Engine misfire ...

[eight-six]

Guys,

Jut trying to get as big a picture as possible as to what could be wrong with my car.

The car is running dual Carbs, still utilises the stock ECU etc for ignition. It’s all set up and starts and runs and goes well some of the time. But ignore that for the time being.

I have started to suffer some random misfire which is almost like the car is hitting some sort of rev limit. Only happens in 3rd / 4th onwards at high RPMS – can be anything from 5k onwards.

I am trying to work out why. I have ruled the carbs out for the time being. They are properly set up so no issues there.

Main suspect at the moment is the ignition system. Everything is good. Except the igniter which wasn’t testing to the specifications listed in the Toyota workshop manual. So I have replaced that.

Can anyone tell me … would a dead igniter potentially cause this sort of issue? The fault is intermittent which usually suggests something electrical.

On another note, if the O2 sensor detected rich running would the ECU try and advance the timing to lean out the mixture? The O2 sensor is hooked up (just disconnected it now for testing) but I am just wondering if the ECU is trying to advance the timing at higher RPM which is causing the miss as its throwing the timing right out.

The O2 sensor is the only sensor that is still hooked up to the ECU which could adjust the running, MAP and TPS are disconnected.

[firehawk]

Hmm, I have a similar setup with 44mm Mikunis and the standard ECU for ignition. Do you have the TPS and MAP plugged into the wiring loom but not connected to the throttle shaft/vacuum?

I had problems with misfiring at first when I had the MAP connected electrically but not into manifold vacuum. The ECU didn't provide enough advance at above 3500rpm and I had fire spitting out both from the carbs and from the exhaust. The hotter the engine got, the worse the misfiring became. Providing vacuum, albeit only from cylinders 1 and 4, to the MAP sensor fixed the problem. The same problem also occurred when the CEL is lit due to missing sensors and the ECU falls back to base advance (~10deg BTDC as standard) only.

If the TPS is connected into the loom but just dangling there, the idle switch is always closed and therefore the ECU uses only an extremely simple advance map intended for overrun and idle. Locking the TPS to a slightly open position would put the ECU off idle and and into the actual advance map.

To get rid of the CEL and error codes, I have modified my ECU with a few resistors so that it sees a constant +20C intake air temp and a ~80% throttle input. It has no pin for O2 sensor and uses the idle mixture pot instead but similarly, the O2 sensor also has effect on the fueling only when on idle. Off idle, the ECU only considers RPM and MAP when it determines the ignition advance so static readings from the other sensors doesn't really matter. The ECU is happy as long as the sensors return a sensible voltage/resistance

[eight-six]

Hey Jani,

Thanks for the reply.

Well I am not holding my breath but I took the car out last night with the new igniter installed and also with the O2 sensor unplugged; and on the few pulls I did do I didn’t hit the invisible rev limiter. Car seems to pull cleaner and my butt dyno says the car is quicker as a result. Whereas before the power would feel quite linear to the redline – it now seems to have a surge once it reaches 4k, and the cams come into play. Engine note sounds keener as well.

Quote:Hmm, I have a similar setup with 44mm Mikunis and the standard ECU for ignition. Do you have the TPS and MAP plugged into the wiring loom but not connected to the throttle shaft/vacuum?

Both are disconnected. When I got the car originally the MAP wasn’t connected and the TPS module was connected to the loom but not doing anything.

Quote:I had problems with misfiring at first when I had the MAP connected electrically but not into manifold vacuum. The ECU didn't provide enough advance at above 3500rpm and I had fire spitting out both from the carbs and from the exhaust. The hotter the engine got, the worse the misfiring became. Providing vacuum, albeit only from cylinders 1 and 4, to the MAP sensor fixed the problem.

The engine seems to pull pretty cleanly. Discounting the magic redline the engine is a bit boggy lower down as to be expected with big carbs, and lumpy cams but once it clears (circa 2k) it just takes off and runs perfectly. Even better now I pulled the O2 sensor and changed the igniter. I am going to have a go at hooking up the Map and the TPS to see what it does to the running. I will customise a old fuel rail for an inlet plenum to take my vac reading from and see what I can do for the TPS.

Quote:The same problem also occurred when the CEL is lit due to missing sensors and the ECU falls back to base advance (~10deg BTDC as standard) only.

The CEL is on permanently with mine, but it doesn’t seem to affect how the car runs. I think what the guys in Japan did was to remove all sensors the ECU used to control the ignition so it’s pretty much locked out. The Diz has been notched forward 2 teeth and then set to max advance just to get the timing right as the cam timing is very advanced. Maybe that’s the ida the Japanese guys used, throw it onto the base map at 10deg BTDC and then wind in as much advance as possible?

Quote:Off idle, the ECU only considers RPM and MAP when it determines the ignition advance so static readings from the other sensors doesn't really matter. The ECU is happy as long as the sensors return a sensible voltage/resistance

But if the O2 sensor was broken and returning duff readings, or reading a rich condition at higher RPMS (as you expect with carbs) could that not influence the ECU and in turn the timing gets advanced? Are you sure O2 doesn’t have an influence on the running past idle? I only say this as I had a faulty O2 sensor previously on my old SW20 turbo and the car ran like utter crap (driving and idle) – which would imply that the ECU on the SW20 did rely on the O2 sensor to ensure proper running.

Even if its only minimal adjustments in the timing its enough to throw the timing out completely?

Cheers

James

[eight-six]

In case anyone need to know for the future I swapped the igniter which wasnt testing correctly and also unplugged the O2 sensor. Car seems to have regained its orginal punch.

[firehawk]

eight-six Wrote:The CEL is on permanently with mine, but it doesn’t seem to affect how the car runs. I think what the guys in Japan did was to remove all sensors the ECU used to control the ignition so it’s pretty much locked out. The Diz has been notched forward 2 teeth and then set to max advance just to get the timing right as the cam timing is very advanced. Maybe that’s the ida the Japanese guys used, throw it onto the base map at 10deg BTDC and then wind in as much advance as possible?

Ok - with the CEL permanently lit, the ECU is not providing any ignition advance and triggers the spark with the base advance only. So by rotating the distributor as you described, you set a lot more base advance, causing the spark to be triggered at a static angle of, for example, 32deg BTDC. This isn't a best possible solution but works reasonably well, especially on race cars which see mostly full-throttle acceleration. I know that several AE86 rally cars in Finland run this way.

When the CEL is not lit, the ECU either uses the normal advance map or the simple overrun map - depending on whether the switch inside the TPS is open or closed. The overrun map is pretty much a static ~30deg advance when above 2000rpm. The normal map is probably the best solution but requires that the ECU can detect all sensors and that the idle switch is always open. The ECU does set error code 11 but doesn't go into limp mode (not a fatal error, so CEL won't get lit) so it'll provide a variable advance throughout the rev range.

JMR_AW11's articles on the 4A-GE ECU secrets were quite helpful in making the ECU play nicely along with carbs but unfortunately it seems that they're no longer online..

Quote:But if the O2 sensor was broken and returning duff readings, or reading a rich condition at higher RPMS (as you expect with carbs) could that not influence the ECU and in turn the timing gets advanced? Are you sure O2 doesn’t have an influence on the running past idle? I only say this as I had a faulty O2 sensor previously on my old SW20 turbo and the car ran like utter crap (driving and idle) – which would imply that the ECU on the SW20 did rely on the O2 sensor to ensure proper running.

Actually, you're right - the idle mixture pot (and thus apparently also the O2 sensor as they use the same ECU pin) has effect when below 3600rpm and partial throttle (light to medium engine load). However, according to JMR's documents, it only has effect on the injector pulse width so the ignition should work normally even if the O2 sensor is outputting nonsense to the ECU.

[eight-six]

Quote:Ok - with the CEL permanently lit, the ECU is not providing any ignition advance and triggers the spark with the base advance only. So by rotating the distributor as you described, you set a lot more base advance, causing the spark to be triggered at a static angle of, for example, 32deg BTDC. This isn't a best possible solution but works reasonably well, especially on race cars which see mostly full-throttle acceleration. I know that several AE86 rally cars in Finland run this way.

Yeah, it’s obviously not ideal for road use as it means the engine is quite flat at anything other than full throttle and long hills are tedious, but then I only use the car for track days and drifting so not massively fussed abut how it runs on the road.

Quote:When the CEL is not lit, the ECU either uses the normal advance map or the simple overrun map - depending on whether the switch inside the TPS is open or closed. The overrun map is pretty much a static ~30deg advance when above 2000rpm. The normal map is probably the best solution but requires that the ECU can detect all sensors and that the idle switch is always open. The ECU does set error code 11 but doesn't go into limp mode (not a fatal error, so CEL won't get lit) so it'll provide a variable advance throughout the rev range.

I suspect that’s why all the sensors were disconnected then as it was easier to unplug all sensors and lock the timing out than try and connect up the stock TPS etc. Potentially though would the fixed timing route result in more peak power over the variable timing option – but the variable timing option resulting in better driveability?

Quote:JMR_AW11's articles on the 4A-GE ECU secrets were quite helpful in making the ECU play nicely along with carbs but unfortunately it seems that they're no longer online..

I will check the IMOC as I think that some of his stuff is saved to the forum.

Quote:Actually, you're right - the idle mixture pot (and thus apparently also the O2 sensor as they use the same ECU pin) has effect when below 3600rpm and partial throttle (light to medium engine load).

Ahhh now as the ECU isn’t taking a reading from the TPS that could be further confusing it? But then it would be getting an RPM signal.

I suspect the issue was the result of the igniter as another issues I have had previously (revs hanging occasionally – but not because of the throttle / bad starting and occasionally lumpy idle “seem” to have gone away)

As a further test I will connect up the O2 sensor and take it for another run to see if it recreates the problem so I know for sure.

Car is going to be on the rollers this Saturday so it will be interesting to see what sort of power its producing.

[F0Bman]

Check your timing, it could be way off and then check your jet sizes.