bob13542

First off, Hello everyone! This is my first post here, I apologize if it's in the wrong place.

I am having a very weird issue with my 1990 Miata. I bought the car roughly 2 months ago, and have been doing a lot of work on it to get everything running, and more importantly keep everything running. (it's broke at least 4-5 various things at this point, but so far until now I've been able to diagnose and fix them).

Background Info:
-Miata has had many modifications done by previous owners (not sure what all of them are, and I'm still finding things that have been tampered with here and there)
-At one point it had an aftermarket alarm system (I removed a whole bunch of wiring that was spliced into the ignition switch including an immobilizer relay
-general mods such as aftermarket steering wheel w/quick release, stereo, speakers, seats, shift ****, air filter, wheels
-Aftermarket Megasquirt ECU with AEM-X wideband (installed by me 3 days prior to this issue occurring, and during those 3 days the car ran better than it ever had before while I've owned it)
-While installing the Megasquirt I deleted the MAF and replaced it with the commonly used GM air intake temp sensor (Megasquirt also has the built in MAP, with the vacuum line going from the intake manifold to the ECU
-While I've owned it car has always had somewhat erratic idle that I've been trying to track down (never would stall before, but would sometimes idle at 850, sometimes 1k, maybe once in a while 1100 but always in that range


On to the Issue
-I was driving 40mph in 4th gear one street away from my house, when all of a sudden I heard a click and the RPMs on the tach dropped to 0. Within a second I heard another click and the RPMs shot back up to around 3k, where they should have been. When this happened, I felt the car shudder, not necessarily jolt extremely strongly, but while the rpms were 0 it definitly didn't have any power or anything. This makes me think the engine did indeed stall, rather than a weird rpm signal. This happened 3-4 times within the span of 15 seconds. I was able to get the car home by coasting around the corner into my complex (occurred 1/4 mile from my house) and driving in 1st gear at 10mph to my house.
-an hour later I decided to drive around my complex datalogging with tunerstudio to see if I could replicate the problem. I made it 3 houses down before it stalled, and did not in fact restart itself which I thought was going to lead to a rather difficult push up into my carport. I could get it to crank, but not start. After 10-15 minutes of me trying to decide what to do, I tried to start it again, and to my surprise it started. It stayed on just long enough for me to back into my carport, where it abruptly died again.
-I've learned that if I wait 30 minutes or so, it will always fire right up without much issue (1st crank of turning the key). However after idling for a minute to several minutes it will die. It will never restart right after stalling unless you wait a minimum of 20-30 minutes.
-When it won't start, I cannot observe a spark either by taking a plug out keeping it connected to the wire, while grounding the threads, or by connecting an inductive timing light
-After cranking it without it starting, I can smell fuel in the cylinder but the plug doesn't appear "wet"
-I can hear the fuel pump prime if I bridge F/P and GND in the diagnostic box
-On tunerstudio logs (attatched) I can see the rpms drop to 0 when it stalls, and at the same time there is a sync loss with reason 39 (CAS?)
-I ran a composite log, and I can see an output from the trigger and toothtime values (Not experienced enough to know if they are proper values, and I haven't been able to find a known good graph or set of values to compare mine to)


Questions...
Is it possible the CAS is bad even though I still smell fuel in the cylinders after cranking?
Is it possible the CAS is bad even tough it starts up without fail every time if I wait at least 20 or so minutes?
Could it be the ignitor, or ignition coils, or I guess plugs and/or wires, even though similarly it starts up without fail if I wait at least 20-30 minutes?
If the toothtime or trigger values are indeed abnormal, is it possible they are wrong because something else is failing and stalling the engine, other then the CAS, or are they proof of a bad CAS (or wiring to the CAS)?


Basically my issue is a new CAS appears to be anywhere from $200-350, and I cannot afford to spend that unless for sure the CAS is the cause of this problem. Similarly I'd rather not buy a new ignitor or coils just to see if it might solve the problem, without knowing if they are the cause.


Sorry for the long wall of text, I've just been trying to be as descriptive as possible, and include every detail I can think of in case one of these small little details points right to a problem I've overlooked. Let me know if there's anything else I should check, and thanks for reading if you made it this far!

gooflophaze

Welcome. It's probably the cas. No one pays $200 for a replacement, plenty are floating out there used. Where are you located?

bob13542

Thanks! And yeah, shortly after posting I was looking around on eBay and found several from $60-100, so I think I must have been getting the cost confused with a different part. I'm located near Portland Oregon. And thanks for the reassurance! this is my first time trying to diagnose a no-start like this, so I didn't know if it still could be the CAS considering it starts after sitting for a bit.

gooflophaze

You've got a lot of good information in the original post, and while I could throw out a few "it could be" scenarios - the behavior screams CAS to me - fails when warm, sync loss, works for a little bit when cooled down. CAS failures I've seen either act like this, or they just fail completely.

bob13542

That's good to hear! (not that it failed, but knowing what failed haha). I'll go ahead and order a new CAS tomorrow morning once I get my paycheck deposited.
Thanks so much!

bob13542

Just a quick little update. I found a scrapyard near me that supposedly had a CAS available. I drove the hour to get there, only to have the guy attempt to charge me $150 while still expecting me to remove it from the vehicle myself. I instead proceeded to order one off eBay for $85, which will supposedly arrive by Friday. I'll update here after it's installed, to let everyone know if it fixed the problem.

gooflophaze

And I'll throw this out there as some additional information - it is possible to rebuild them if you're handy with a soldering iron and digikey.

Our Friend CAS
https://forum.miata.net/vb/showthread.php?t=464994

bob13542

I took a look at those articles, and they look rather interesting. They seem to be describing the possibility of rebuilding one of the 1994+ hall effect CAS sensors. Any idea if it's possible to rebuild an optical one?

Also, I used the information there on pinouts to test mine. I got it hooked up to a PC power supply, and I can see the SGT/NE signal (the one that cycles from 0-5v 4 times per revolution), however the SGC/G signal appears as a constant 0v no matter how many times I rotate the CAS. This definitely appears to be the problem to me. I'm just wondering why the car would sometimes start, and then die shortly after, when now I can't seem to get any signal at all. Maybe it was on it's last legs, and the fact that I removed it broke it completely? Maybe at the mid 50s temperature outside it worked, but bringing it into the 70 degree house caused enough of a temperature rise to stop it from working? Anyway, when the new CAS arrives I'll put it in and I'm sure it'll probably start right up. I'll make sure to report back here once that happens.

Joe Perez

iTrader: (8)

EDIT: I'm retarded. See full explanation in post #31.



Haven't looked at the logs, as I can't find my license code for MLV, but the note of a "click" makes me think that this is not a CAS problem. Sounds more like power supply issue. (Keyswitch, main relay, etc.)

Is the processor resetting every time this happens? IOW, do you see zero RPM in the log during the outage, or does it just skip over the event?


EDIT: just looked at the raw data in Excel. Do you have a full log of the event happening, rather than just a composite trigger log?

bob13542

I believe the stallissue.msl contains a full log of a time that the clicking and stalling issue occurred. I thought it was a little weird that there was clicking too, although after removing and testing the CAS it does appear to be bad I think. I wasn't seeing any voltage from one of the signal wires, when from my understanding it should pulse between 0 and 5v twice per rotation. It is possible that there is another issue with the car as well, as I've had countless things die over the month and a half of owning it.


As for seeing zero rpm, the tunerstudio log definitely reports 0rpm as it stalls, and at the times that it would restart itself instantly, I'd see the rpm dip to 0, and rise back up. I also see the lost-sync counter increase whenever that occurs. I remember that while driving 40mph originally, while the rpm would drop to 0 the speedometer would still show proper speed. Not sure if the speedometer is based on a cable, or electrical.

bob13542

Here's a tiny bit more background information, a few things I thought of that might be related, but I'm not sure.
-A few weeks after getting the car, I started having an issue getting it to start. When I'd turn the key, the car would make an audible click, but not attempt to crank. I replaced the starter, which seemed to fix the issue for a little while. A week later it started doing the same thing, so I tested the battery. The battery failed autozone's battery test, so I replaced it. A few days later the problem returned. It would ocasionally click and not start, but I always could get it to start by trying several times. During the megasquirt install, I found the remnants of the old alarm system, and removed that. I haven't had the click and not start issue since, however everything was only working for 3 days before this happened. The click I heard while driving when the rpm would drop to 0, sounded very very similar to the click I was hearing when it wouldn't start or even attempt to crank. I am not sure if this is relevant, because a click could be any relay. The part of the alarm system I removed included what was most certainly a immobilizer relay that was spiced in between the starter cable coming off the ignition switch. I think it was that relay I was originally hearing when I would get the click but no start/crank issue when I first got the car. That relay can not be causing the current issue now as it has been removed, and the car ran perfectly for the 3 days after it was removed, and before this issue occurred.

Joe Perez

iTrader: (8)

Yes, I see that now. Finally dug up my old MLV license (it's sad that I've forgotten what all the file extensions mean) and I see valid data from the MAP and CLT sensors during the event.

Based on that, I rescind my previous comment. This is not a processor reset or a power-loss event.

Still, the clicking bothers me.

This is a 1.6 NA, so the fuel pump is normally controlled by a switch closure in the airflow meter. Have you removed that box and / or wired the fuel pump relay directly to the MS? If so, that would explain the click.

Edit: The MS is in cranking mode during the event. This is damned odd, as that should be impossible with 0 RPM.


Edit 2: I'll note that a similar problem, which was in fact CAS related, was what drove me to purchase a lathe and fabricate a crank trigger wheel about 10 years ago, as chronicled here: https://www.miataturbo.net/megasquir...ss-last-20241/

Sadly, most of the pictures in the thread are lost to time. This is what the final solution looked like:




Note: I do NOT recommend that you do this. This was the extremely hard way to solve the problem, but I was bored at the time.

I'm leaning back in the direction of your original suspicion.

bob13542

That looks really cool! As for the airflow meter, it has been removed. I now realize I forgot to mention about that. The MS install guide I followed mentioned to replace it with an air intake temp sensor, and the vacuum line to the MAP sensor in the ECU. It also had me remove a fuse from the fusebox, the ST-SIGN one if I remember right. From my understanding removing the fuse was supposed to disable the AFM fuel cut switch, and the two other sensors take the place of the missing MAF sensor.

However, considering the car ran fine for 3 days, I'd think it'd be weird if it's the fuel pump cutting out. because of the missing AFM.

But yes, the clicking seems really weird because that makes it sound like more of an electrical possibly power related issue rather than a sensor. I'm starting to wonder if the CAS may in fact be bad, but not be the only thing that's bad.
Thanks for taking the time to check out the logs, I really do appreciate it. I knew how to take logs, but even after looking through them several times I was still having issues using them to find out what was wrong.

bob13542

One more update. From what I can tell, the internal electronic components of the 1990-1993 CAS are available online, under the name J925 Auto Ignition Module for T1T49171. I already ordered a full CAS, so I likely won't be testing the process of rebuilding the CAS, but it does look possible, and relatively easy. The module is held in by a handful of screws. I've seen this ignition module available for $20-40, however it has a relatively long ship time, so going that route would result in more downtime than ordering a full CAS off eBay with quick shipping.

Maybe someday I'll order the parts and rebuild my old CAS, and test it to see if it works.

Joe Perez

iTrader: (8)

Note sure if it's appropriate to thank me. All I've done so far is to say "I think you're wrong. No, wait, I'm wrong. This is weird as hell and I really have no idea what's going on here."

The fact that the MS thinks you're going into cranking mode during the event, despite a 0 RPM indication, is weird. It's computing a fuel injector pulsewidth and everything (though I note that the PW changes throughout the event, and PW means nothing if no injection events are being called for, which we can't see in the log.)

When it comes back, is there a *pop* as though a bunch of unburned fuel in the exhaust has been ignited, or is it just an "Oh, the engine is working again" thing?

bob13542

As far as I remember, I didn't hear any loud pops. It definitely didn't sound like the popping from hitting a rev-limiter or anything. It just felt like a sudden loss/regain of power, like the car stuttered for a second. The only real sound I heard was the clicking.

This is definitely very weird. I'm just hoping the CAS fixes it, as I want to be able to drive it again haha. Only so many more nice days left this year before it starts getting cold and rainy again.


This car definitely has a lot of issues lol. The engine is a early 90 so it has the short nose crank (no wobble yet thankfully), and has close to 220k miles. The transmission is one of those early ones that's prone to get stuck in reverse (already happened once). If I had known all this was going to happen from the beginning, I probably would have picked a different Miata, but at this point I have so much time invested that I'm dedicated to getting this thing working and reliable again. I just tell myself "hey, there are only so many things that can break. Once they're all replaced, hopefully it'll stay running for good". Besides, all of the NA generation are getting old, so who's to say I wouldn't have had similar issues with other Miatas if I had just bought a different one.

Joe Perez

iTrader: (8)

Well, that's just confusing-er.

Also, just in case you are in the mood to be mildly unsettled, that last reply was written by a semi-naked man sitting on the one and only toilet on the entire 93rd floor of the John Hancock tower, thinking that maybe he shouldn't have gone quite so heavy on the crushed red pepper flakes in the previous nights' dinner.

It's a constant struggle between "I wish this dish had a bit more flavor" and "I wish it didn't hurt quite so much when I poop."

Because, you know, we believe in full disclosure here at MT.net.

bob13542

I only have two things on my mind now. Firstly I haven't even been on this forum a week and I'm already loving it. Secondly, who the hell designs a tower with only 1 toilet on an entire floor?

Joe Perez

iTrader: (8)

John Hancock is a weird building. The lower half is all commercial office space, the upper half is all luxury residential condos, and the upper-upper fifth is all machinery, broadcast and public-service RF equipment. From 93 up to the roof is mostly just industrial space which is rarely occupied by humans, so there's just a single stall for the whole floor.

(Yes, I'm aware that that's 120%. It's a tall building.)

If anyone here finds themselves already past the two layers of security needed to get to Floor 16, and has a forged access card which allows them to operate elevator car #4 to get from 16 up to 93+, and also finds themselves in urgent need of a restroom, the access code for the door to all restrooms from 93 and above is 1-3-5.

Marioshi

That's one secure *******.