I have a '97 turbo Miata with an old Flyin' Miata ObiWan ECU. The wide band is an AEM unit which goes through a sequence of lights and messages once I turn on the ignition. A few drives ago I pulled off of the highway for gas and got the car filled up, only to find that it would not start again. The only thing I noticed odd was the AFR gauge start-up sequence.
Typical Sequence:

1. Sequentially display numbers 1.11-9.99
2. AEM
3. U49
4. P00
5. Gradually, sequentially rises from 10.0 - 18.0
6. - - -

Once I see '- - -', I crank it and the car starts first time, every time

On the day of reconning the sequence was

1. Sequentially display numbers 1.11-9.99
2. AEM
3. U49
4. P00
5. Started at around 14.6 and held there, very slowly ticking up.

I waited for as long as it normally takes for the gauge to read '- - -' and I think it got to about 16. Cranked the car but no luck. Sounded like the engine was choked. I ended up waiting for the gauge to read '- - -', which was about 15 minutes of occasionally turning the ignition on and off again. Once it was displaying '- - -' I cracked it and it started with a bit of throttle persuasion. The car ran home and idled fine while doing so. I really couldn't find much wrong with it at the time.
Since this day, the gauge has never read anything lower than 14 on start-up (cold or hot). And once I see '- - -', it starts fine. The new issue that has caused me to start worrying, is that the car is a little smokey now. Blue in color, I think. I've put and few hundred miles on it since the gas station. It's even had a little puff out of the engine bay when I was idling at a light. Where could that have come from? Crank case? Turbo inlet?
I'm starting to think I real issue with the car that's a slow burner. Any help would be greatly appreciated!

 

Unsure on your root issue, but a wideband almost never causes a no-start condition. In fact, the ideal way to make your wideband last is to program it to turn on only at higher temperatures, like 10-20 seconds after start up, or once coolant reaches a certain temperature, like 120 or 140.

My point is, it won't cause no-start. You want to look at compression, fuel pressure, ignition, injection, ecu, etc. Don't hyper-focus on the wideband.

And while ECUs like your old Link, hydras, MS1, etc can be great ECUs, you'll find it difficult to get help with them when they go wrong. I did find a thread from 14 years ago where a Link like yours sold for less than $400. Even then it was a ~10+ year old ECU.

Start by doing a compression/leak down test, which is a great time to install some fresh BKR-7E plugs gapped to ~.025", and installing an in-line fuel pressure gauge is always good for diag.

 

Thanks for the support, curly. I'm taking my sweet time on this due to tool inadequacy and other priorities; but here's what I've gathered so far.
I haven't yet understood entirely what a good compression number is for this particular build. I understand the car has 8.5:1 84.5mm pistons, but I haven't had the time to guesstimate its potential compression.
My compression test results are as follows:
C1 - 175 psi
C2 - 170 psi
C3 - 170 psi
C4 - 170 psi
My leak-down test results are as follows:
C1 - 100/93
C2 - 100/94
C3 - 100/97
C4 - 100/95
I think my next step will be trying data logging a highway pull and some cold and hot idling.
Fuel pressure and injection after that I think.

 

Those are good compression and leak down numbers. I would want 160+, and less than 10% leak down on each cylinders.

Let us know what else you find.

 

I had the car idling a little yesterday after installing new plugs, and it was running pretty sweet while cold. After about five minutes the mild blue smoke came out of the tailpipe again. Additionally, during the work I was doing, I took the turbo blanket off so didn't brush fiberglass into my arms. This coincidently allowed me to discover a constant wisp of smoke coming from the exhaust side of the turbo. I'm starting to think all my issues are coming from the turbo. Perhaps a leaky gasket or maybe it's about to give up entirely.

 

I'm still investigating my smokey car. I have pulled to turbo and cleaned it up the best I can. I thought it best and got the rotating assembly refreshed so it wouldn't be question when further investigating my issue. I was pretty much ready to reassemble everything this weekend until the step of priming the turbo with oil. I was suspecting the turbo to be 'blown' in some way and so I wanted to confirm oil was coming through the feed line. With the oil feed line just dangling, I cranked to car half a dozen times with not a single drop coming out. I found some advice that will let me get a faster crank my removing the plugs - makes a lot a sense - and it worked. Engine spinning nicely and yet not a drop of oil from the feed line. I'm looking for some advice/tricks that can help come to the conclusion on what's happening here. I don't have a convenient way of checking if the line is blocked because I think I'll need to take the entire plenum off to get at it. I think the feed line comes out just above the oil filter, but I honestly can't see exactly where. This feels like a bit of a rabbit hole that will end up with me pulling the engine. Any ideas why I have no oil while cranking the engine over that aren't catastrophic?

 

Miatas will still make oil pressure with ~1 qt in the oil pan, so don't worry about losing too much oil with a test. Have someone start the engine while you hold the feed line in a container. Let it run for 5-10 seconds after you see oil pressure on the dash before you start panicking.

 

I got the car put back together this weekend. Turbo was cleaned up and a new rotating assembly put in. Refreshed the oil feed and drain lines too. Long story short, the car has never run better in my time owning it. Obviously, something had worn out in that turbo. The AFR gauge still has its differing start-up sequence but isn't hurting the cars performance in any way noticeable. Any precedent for an WB sensor degrading from burning oil upstream?
I'm dipping my toe in the data logging now to see what's there. Took me a while to find the right software and manual (DataLogLab).
Thanks for a past input again, curly.