Happy 4th Miata tinkerers. Anybody successfully install a stand alone ECU on an NB Automatic Miata?

In June, I meticulously installed a Flyin Miata turbo kit in my 99 Miata with Aisin 03-70LE transmission. I pulled the motor and tranny, and installed new valve stem seals, a new head gasket, new oil pan seals and new front, rear and cam seals. Everything *looks* great.........

Everything works great, EXCEPT low speed throttle transition i.e., when you're at or below ~20 MPH with the shifter in "D" or "R", and get off the throttle, AFR goes to ~9:1 and stays there. The ECU can't recover and the damn thing dies every time. You can prevent this by driving with the shifter in "2" or "1" (retaining drag on the engine, unlike when you're in "D", when the engine freewheels on decel), or by quickly shifting into Neutral and fethering the throttle, but this is stupid; you don't retain an automatic so that you can start driving it more like a manual car....or figuring out how to work around bugs in your non-configurable piggyback controller.

I've asked Flyin Miata to help figure this out but all that's brought is a suggestion to advance the timing from 16-deg BTDC and turn the idle bypass out a bit, like it says in the kit instructions; I've restored factory 10-deg. timing and mucked with the idle air screw and it doesn't help. Note that I've smoke tested the intake system and the only thing that showed any smoke was the EGR valve, so I replaced it...despite the possibility that it's normal for it to smoke a little. Compression test shows 150-154 PSI.

Everybody tends to say, "Just install a stand alone" or, put a manual in it. I'm keeping my automatic. And the problem with a stand alone/ECU replacement: In the factory service manual (pg. 495) it clearly states that there's serial communications between the PCM (ECU) and the transmision control module (TCM). Nowhere does the manual describe what's in this serial communications or how it's structured, which likely explains why every advertisement for the MS3PNP says, "manual transmission only." In other words, whatever info is conveyed over Miata ECU pins 1K and 1N is a mystery...a secret.

If you have a turbo on an NB automatic car, what did you do to make the car as easy to drive as it was before the turbo?

 

Connect with Pat of patsmx5 fame on here. he may be able to offer some guidance on that combo.
He did a 99 with a megasquirt and an aisin automatic a while back. He moved on to a c4 transmission from there and eventually went full on race car with it. I did a 99 with a early jatco 4 speed automatic, but mine was mostly manualized valvebody. I've since moved on to c4 transmission as well.

 

I am about 99% sure you can just swap in an aftermerket ECU and things will work fine. I know I have had an NA8 customer do it, and I believe a few people who have done it in NB's on here in the past.

 

Background. My car was originally a 5speed manual car. When I converted to a mazda auto, I pulled the entire vehicle wiring harness, and swapped it to a 99 auto harness. Pedals, Trans computer, everything. Nightmare of work, harness changed from headlights to tail lights.

Then I kept the stock auto ecu and harness and trans ecu in place. I cut the wires to the injectors and coils, and MS3 PRO ran those. I patched into the crank/cam/tps and fed those into the MS3 PRO so it saw what the mazda ECU saw. I added seperate CLT/AIT sensors for the MS3 PRO. Mazda still ran EGR/EVAP/IAC, etc.

This way, I lost nothing on the mazda side, everything worked exactly as it should. And it worked perfect. Never had an idle dip or stall, not once.

Your problem could be several things. But here's what I'd do first to start troubleshooting.

Question zero, I assume before the turbo swap it drove fine?

1. Tie wastegate open on turbo, put stock intake pipe with MAF back on the car. See if problem goes away. If it does, you have an air leak.
2. Still have the stock narrowband installed in the stock location feeding the stock ECU? If not, fix this.
3. Did you touch the idle screw? You should not have to, it was already correct before.
4. Any codes? Have you checked to see if all the sensors look good on a scan tool?
5 What is the grams/sec showing on the MAF when it's fully warmed up, at idle? I forget what it's supposed to be (mazda manual has spec), but this is a critical thing to check and if it's not right, you have to fix this.
6. If you put stock coil pack back on, and new plugs gapped to .044, does problem go away?
7. Is the kickdown cable to the auto trans installed and adjusted properly?
8. Try pinching off/plugging off the recirculation hose for the BOV, does the problem go away?

 

Thanks, Gents. I do appreciate you taking time to respond. Hearing from anyone that's been successful with a turbo auto is appreciated. Re: your questions @patsmx5

0.5. before the turbo swap it drove fine?
Yes. Perfectly. This car's been in the family for over two decades. It's always been maintained, mostly by me, and has always run perfectly.
1. Tie wastegate open on turbo, put stock intake pipe with MAF back on the car. See if problem goes away. If it does, you have an air leak.
I might do this. I have pumped smoke into the intake until it came out of two places i.e., the cap I stuck on the MAF and the EGR valve. Out of desperation, I installed a new EGR valve. There is zero smoke coming from anywhere else, before it pushes its way out the temporary MAF cap.
2. Still have the stock narrowband installed in the stock location feeding the stock ECU? If not, fix this.
Yes. Brand new narrow band sensor in the first of three bungs after the exhaust manifold. Wideband in the second bung (for AEM gauge), and the second narrow band if after the cat (OEM location).
3. Did you touch the idle screw? You should not have to, it was already correct before.
Idle air bypass screw...not the throttle plate-stop screw. I've only messed with the bypass in accordance with the FM kit instructions...with no avail for this persistent rich-on-low-speed-lift-off problem.
4. Any codes? Have you checked to see if all the sensors look good on a scan tool?
I've checked for codes. None. Short term and long term fuel trims are in single-digit percents. Although I haven't systematically checked other sensor behavior; I'll do that.
5 What is the grams/sec showing on the MAF when it's fully warmed up, at idle? I forget what it's supposed to be (mazda manual has spec), but this is a critical thing to check and if it's not right, you have to fix this.
I'll check that. Haven't...yet, partly because the car runs great under all other conditions. I'll report back regarding what I see.
6. If you put stock coil pack back on, and new plugs gapped to .044, does problem go away?
I might try this if other things don't work. For now, I'll open the plugs up from their current 0.28" gap. and see what happens.
7. Is the kickdown cable to the auto trans installed and adjusted properly?
That cable is exactly where it was prior to installing the turbo kit...and the car/trans works great under all other conditions e.g., cruise, stomping on it, decel above 20 MPH, decel below 20 MPH while in first or second shifter position.
8. Try pinching off/plugging off the recirculation hose for the BOV, does the problem go away?
I'll try this too, and report back.

As it stands...if I can't make the car behave just like it did before the turbo kit - upon decel in "D" (or "R"), at cruise speeds of 20 MPH or below - I'm likely going to install a Speeduino Micro for fuel and spark, inspired by @patsmx5 experience/post, which I appreciate. The Micro isn't much bigger than the Voodoo Box, and I doubt anyone would notice the difference...even if they looked under the dash. And since OBDII port would still be there and active, I suspect that would be just fine.

 

I checked again for vauum leaks. MAF signal shows 4 g/sec at idle and 27 at 3,000 RPM in neutral. AFRs hold pretty steady at ~14.7:1 at both RPMs. I can't find any vacuum leaks and I currently have the timing set to factory specification (10-deg). This is puzzling. I have a theory, and it's nudging me closer to 'parallel Speeduino'.

What we know: The Miata's factory "PCM" (ECU) expects a specific relationship between MAF readings, throttle position, RPM, and manifold vacuum.... from its normally aspirated engine. With the longer intake path and turbo, this relationship has fundamentally changed. Even at low speeds, the turbo is spinning. While the wastegate is set to 8 PSI, even below at slow cruise speeds, environment inside the intake manifold is closer to atmospheric. And the MAF will be reading more air mass than the stock NA tune expects for any given RPM and throttle position. I think the stock ECU is not able to correctly interpret these conditions, and if we believe Flyin Miata, the Voodoo Box isn't doing anything under slow speed transient conditions. Assuming this is true, the stock ECU is still injecting fuel based on its NA air density models.

The ECU, already confused by the higher airflow from the turbo, receives a delayed MAF signal as deceleration begins. This makes the ECU's fuel cut-off strategy less accurate, leading to more pronounced rich condition. The ECU's delay in recognizing the reduction in demand, combined with the MAF reading residual air in the extended intake, exacerbates the temporary rich condition.

I think it's important to note, too...that I moved the front O2 sensor more than a foot further down the exhaust system. The car had a CA two-cat exhaust system on it; the front O2 sensor was right up by the head. Moving it farther down onto the Flyin MIata downpipe introduced another delay (latency) in the feedback loop. Now, the ECU gets its O2 signal later...and so it's slower to react to changes in AFR. If the engine momentarily goes rich, it will take longer for that change to register at the displaced O2 sensor and for the ECU to make the necessary fuel trim adjustments. I think this increased delay causes the ECU to "overshoot" its corrections. I think this is what's causing the significant and prolonged rich condition during the initial phase of deceleration...and without the load from the rear wheels on the engine, the ECU doesn't have time to recover.

If this theory is sound, I think it means: "let a Speeduino control your injectors based on your wideband". If further testing yields anything interesting, I'll share it. And if I install a Speedunio, I'll let you all know what happens.

 

This is a problem. I spent 15 min looking this up, but here's what it should be. From mazda, With an automatic. MY = 1999.


So for whatever reason, you are wayyy over-reporting on the MAF, which perfectly explains the rich condition. Install the factory intake and see if the problem goes away. If it does, air leak. If not, bad MAF likely.

None of this is correct, except that the vodo box is out of the pic.


Moving the sensor is a bad idea, but it's not the cause of your problem. You should fix that once you fix the MAF/air leak problem.

I had a bad MAF one time that reported a value in error. I bought it, and it was a used MAF, but I suspect it was for an 01 not a 99, or it was just bad. But eventually I looked into the g/s (like I recommended here), saw it was reporting a value out of spec, replaced it, and problem went away.

 

I would seriously consider trying the first suggestion pat made to tie open wastegate and put stock intake and maf configuration to test.

That would not only show if leaks, which you say you don't have, But would show if there is an issue with the blow off valve return location on the flyin miata intake. That location may be too close to the maf sensor. If when you close the throttle and have some turbo speed there may be some feedback from the bov return hose onto the back off the hot wire sensor tricking the maf sensor into thinking there is flow into the engine, when in reality it could be backflow through the bov outlet. I'm just making a wild guess from looking at the pictures and the angle and proximity of the bov return hose to the maf sensor. I could be wrong though as its just a quess from pics. If that turns out to be the issue the fix could simply move the bov return closer to turbo inlet.

 

My apologies; I made a stupid math error...converting lbs to grams. Here's the MAF results, within appropriate proximity of the factory manual specs.

0.008 lbs = 3.6 grams

0.022 lbs = 10.0 grams

I spoke with a PhD E.E. earlier. Not a 'car guy', but he's quite broadly, deeply intelligent and tends to be "hands on". His theory: the compressor side is an obstruction, and when the throttle snaps shut, not enough air gets through (as assumed in the OEM tables), causing the rich condition...and since the ECU only has a narrow band to work with, all the ECU knows is that AFR isn't "stoich", and it can't "hunt" fast enough to find it again. In other words, it's just a low speed transient condition -- caused by the presence of the turbine -- that the Voodoo Box wasn't designed to handle on a car that "freewheels". If this is correct, I suspect manual shift cars display similar behavior for people that tend to push in the clutch as they come to a sfop....rather than down shifting. Anway...that's all I can do on this thing for another two weeks. I'll report back if/when I have anything interesting to report. Thanks again for your collective suggestions and feedback. I appreciate it.

 

I give up. But I suggest you reread my post.

 

I read your post closely. A smoke test doesn't show any smoke except where we expect it e.g., from the cover on the MAF, when smoke is pumped into the intake. And I looked at the service manual as well. I can see that MAF spec is 2.4-3.4 at idle and 8.3-9.6 at 2,500...The MAF on my car is showing 3.6 at idle (about 6% out of range) and 10.0 (about 4% out of range) at 2,500. I mistakenly indicated "27" in my prior post; that was wrong/my mistake. The observastions shown on the scanner screenshots don't look like a "bad" MAF in a cool, humid area of coastal CA... but then I don't have enough experience with MAFs/MAF meausrements to know. I may smoke test it again, to confirm again that there's no air leaks...but nothing's changed except the installation of a new EGR valve. And if you're suggesting that 10.0 instead of 9.6 at 2,500 RPM and 3.6 instead of 3.4 at idle means, "replace the MAF", perhaps I'll go buy a new MAF sensor...after applying some of this MAF knowledge and guidance.

Also, the BOV is plumbed exactly as specified in the FM kit. There's an internal aluminum 90 that points any return air downstream, away from the MAF. Perhaps I'll ask FM support if they've ever seen/heard of a problem with where it returns air at sub-20 MPH speeds upon the throttle closing.

 

Try Pat"s suggestion #1 or at the very minimum his suggestion #8 from post #4. The suggestion #8 can be tested easilly with no real changes to anything other than blocking/pinching the bov return. Some of the The back flow from the bov return being so close and angled towards the back of the maf could be reaching the maf causing a false rich condition

 

Easy, peasy. I'll give it a shot @Newaza Enjoy your week.

 

Yeah, piggybacks can't fix that decel stall, it’s a known issue on boosted autos. I had to run a split setup: factory ECU for the trans, MegaSquirt for fuel and spark. It’s not plug-and-play, but it works. Standalone ECUs don’t talk to the TCM, so full replacement kills auto control. No clean fix yet, just workarounds.

 

I've come back with my tail between my legs, to a degree. Previous readings with the original MAF were 0.008 lbs (3.62 grams) at 800 RPM and 0.022 lbs (10.10 grams) at 2,500 RPM i.e., indicating higher than the high end of the factory service manual-specified range, at both specified RPMs. Of course, from the ECU's perspective, more air means more fuel please....the cause of the problem I'm experiencing. So, upon @patsmx5 advisement, I shot the parts cannon and bought a new MAF. The readings with the new MAF are 3.18 grams (12% less i.e., Better) at 800RPM and 8.16 grams at 2,500RPM.

I also adjusted timing back to 10-deg. BTDC (Flyin Miata recommends retarding timing by six additional degrees) and, I turned the idle-air bypass screw out 1.5 turns from where ~900 RPM is achieved, with the GND and TEN terminals connected in the "DIAGNOSIS" connector (i.e., when the IAC valve is inactive).

Altogether, the super-rich-upon-throttle-lift problem is significantly less. It still occurs, but under a narrower range of low-speed, low-load conditions. Most (not all) of the time, the ECU tends to recover from the rich condition. The car is minimally drivable now...but it does die occasionally if the throttle isn't gently opened when the rich condition occurs. The car isn't as easy to drive as it was before the turbo kit was installed, and that's annoying considering about $6K was spent on the kit and various other parts-replacement during the install.

Next, I'll try pinching off the BOV return line. Perhaps that'll improve things even further. And I ordered a Speedunio. That's going to live in parallel withe factory "PCM" (ECU), and it'll control the injectors and timing.