD now. 
Re-re-solder --- shoot for a nice even 333F. 
The part that is somewhat confusing is that IATs are negative temperature coefficient. That is, resistance goes down as temperature goes up. So shorting the lead should make the temp go UP.
This is unless you have an 80s Audi with Motronic, in which case you have a positive temperature coefficient sensor because German, and the only place you can really find one is Jim Ellis Audi and they will tell you that the sensor is a few weeks out because international backorder because who in their right mind buys sensors for 30 year old Audis, and then the next day when you think to yourself "this is a sign that I need to Megasquirt this car too, they will recieve that as some sort of collective unconsciousness Bat Signal and immediately ship that shiny, candy-like $170 sensor.
(Anybody wanna buy an Audi temp sensor? 
)
As soon as laptop has some more charge, I'm going to experiment with unplugging the sensor and seeing what it does.
Still waiting for laptop to charge some more, but I had a kind of revelation and dug up my Haynes RX-7 manual.
GM IATs at 68 degrees is ~2.4k ohms, 77 degrees is ~2000 ohms. It's something like that outside.
According to the Haynes, the manifold air temp sensor's range is like this:
68F: 41.5k
122F: 11.85k
185F: 3.5k
I wonder if the guy I bought it from set it up in hardware to use the Mazda temp sensor. Extrapolating that curve, I can see a 2K resistance being read as 233F.
Bugger. I don't think I have any Mazda temp sensors.
And just think, that poor solder connection is why it was able to run in the first place.
edit:
R4 is different than all the others. R4 is supposed to be a 2.4k ohm resistor, which is what the IAT (MAT) should be at 68 degrees. R4 is orange-orange-orange-gold, which is... lemme find a chart... 33k ohms.
Well, this explains a lot.
Just to put a done stamp on this: Unplugging did result in -40 temp reading. Went for a drive to see what the gamma correction is on -40. (It's 125%, for record)
WOW it was running rich :) White smoke on cold idle and it was bucking even at 75mph uphill, with the O2 monitor reading 10:1 even with the raw air getting into the exhaust from the misfiring. Until I get the resistor swapped out, I dropped my injector on-time to .9 and my req_fuel from 6.4 to 5.0ms. It's still not right but it's better than it was before.
I'm really glad that my 10 heat range NGKs didn't foul with all that fuel. MSD ignitions rule.
... 10 heat range plugs? WUT.
What sort of MAT sensor are you using?
Do you get something of a reading on the MAT gauge if you check the input on a Stimulator?
Swank Force One wrote:
... 10 heat range plugs? WUT.
it's a rotary thing. The plugs run really, really hot. Stock is 7 and 9. I had been running 9s all around, but was still getting some top end missing when at WOT for long periods of time and high coolant temps, so I went to 10s.
Next step is the 11.5s and 13s but those cost a lot more 
Matt: I built the car around a GM IAT. Silly me, I'd assumed that the guy who built this 3.0 board was also building it for GM sensors, or at least was planning on doing the corrections in software.
I don't have a stim. My electronics ability begins and ends at making horrible cold solder joints so I buy pre-built units. Honestly, I'm surprised that I was able to fix what I did.
Gotcha, forgot it was a rotary.
I get people calling me crazy for running ZFR7F-11s sometimes.
If they don't foul, then they're hot enough!
The only thing that "got" me about the cold plugs (besides the whole idea of putting strap style spark plugs in a rotary) is that the porcelain is so short that you have to run tighter gaps than normal to prevent the spark from just shorting across the porcelain. I used to cut down Autolite plugs and run a .040 gap, these R5671A-10s have to be gapped at .020! Some run them as wide as .022! Cripes that is narrow. I was worried that the tiny gaps would result in poor low load drivability but it feels the same as with the stock open-faced 7s and 9s. As a bonus, I can buy four of these for what two of the normal plugs cost.
People run the fine wire 11.5 and 13 heat range plugs at .010!
Knurled wrote:
I don't have a stim. My electronics ability begins and ends at making horrible cold solder joints so I buy pre-built units. Honestly, I'm surprised that I was able to fix what I did.
In that case, you will either need to improve your electronics ability or need assistance from someone else who has better ability here.
No worries. I'm farming out the repair to a guy we have do all of our electronics.
