3k to 4k rev jump
 

1991 1.6L, automatic, intake, exhaust, AFR gauge, stock ECU.
So I'm experiencing a really weird tachometer fluctuation that seems to be new. When I'm warmed up, OD off, and I really get on the throttle, my revs increase in a normal predictable way up to 3k rpm, spike from 3k to 4k way faster with no additional throttle input, and then after 4k stay normal again. They also seem to bounce around in that range. When I'm on the highway cruising, they'll do the same thing but again only between 3k and 4k, everywhere else my tach is rock steady, but it wavers and bounces in that range. My oil pressure seems to waver with the tach. I know some people have dummy gauges, but I'm pretty sure my 91 should have a real oil pressure gauge, which makes me think that the engine rpm is actually wavering BUT, I don't hear a discernible exhaust tone difference, so that makes me think the RPM isn't actually wavering and it's a gauge thing. My AFR's seem to be all over the place when this happens too, bouncing between 14.1 and 15.3 while steadily cruising, so I know the ECU searches, but that seems to be an overly large range. I'm still on the stock ECU too, no changes other than wiring my AFR gauge in, and my AFR gauge works normally in every other RPM range too.
My initial searching led me to think electrical, so I checked all the connections and grounds I could find, none seemed bad or corroded, but I cleaned them up regardless. Currently I'm wondering if the AFM flapper is just loose and bouncing around, that RPM range coincides with my worst NVH frequencies, so I was thinking the vibrations could be moving it. Other than the RPM change there don't seem to be any other differences, there's no noticeable dip or gain in power, no unusual noises or anything. Does anyone have any ideas? I've attached videos of my gauges that I think accurately illustrate the problem. I don't think the OD has anything to do with it. It does it less with OD on, but that's because the car just shifts earlier so it doesn't reach the problem RPM except while highway cruising.

You have your helmholtz resonator deleted.

Yeah, but I've had that deleted for 6 months or so now, and the problem is a few weeks old at most. Also my intake should act like the ARC intake with the larger center chamber that allows the pressure waves to destruct out the negative effects of a straight tube. Additionally I've only heard of a 3-4 ft/lb torque drop as a result of a helmholtz delete, never an RPM spike, and definitely not wavering RPM. And my car didn't have any noticeable change in power while these fluctuations were occurring, if the RPM was actually bouncing 5-800 RPM on the highway, surely I would feel bucking of some sort? I don't discount that my intake my have an effect, but I'm doubtful it's due to the resonance of the intake front. I should also theoretically be feeling a smaller fluctuation around 1500-2k and 6k+ because the same frequency will also result at 1/2 and 2x the RPM, but neither are present. Tell me if I'm wrong.
Edit: I assume you're talking about the intake helmholtz, my exhaust never had one

RPM isn't actually fluctuating that much or we'd be able to hear it. Gauge is reading incorrectly.

So back to electrical? Where all should I check? There's the 3 engine bay grounds. Maybe grounds under the dash? Connections to the cluster? Does that fact that it only happens at a certain RPM point to any specific place? I'm assuming it's like a 0-5v signal or something and it's having some weird issue carrying the voltage in the middle of the range? Electrical is not my forte.

It's going to be an AC signal not 0-5 volts like you get from a typical sensor. To measure it properly you'll need a O-scope.

Signal for the tach at the cluster is yellow/blue pin 1H.

You can get it in the Diag connector as well.

that signal originates in the ignitor module

Don't have an oscilloscope, any way around that? where would I even be checking? signal at tach? further down the line? What is a "good" reading? should be 12v pulse, am I looking for pulse frequency?
Pulled the ignitor, leads look fine, no sign of burnt out wiring or worn out connections, connection is firm. Thought I could disassemble and check the board for burnt out bits, but apparently it requires dremeling it open, so no, rather just pick up a new one if we determine that's the issue.
Could I utilize someone's spare cluster for testing? Any issues with mine having everything in different units as far as what the ECU wants? I suppose it turns the needle all the same regardless of what's on the faces.
per this thread: https://www.miataturbo.net/ecus-tuni...ecifics-58571/ It looks like the tach signal comes directly from the ignitor without going elsewhere. Would my ECU be seeing a bad RPM signal, or is this just the tach itself? Honestly if it's just the tach, then fuck it, I just listen to the engine most of the time anyway, the erratic needle is what caught my eye.

Not really any way around the O-scope in your situation IMO. You can use a DMM to see that signal is getting there but that doesn't really help you as you need to see if the signal is fluctuating erratically. I guess you could wire in an aftermarket tach to confirm the fluctuating RPM signal.

If the igniter is the issue and not the cluster then the ECU should be getting the same janky signal. If it runs and shifts fine then it's probably not worth worrying about.

Alright, ignore the problem it is then haha
I'll see if somebody local has an O-scope I can borrow. Would I be checking this signal while driving? I assume I'd be looking at a somewhat linear frequency increase then when I hit the 3k-4k range I'd be expecting to see large fluctuations? This would confirm that I'm getting errant signal, meaning bad igniter? Could this be done with the car in the air, or do I need to be moving? Haven't tried to replicate the problem in the garage yet.

Square wave. Should look like that. If signal from the igniter is the issue I'd expect erratic waveforms at the problem RPM. If the tach does it's dance in neutral then test in neutral. I wouldn't really want to spin things to 4k in gear on jack stands.

Basic O-scopes sufficient for pretty much anything automotive/audio are super cheap if you assemble yourself. I've built 2 of these and have no complaints. Not currently available from a US warehouse so shipping will take forever.

https://www.banggood.com/Original-JY...r_warehouse=CN

Nice, that was super helpful. I'll see if neutral works, that would be ideal. So I'd be tapping into the 1H and ground going into the tach connector I assume? Would any ground point work, or do I need to be on the tach's ground wire?
Soldering is beyond me honestly, but that is a great resource.
Thank you.

Edit: Also If I see the tach fluctuating but the signal is normal, this would indicate a bad cluster? or at least tach?

https://cimg2.ibsrv.net/gimg/www.mia...7d4785d482.png

Either 1H at the cluster, or IG- in the diag connector like Brain's illustration shows. The diag connector would be easier if it does it in neutral. Any chassis ground should be fine.

Yes, what you described would point to an issue in the cluster.

Oh, that's even more convenient. Could I run a wire from the IG- to inside my car and connect the oscilliscope to that so I don't have to tap my tach wire if I have to be driving? Would the wire length throw off readings from the resistance?
Seems like I have actionable things to check out, I appreciate the help. Will update with success or failure upon finding an o-scope.

A wire run inside should work fine. Wire length shouldn't change anything.

This is not true. On the NA, the ignitor tach signal ONLY goes to the cluster. The ECU derives engine speed via the CAS. They are totally separate. I recommendyou try to borrow a known good ignitor on FB or from a local club first since you don't have the proper diagnostic equipment.

You can get chinese 2 channel USB scopes on ebay for something like $50 or less that will do the job if you choose that route. I use an Owon 4 channel VDS3104 at home, which is more $s, but a pretty solid unit.

Oh, and there is a cap wired across the ignitor. I always assumed it was for noise suppression, but it could be tuning a resonance. You might A-B test that as well.

Would COPs and a resistor in the DIAG box fix this?

I assume he just forgot a "not" at least that's how I read it based on the conclusion that it's not an issue as long as the car is driving well. I'm trying to drum up and O-scope locally. Is it possible to fry someone else's igniter if there is an issue causing my igniter to go bad? Or is this just a wear and longevity issue for the igniter?
The little two wire connection mounted next to the igniter that doesn't seem to go into anything?

If the issue is the ignitor, then COPs could fix it. If you have a 2 pin connector open next to the ignitor, your cap might be missing. The cap is a black box with a metal tab attached to the same bracket as the ignitor. Mazda would probably only needed a failure rate due to tolerance or predicted aging of like 1/10000 to add that cap to every car for this reason, so who knows. Maybe that is why it is there.

as far as breaking other stuff you connect, who knows. I would suggest your ignitor on another car to test that way, but if it is resonance then the coil packs provide the inductive component and you won't be able to get the same impedence characteristics with a different set of coils. I think it is unlikely that you would damage a cap if you tried that.

Here are pics of the cap and the bracket. My install has eliminated the bracket.



https://cimg8.ibsrv.net/gimg/www.mia...6bb833100f.jpg


https://cimg4.ibsrv.net/gimg/www.mia...3ede29aa8a.jpg