Forked from the "GSL-SE vacuum chamber question" thread per the advice of @LongDuck

For the most part, the car runs great, but every so often it goes into "stumble mode", no power, car basically limps to the roadside and stops. I've had it happen at low and high speed cruising - most times, it's been while the car has been running for awhile, so it could be temperature-related, but I've had it happen a couple of times after cold startup within 100 yards after leaving my house. After resting, the car always recovers and runs fine until the next time it happens. I've had it go 200-300 miles in between "stumble mode" instances - feels like a sudden lean mixture situation, based on my experience with other cars.

Over the past couple of years I've been going through all the systems:
1. Car is bone stock, other than I run pre-mix (previous owner removed the OMP). ~85K miles. Dowel pin leak about 5 years ago, engine was rebuilt/resealed by Banzai Racing. Banzai stated the engine internals were in great shape, thought it was probably the result of running pre-mix, only needed to be resealed.
2. New OEM fuel pump and filter, dropped the tank, very minimal rust in a couple of places, no debris in the pickup screen. New rubber hoses in the fuel line from the tank to the intake (except the two under the dynamic chamber). Fresh fuel/pre-mix/stabilizer.
3. New igniters/thermal paste, new coils. Plugs/wires were done in the not too distant past.
Still had the problem occur last fall after the above work.
Currently going through the intake-related items - cleaning throttle body, new pressure regulator and pulsation damper, replace any bad vacuum hoses, sent out injectors (ND units, don't think they are OE) which tested/balanced fine. One of the fuel rail mounting bolts was finger-loose, but no other anomalies noted. Thermo sensors test ok, replaced the delay valve which was out of spec. Bench tested the BACV and both parts of it tested ok. I'll end up going through all the other fuel/emission sensor testing in the FSM once I have everything back together.

Searched the threads here and elsewhere over the past couple of years for someone else that might have had a similar condition, but didn't turn up anything that seemed to point out something I haven't addressed. Sorry for the long post, I would certainly appreciate any pearls of wisdom you might have related to this.

 

@t_g_farrell replied:

To me this sounds like ignitors over heating or failing. When this happens try swapping the leading and trailing ignitors. I know you replaced them but its the first thing that popped out to me about it.

It did to me initially, too, which is why they were the first thing I replaced a couple of years ago. By swapping the leading and trailing ignitors, assuming the problem goes away, what does that tell me?

 

Leading ignitor is bad if it follows the ignitor. Make sure the cpu paste is good and making good contact, that helps draw the heat away from the ignitor. Dialetric grease is not a good substitute as it doesn't help with heat transfer and also prevents it getting a good ground. I always use computer CPU paste or the paste that sometimes comes with the ignitor. I don't have a 5 letter, so I'm just talking generally based on your description.

 

@LongDuck replied:

Possibly Ignitors, but I'm thinking Fuel Filters (3 of them). Any rust AT ALL in the tank is going to be a problem.

The outboard fuel filter by the pump was replaced a year or so ago, along with fuel pump (which came with a new inlet screen obviously). Not remembering a 3rd filter - the previous owner had hacked in a generic NAPA filter between the tank outlet and the pump, I removed that and replaced it with a 100 micron stainless steel cleanable unit - debated about whether to go back to "OE" and just bypass it with the rubber hose, but I figured the added filtration between the tank and the pump might catch any rust. Haven't checked it, I will disassemble it and check to see if it has accumulated any rust. As I stated earlier, the tank looked pretty clean when I dropped it during all the fuel-related work. A few spots of rust, couldn't really check behind the baffles. I didn't see anything in the tank pickup area or screen and the fuel I drained out was clean, no sediment.

 

Thanks - just to make sure I'm understanding, if I swap the ignitors and the problem goes away, the leading ignitor is bad? Otherwise it would be (might be) the trailing ignitor? I will also pull them and re-paste them. I only use CPU-grade thermal paste, spent a lot of years in the computer hardware engineering business. Anything particular about areas to paste/not paste? I just put a thin coat along the entire back pad of the ignitors if memory serves.

The other thing in the back of my mind is hidden wiring damage, possibly at the ignitor connectors or elsewhere. Didn't see anything obvious during the ignitor swaps, but maybe I need to do some resistance measurements of those leads.

 

Did some research, so now I understand that the leading ignitor is the one that will cause poor running if its bad.

 

So the idea is to see if the problem follows the ignitor or not. If it doesn't then you need to move up the line and look at connections. It could be a dizzy issue but given it has to be hot to happen, it smells like an electronic part somewhere getting over heated. The only other difference is this is a fuel injected rotary, so I would also be very suspect of the fuel delivery as well.

 

That's kinda why I was focusing on the fuel system. I would plumb around the add-on Fuel Filter that you've installed and see what changes - maybe nothing. The Fuel Pressure from the tank to the pump is low, but the volume is high. Any restriction upstream feeding the Fuel Pump, you're going to feel it. That's why Mazda put a plastic screen filter (*coarse) in the Fuel Line going to the Pump, and left it at that. The Fuel Pump has a metal screen filter just inside the Pump inlet tube to catch anything big enough to bind the Pump - but this doesn't restrict the flow much at all.

The Fuel Filter itself has a fine mesh screen in the inlet which is where you'll find small sediment like rust will build up and restrict flow. The first symptom is reduced power at the top end, inability to accelerate road speed - in spite of being able to accelerate up to that speed just fine - and does NOT show symptoms at idle free- revving, because there's no load. Once you add load to the engine, you add Fuel Injector Pulse Duration and any restriction of volume or pressure causes lean mixture and loss of power.

Beyond the simple things above, the SE relies on a Fuel Pressure Regulator in the form of a vacuum-assisted valve on the Fuel Rail which restricts the RETURN line after the Injectors to bump PSI at the Injectors from ~44 to 60+ psi under WOT. This vacuum hose is signaled from the RE-EGI intake box at the top of the engine and actuated through a solenoid valve of the Rats Nest. If the pressure regulator is bad (*or the vacuum hose going to it from the Rats Nest actuator), it won't restrict the flow, you won't see higher pressure at the Injectors, and the engine will starve at high RPM. Repeatable.

What you're seeing with intermittent onset is confusing, though. The ECU on an SE has no "Limp Mode" - it's not that sophisticated. Does the engine shut down entirely, or just run poorly? What happens when you shut it off and restart it immediately? How long do you have to wait before you can start it and it'll run properly?

 

Also, THIS is why I suggested starting a new thread. Every reply adds detail to help the algorithm suggest related topics. Scroll down, and have a read on these:

 

Thanks for all the analysis, LongDuck - some comments/responses:
1. 100 micron filter added between the tank and the pump. 100 micron is pretty non-restrictive, and the car does run very well ~90% of the time. No loss of power at all, pulls hard all the way up to 8000-9000 RPM. To plumb around the filter, I'll have to drop the tank and run a straight 1/2" hose in that line per the OE configuration. I'll keep that in mind as a step to take if the problem still arises after all the work I'm currently doing. Might even take the tank to get cleaned/sealed, because "might as well" with it dropped.
2. Fuel pressure regulator - I have installed a new one (aftermarket, OE NLA) as part of my current work with the dynamic chamber removed. New fuel pulsation damper (OE) as well. New vacuum hoses under the dynamic chamber, too. "Might as well" with everything under the dynamic chamber accessible. Got a vacuum chamber on it's way to me too, so I'll revert that vacuum circuit to OE configuration.
3. The "limp mode" is the engine runs poorly, doesn't completely die. Crappy power at all RPM, pedal to the metal with little to no acceleration, or jerky acceleration, but the engine still runs, technically. One time it happened I was on the highway, about 15 miles from home, and I was able to finesse it back to the house, even at highway speeds. I haven't done much testing with how long it takes to recover - one time, it went into stumble mode for a couple of miles, I pulled over/shut it down, it sat for about 10-15 minutes, and ran fine after that. Another time, I left my house, engine was still cold, 1 mile away it went into "stumble mode" for about 20 seconds, cleared up and then everything was fine for another hour's worth of cruising.
I fought a similar issue with my 1991 Z-28, which does have an actual "limp mode" with the spark control circuit. Stranded us a few times, only to fully recover after a short rest. Took me a few years to finally track that down to a faulty EPROM (MEMCAL) in the CPU box. Only people like us with OCD (Old Car Disease) would put up with such behavior. But - the end result is worth it. The GSL-SE puts a smile on our face, at least 90% of the time.

Thanks for all your feedback - at this point, with all the changes I have pending, I need to get the car put back together and run it for a while and see if the problem recreates itself. Maybe some of the myriad changes I've made have addressed it. If the issue rears its ugly head again, this forum has given me several things to try/check. I'll keep testing status updated here, and I certainly appreciate everyone's expertise that has chimed in with suggestions. People like you help keep the hobby alive. :-)

 

Well, here's what I think is happening - make of it what you will; under certain driving conditions, you're somehow getting fuel starvation to the Fuel Pump, which then cavitates and forms turbulence inside the Pump which in turn kills fuel volume and pressure. This results in your stumbling and lack of power, because the Pump isn't working effectively anymore. There shouldn't be any restriction to flow of fuel volume before the Pump, but you have 2 filters in line - one of which Mazda intended. There's a restriction, however small, with anything you put in line and it could be the cumulative flow restriction which 2 smaller restrictions create. Hard to tell.

OR,...

Your Fuel Injector electrical plugs are somehow losing connection to one or both Fuel Injectors, which oddly works itself out over time. This is a long shot, but the green plastic connectors get old and brittle and easily break, especially if installed and removed occasionally, like during maintenance. New electrical connector plugs are available from Borg-Warner and have to be spliced into the original harness. On that note, the OEM Fuel Injectors are 2×580cc, low impedance injectors. The simple ECU fuel map is for these specific injectors which are ONLY used on the SE in the rotary world. Later EFI injectors are either lower or higher volume, and 4× being staged to fire at different times. If you have different aftermarket injectors in there, though they test fine, they might not be flowing properly according to the designed duty cycle provided by the ECU map, but you said it otherwise runs fine (???). Again, this is a long shot.

Also worth considering...

The ECU uses the Trailing Coil firing spark as a timing signal for when to fire the Fuel Injectors. This is a white wire connecting to the bottom spade terminal on the Trailing Coil. If the Trailing Coil isn't firing properly - for whatever reason - 1) your tachometer will stop working, and 2) your engine will die completely because the ECU can't determine when to fire the Fuel Injectors = no firing at all. The next time it stumbles, try shutting the engine down and mmoving the white spade connector from the Trailing Coil to the Leading Coil (same position), and see if that fixes it. What you've done is to give the ECU the Leading Coil as its EFI signal, which removes the Trailing Coil from the equation. If that resolves it, you can start inspecting the Trailing Spark system (Coil, Ignitor, Lines, Plugs).

 

Regarding the fuel pump cavitation theory - one other thing just came to mind in case it might be related. I've noticed that pressure seems to build in the fuel tank at times. Removing the gas cap, I sometimes get a big pressure release. Similar to late-model cars with sealed systems - might be normal? I've also noticed on occasion when filling the tank, it will spit back through the filler neck and I have to operate the pump nozzle at less than full bore in order to avoid it. I replaced the cut and check valve in case there was some sort of evap blockage but it didn't solve the issue. I tested air flow through the piping from the charcoal canister back to the tank, and there is air flow but some resistance. Possibly a partial blockage? I can add that to the list of things to check.

The fuel injector plugs looked solid when I removed them for the recent cleaning, plastic connectors/spring clips were still intact, connections seemed tight. The injectors are Nippon Denso "green top", no part number available. Attached is the report from the recent cleaning/testing - the flow rates seem much lower than that OE ones you listed.

Once i get the car back on the road, if (when) I hit stumble mode again I'll try the coil wire swap (and the ignitor swap) and see if either helps narrow down an ignition issue.

Thanks again for all your analysis!

 

I think you're onto it with the most recent information about the tank not venting. If the fuel tank cannot vent to the Charcoal Canister, then suction will form in the tank which prevents smooth fuel flow to the Fuel Pump. Your symptoms of the tank spraying fuel out at you when fueling is key - it should not do that, and either your fuel vent line going to the Rollover Valve (aka, check & cut) immediately forward of the tank OR one of the connections to the Charcoal Canister are getting blocked at random and causing fuel starvation. This would also speak to the intermittent nature of your symptoms.

On Fuel Injectors, your test report shows a 15sec Dynamic test spray volume, which when multiplied by 4 to get to 1 minute of flow equates to ~430cc of volume, per minute. That's too low compared to the 580cc flow rate of OEM SE Fuel Injectors. While your Injectors tes as Low Impedance, which is correct - they're probably not flowing enough fuel and the engine is running lean, and you're pre-mixing,... which adds risk. The O2 sensor on an SE is only good for Closed Loop Injector Pulse Duration fine trim in either 5th gear or with Cruise Control engaged (*IIRC), so it will not adjust fuel volume at all loads and RPM based on O2 feedback. It's a very simple system, remember. But,... If those Injectors are working fine for you, who knows. Leave them alone, I guess.

If it were me, I would remove the extra Fuel Filter you installed and while you're there, double check routing of your vent lines to be sure they're clear and going to the right place, per the FSM. The Charcoal Canister is NLA, but a suitable replacement can be found through search. Honestly, it's probably not the problem, as it's just a trap to prevent vent to atmosphere so your garage doesn't stink like gasoline, but if it's clogged, it's clogged.

P.S. - you don't have to quote posts every time, as it makes scrolling more difficult for readers...

 

Got everything put back together - found a broken nipple on the purge valve, so fortunately the rats nest I bought for the vacuum chamber had a good purge valve as well. Checked the charcoal canister - I do get a small amount of air passage through the tank vent, so I'm thinking it's not plugged, at least not totally. Set the TPS and idle speed and took it on a 2 hour/78 mile run at mixed speeds with no issue. Still need to work a bit on the idle stability, but the car ran great. Filled it with 10 gallons of gas at the end of the trip, and no burp back through the filler neck. I'll update here again when I have good or bad to report, but hopefully the problem has been solved.

Thanks again everyone for your insight!

 

Well, unfortunately ran into stumble mode again. The car had been performing flawlessly for ~150 miles - today I took it on a 50 mile round trip, and about 10 miles into the drive at 60ish MPH, it started stumbling. Intermittent again, but I decided to get off the highway and take back roads the rest of the journey. On the back roads, around 30-40 MPH, 2500 RPM, it did ok - if I tried to step on it, though, it would go back into stumble mode, around 3000 RPM and above. Pull up to stop signs or lights, it dropped to 800ish RPM and idled fine, and would get going again with light throttle. Ran like this for the next 10 miles, then decided to pull over and try the igniter switch suggestion to see if that had any effect on the stumble mode. Stopped the car, switched igniter cables, tried to start the car - it would crank but not fire. Bunch of smoke from the exhaust. Swapped the igniters back to original, same issue - car cranked for about 30 seconds before finally starting, more smoke from the exhaust then the exhaust was clear once the car was running. Car was idling fine, a little high at 1100 RPM vs. 800, so I continued the next 5 miles to the destination. Car rested for about 30 minutes while I did my business, started up fine for the return trip, no smoke, and made the 25 mile return trip with the same intermittent stumble behavior. Light throttle, 30-40MPH, 2500 RPM, ran ok - under moderate or heavy throttle, sometimes it would stumble at 3000-4000 RPM, other times it would run up to 7000 RPM with no issue.

Still have more of the earlier suggestions to try - I'll update here with any news. Unfortunately I only have another 2 months of driving season before it will get parked for the winter.

 

I'm used to a 12a with a carburetor so I'm not sure on the SE... but I think you have to actually swap the igniters themselves not the cables that connect to them. I think just switching the connectors would give you really weird timing.

 

The ignition system on an SE is identical to any 12a car. To effect the Ignitor swap, the best method is to actually remove the Leading and Trailing Ignitors from the Distributor base using a flat blade or cross blade screwdriver (*noting that the Trailing Ignitor leaves you very little room to get to the 2 screws because of the Alternator getting in the way. For this, I use a screwdriver bit and a matching ~5mm(?) box-end wrench that fits the bit.). It's a good idea to mark your Ignitors using a whiteout pen or some tape, so you know which is going where.

Once the Ignitors are swapped, then observe behavior. An SE will NOT run without a working Trailing Spark (*Ignitor & Coil), as the ECU uses this as a signal on when to fire the Injectors. It also will not start or run reliably without a Leading spark. On an SE, Trailing Ignition is also what gives you Tachometer signal. Mazda designed this so if you ever lose Tach, you'll start poking around and eventually realize you lost Trailing Ignition which cleans up emissions and adds a bit of power.

If the swapped Ignitors fix the issue, your Trailing Ignitor is bad / intermittent (*now in the Leading position at the front of the Distributor base), and your Leading Ignitor is probably okay (*now in the Trailing Ignitor position). This is why I keep a spare Ignitor in the glove box...

As a simple, easy, roadside test - on an SE you can also just move the white lead wire for the ECU's EFI trigger from the Trailing COIL to the same position on the Leading COIL - though your Injectors will be firing out of phase, the car should start and run, indicating that the Trailing COIL (*or the Ignitor which ignites it) is bad. I've used this trick before to get home.

 

Thanks Toruki and LongDuck - I'll try the ignitor swap, and see what happens. The ignitors, coils, plugs, wires, cap, rotor are all less than 2000 miles old. Any guidance as far as applying the thermal paste to the ignitors? CPU heat sinks typically work best with a thin coat, which is what I did when I replaced the ignitors a few years ago.

 

A light smear is all it takes; recognize that Mazda didn't even include it from the start, so it was devised as a way to extend life of NLA parts by enthusiasts like us. I never have, and I live in the desert. I'm still running my original Leading Ignitor at 252k miles. Let us know what you find out,

 

Haven't switched the ignitors yet - still in their original positions. Checked the AFM per the factory service manual - everything seems to be in range, the "E1 <-> Fc" measurement when the plate is fully open measured around 10 ohms, spec is 0 ohms per the FSM.

Also had ECU checking on my to-do list, so I took the car for a short ride to warm up the engine, everything ran fine, did the checking of the ECU per the FSM. Found this "****" under the carpet, and figured out with Mr. Google's help that it goes where the red arrow points. Not sure yet what that component is, or the significance of the "****". The only questionable ECU measurements were:
Pins "B,D,S,T" (grounds): measured 2.1mV, 1.7mV, 18.4mV, 21.6mV respectively, spec is 0V. Might need to check wiring or chassis ground points.
Pin "N" (O2 sensor): measured 0.4V, spec is 0V
Pin "b" (Relief solenoid valve control unit): measured 8.83V, spec is approx. 12V
Pins "c,f" (Checking connector): Pin "c" measured 62mV, pin "f" was flaky, meter registering constantly changing voltage. Spec is 0V.

After doing the measurements, I started the car, let it idle for a few minutes, idle was good but it eventually stalled and I couldn't get it restarted after cranking for 20-30 seconds. I'll revisit it later to see if this behavior is temperature-related.

 

The "****" is just a filter. Stick it back on the stub.

The stub is the air port for the Altitude Compensator switch - that black, oblong plastic thing mounted to the ECU rail. It's function is to measure atmospheric pressure so that the ECU can compensate for altitude changes, both during high altitude starts, and also for trips that cover large variances in altitude like driving from the seaside up to the top of a local mountain. These don't go bad, because most of the time they're never in operation other than a reference reading for the ECU fuel map.

Point being, the Altitude Compensator isn't part of this problem. Keep looking, and none of the other electrical signals seem particularly worrisome to me, either.

You're diving pretty deep into the Electronic Control Systems; these cars are not that complicated. For most of the electronic signal systems - it either works, or it doesn't. If it were me, I'd be swapping the Ignitors to look at logical, functional systems.

 

I went back an reread this entire thread and this part caught my eye: (*you wrote) "Light throttle, 30-40MPH, 2500 RPM, ran ok - under moderate or heavy throttle, sometimes it would stumble at 3000-4000 RPM, other times it would run up to 7000 RPM with no issue."

This is almost certainly fuel starvation. If it's intermittent, then it's intermittent fuel starvation, likely caused by crap in the tank hitting the pickup sock, or could be the pump getting an obstruction of flow from upstream.

Did you REMOVE your extra filter you added? It's important,

 

I haven't yet, but I think the fuel level in the tank is now low enough so I can drain/drop the tank, remove the extra filter and replace the line. I'll also check the filters for any contamination, and probably see if I can send out the tank to be cleaned/sealed just in case there is rust past the baffle. I saw very little in the pickup area when I dropped the tank the first time, and nothing in the pickup filter, but I couldn't see past the baffle. I also have a set of OE injectors due to arrive this week - if they test ok, they would remove any questions about the under-flowing non-OE injectors that are in there now. Maybe tank contamination, extra fuel filter, and under-flowing injectors are all contributing to an intermittent fuel starvation condition.

The "failure to start" issue is new - could be flooding, but I don't get any raw fuel smell. Both times it has happened, the engine has been at operating temperature. I'll see tomorrow when I try to start it cold if it is temperature related. Could be a side effect of the work I've done on the engine fuel system, or maybe a sensor got disturbed during the DC removal/install.

 

Failure to start = Air / FUEL / Spark. We're back to fuel starvation. Getting the right Fuel Injectors in there will probably help. Be sure to inspect the plug connections that bring power to the Fuel Injectors, as these have small spring clips that can break the plastic connector and lead to injectors not firing properly.

I have some threads on that here if you search on my Username and "Fuel Injectors". Good luck, let us know how it goes.

 

Well, it didn't take long to go into stumble mode after the tank cleaning/sealing. Drove fine for ~30 miles, took it to get the state inspection done and the mechanic made the comment that he thought it was running rich just based on the fumes generated during the inspection. Left the inspection station, a mile down the road it just stalls. Cranked for 30-40 seconds, finally got it started again, and then rough running/won't rev past 4000 RPM but I did manage to get it home. Sitting in my driveway, it's idling at around 1200-1500 RPM, where it had been rock solid at 800 RPM for the ~30 miles that it drove well. This matches the experience I had with the previous drive before the tank cleaning/sealing.

On the past two drives, it does seem to be an issue that happens after the car has been running for awhile. I'm sure if I go out tomorrow and fire it up after it has sat overnight, it'll run fine at least for awhile. It's only in the 50s here today, so it's not a hot day by any means, but maybe it's some sort of "hot-soak" type issue. Could try swapping the igniters to see if the problem goes away, could swap in a good set of OE injectors for the aftermarket ones in there now and replace the electrical connectors. I'll also review the thread to see if there are other suggestions made that I haven't tried yet.