[rusEfi] 93 1.6l Miata stock engine (for now)

What engine are you torturing? Is it inside some car?
User avatar
kb1gtt
contributor
contributor
Posts: 3442
Joined: Tue Sep 10, 2013 1:42 am
Location: ME of USA
Contact:

Re: 93 1.6l Miata stock engine (for now)

Post by kb1gtt » Fri Sep 06, 2019 9:43 am

hmmm, what died and why did it die? Then how do we make it more robust against what ever situation that caused that.
Welcome to the friendlier side of internet crazy :)

stefanst
contributor
contributor
Posts: 572
Joined: Wed Feb 17, 2016 12:24 am

Re: 93 1.6l Miata stock engine (for now)

Post by stefanst » Fri Sep 06, 2019 12:32 pm

I was logging at the time it first happened. No trace of anything suspicious in that log.
The NA6 Miata does have a weird sensor ground loop which seems to cause all kinds of noise. So I can't say for certain that we didn't have a signal above 5V on some sensors at some time. But I am under the impression that our signal conditioning should take pretty good care of that.
Also worthy of note is that the MAP sensor, where the problem occurred, is grounded properly. So the signal quality is better than the other sensors.

stefanst
contributor
contributor
Posts: 572
Joined: Wed Feb 17, 2016 12:24 am

Re: 93 1.6l Miata stock engine (for now)

Post by stefanst » Fri Sep 06, 2019 1:54 pm

russian wrote:
Fri Sep 06, 2019 3:35 am
Hardware and reality. Two things we cannot trust :( But we probably do not have resources to deal with limping mode and automatic switch to limping mode :(
In my particular case, I think it would have helped to set the MAP value to let's say 60 when it's out of reasonable range. It would have run terribly- rich as hell at low throttle and super lean at more throttle, but it maybe would have run.

User avatar
kb1gtt
contributor
contributor
Posts: 3442
Joined: Tue Sep 10, 2013 1:42 am
Location: ME of USA
Contact:

Re: 93 1.6l Miata stock engine (for now)

Post by kb1gtt » Fri Sep 06, 2019 4:05 pm

Hmmm, should I just consider the board was zapped while it was on the bench, and it took this long for the ESD issue to produce a noticeable result? I'm a skeptic about that, but also it's possible. I'm mostly trying to make sure there isn't some kind of data here that could help make the design more robust. It would be great if we could get to root cause. However it sounds like it's simply an unknown as it's to hard to track down for sure. If you have any intuitions or similar gut feels about what happened, feel free to let me know.
Welcome to the friendlier side of internet crazy :)

stefanst
contributor
contributor
Posts: 572
Joined: Wed Feb 17, 2016 12:24 am

Re: 93 1.6l Miata stock engine (for now)

Post by stefanst » Fri Sep 06, 2019 8:31 pm

kb1gtt wrote:
Fri Sep 06, 2019 4:05 pm
Hmmm, should I just consider the board was zapped while it was on the bench, and it took this long for the ESD issue to produce a noticeable result? I'm a skeptic about that, but also it's possible. I'm mostly trying to make sure there isn't some kind of data here that could help make the design more robust. It would be great if we could get to root cause. However it sounds like it's simply an unknown as it's to hard to track down for sure. If you have any intuitions or similar gut feels about what happened, feel free to let me know.
You may have a point there. I do swap brainboards around, without care, all the time. No grounding, no static protection at all. This particular board has probably sat in a plastic bin for at least 2 years I think.

stefanst
contributor
contributor
Posts: 572
Joined: Wed Feb 17, 2016 12:24 am

Re: 93 1.6l Miata stock engine (for now)

Post by stefanst » Sun Sep 08, 2019 10:44 pm

So, on to my Ground problems. It appears that, when the injectors are REALLY firing, my ECU ground is floating up by approx 0.3V or so. Since the sensors are essentially (more or less) externally grounded, this really messes with the readings from the sensors. The reading of my coolant temp sensor varies by almost 50*C from idle to full throttle. IAT only varies by about 15*C. Both happen within a few seconds. No way this is a real value.
Vbatt, as measured by rusefi changes by about 0.25V from FT to idle.
Hard to say what the variation of MAF is, since I don't have a means of comparison, but all these signals are super-noisy.
Only my MAP and EGO sensors that are grounded straight to the ECU show low noise and proper behavior.

Is anybody else experiencing something similar with an NA Miata? It looks like this particular style of ground wiring may be unique to the '93. See below:
Miata NA6 93 Grounds.png
Miata NA6 93 Grounds.png (197.99 KiB) Viewed 62 times
All we have is two wires, I'd say about 20 gauge, to supply everything, including ECU, sensors, relays, injectors, valves etc. with grounds. Just an awful design.

To add insult to the small wire injury, The sensors are grounded at the cylinder head, where everything else is getting it's ground from AND at the ECU. So we essentially have two wires, crossing at some point, turning into three wires at another point, crisscrossing and being a general mess.
I'm thinking about trying to find the wire that grounds the sensors to the block and not the ECU and just cutting it.

stefanst
contributor
contributor
Posts: 572
Joined: Wed Feb 17, 2016 12:24 am

Re: 93 1.6l Miata stock engine (for now)

Post by stefanst » Sun Sep 08, 2019 11:19 pm

So, after this rant it occurred to me that Mazda just CAN'T be THAT stupid. So I decided to test my ground connections. Turns out my power grounds going to the ECU are not grounded at all. Near infinite resistance to ground. Something is fubared.
This MAY explain why my stock ECU failed. It may have relied on those grounds. The rusefi is getting grounded via the sensor grounds which are tied into power grounds on the board. So it runs. Just with terribly noisy signals.

stefanst
contributor
contributor
Posts: 572
Joined: Wed Feb 17, 2016 12:24 am

Re: 93 1.6l Miata stock engine (for now)

Post by stefanst » Tue Sep 10, 2019 12:52 am

So this block and another one just like it, as well as the connectors hooking into them are apparently the reason why my sensors don't read so good.
Bad Ground Dist Block.jpg
Bad Ground Dist Block.jpg (52.96 KiB) Viewed 43 times
These are in the passenger footwell on the '93 Miata (pretty sure '94 as well, maybe 90-92 too). It appears that this footwell likes to get wet when the AC drain is blocked or some such. Need to do more research.

Post Reply