Spark Plug Problems!?

[Ian M]

So we have been running an RR engine this year. In the past I have tuned 2 years of F4i engines without too many problems. We are using a Motec M4 ECU.
Our problem this year seemed to be fouling spark plugs, especially on cold starts.
I know this would usually point to running the engine too rich, but I have been leaning the engine and leaning the engine all year / summer and we are losing power and the engine is starting to run rough, especially on the low end, not wanting to pick up. Even when leaning the engine out so much that it heats up rapidly and runs like crap backfiring blue flames, when the car is shut off at any RPM and the plugs are immediately checked, they are always rich with soot.
If the engine is warm and fuel is added to where it makes power and runs smooth (where we normally run it) the thing will start right up without even touching the throttle.
I originally thought the cold start was feeding to much fuel, so I basically shut it off and manually fed it fuel until it tried to start, but if I give it enough fuel to start it smoothly (~25% overall trim at 70F) than if it doesn't start right away the plugs will be fouled and require cleaning.
We are running our dwell time at 3msec, which is what I used for the F4i's which have almost exactly the same coils.
Whatever the problem is seems to be getting worse, so I am wondering if maybe we have low compression (going to do a compression check sometime). Maybe rings or improper seating exhaust valves.
Or would you be more likely to expect a weak spark maybe caused by worn coils.
What is weird is that whatever is causing the problem is fouling all 4 plugs, which is what is throwing me.
Any suggestions or similar experiences would be much appreciated because this has been baffling me for quite some time. I am at a loss!

Intake, exhaust, compression, coils, Bosch ignitor, possibly retarded engine tuner... where to start?

[Jonathan S]

A few things come to mind, although not all explain why it it happens to all 4 plugs and only when cold.

- Bad gas
- Insufficient coil voltage
- Leaky/diry injectors

How many sets of spark plugs have you gone through? Are you sure they are the correct ones for the engine (gap)?

Do you have a lambda sensor?

[Grant Mahler]

Have you verified timing?

[Kirk Feldkamp]

I was going to ask that too... what is the ignition advance at in this problematic idle area? How about fuel pressure? Does the engine run well under other operating conditions? I've had some engines idle mostly ok on initial startup... but then suck everywhere else. I've been amazed how off some settings can be and still have the engine idle.

-Kirk

[Brian Evans]

possibly wrong heat range plugs, try a few steps hotter until you can start and idle the engine OK, then be careful when you start to make power runs. Also ignition advance, try a lot more advance as soon as it starts if you have that option (step the advance curve at 400 rpm or so). I've had some engines that need 20 degrees advance in order to idle, and I had one engine that I ran locked ignition at 44 degrees advance - it loved it! So no preconceptions about what it needs.

Brian

[Ian M]

Thanks for the suggestions! To shed some more light on the subject.
Yes, we do run and log a lambda sensor and have done many dyno / track runs and observed the data. I believe that I have a pretty decent fuel map built because the lambda reads 0.95 from 0 rpm - 4000 rpm (we idle at 3000) wants to surge any lower than that. Anyway, lambda never reads richer than 0.88 thoughout the whole rpm range. Am I correct that above 5000 rpm I should be shooting to keep the A/F around 0.9 lambda?
As far as timing goes, it is similar to how I ran all the F4i engines.
0-1000 about 15 degrees. Ramp quickly to 25 ish at idle. ramp to about 50 by 8000rpm.
Spark Plugs are the factory RR spark plugs.

Thanks for everyone willing to help with a strange problem.

[Grant Mahler]

I meant have you checked the timing advance with a gun to see if you are actually getting the advance you think you are.

As you probably know, there are many settings that could screw this up - checking your advance at several points on the timing map is a good way to make sure your ECU and your engine are on the same page.

Your map sounds like it would work fine, if you are actually getting that.

[Kirk Feldkamp]

In my experience, most engines will idle more happily at a lambda greater than 1. A Honda K20A I just finished working on was 110% happy at 1.15-1.20 at idle. Try leaning your motor out a bunch to see if you can get it to actually lean misfire. Richen it up until the misfire stops, and then adjust ignition timing to set your idle RPM's. If you're trying to get the factory recommended heat range plugs to stop fouling, then you're probably going to have to go leaner than 0.95.

Just so you're not worried, you will NOT hurt the engine at idle if you do this. It's only when you start to load the engine that you want to shoot for your lower lambda targets as you've described.

-Kirk

[neil trama]

It sounds to me like you are using the standard Motec "engine temperature compensation table".

You need to do your own table.

1) Go to "cold start" in the main screen.
2) Set "cold warmup enrich" to -1.
3) Go to "fuel" in main screen.
4) Go to "engine temp comp" and set up a custom table.

The original table will almost certainly be too rich during warmup.

Neil

[Tim.Wright]

We had a similar problem with our car this year. It kept fouling plugs and didn't respond to reductions in fuel.

Problem was traced back to a fuel pump failure. It was only producing a small amount of pressure, as a result the injectors weren't atomising the fuel, just dribbling it all over the plugs.

So it could possibly be a fuel pump, or reg problem.

Tim Wright
Suspension & Vehicle Dynamics
Curtin Motorsport

[Ian M]

Hmmm, the fuel pump never crossed my mind. That is the kind of strange story I was looking for though. I need to get a gauge to put in our fuel line and check it out.
It would seem that to make it run at all though, more fuel than necessary would have to be fed to the engine due to the lack of fuel pressure and I am actually pushing less fuel to this RR than I did to the F4i's. I am definitely going to check that out though, because that would explain the problem at all 4 cylinders and the fact that it seems to be gradually getting worse (aka. pump could be gradually getting worse, this is the 3rd year for this pump).
I am going to recheck my timing this weekend. I have already leaned it out until it lean misfired, I shut it off while running this lean and...the plugs were still dark and soot covered. I was like WTF!???
Son of a biotch, I can't believe I didn't ever suspect the fuel pump or regulator. The symptoms seemed like leaky injectors, but I couldn't explain why they all seemed like they were leaking onto the plugs.
Low fuel pressure and lack of atomization would also explain why the thing starts fine when the engine is warm (easier for fuel to atomize) and why it won't start worth a crap and fouls plugs when the engine is cold even with almost all of the cold start trim pulled out of it.

Thanks Timo, I suspect that you may have hit the nail on the head. I probably should have thrown a gauge in the line as soon as we started experiencing problems.

Turbo Twig,
I'll try to lean her up at idle some, but it seems that if I get over 1.00 lambda the engine starts to surge +- 500rpm? Any suggestions to fix this? Retard the timing a little, and/or make a bit of a hole where I want it to idle?

[Ian M]

Looks like it is actually the coils to blame. Shooting a very weak spark. Replaced one of the coils with a backup coil and tested the spark and there was a world of difference. You could also feel the discharge in the good coil and just a faint pulse on the bad ones.
Still not sure what cause all 4 coils to go bad. I am running a dwell time of 3 msec. I am wondering if we overjuiced them with our battery charger? Maybe the engine got too hot (unlikely)? Poor initial tuning? Not sure?
Investing in a small jumper pack instead of a full size battery charger when new coils are installed. Anyone else ever had any problems with the coils?

[The Gus]

I've experienced many problems with the F4i coils and 3ms dwell time. For some reason the coils were fine on the dyno and crap in the car after just a short while. I highly suspect the two different cooling systems played a large factor and helped overheat the coils significantly quicker in the car than on dyno (no cooling problems on dyno, but many in car: 84 vs 110+ C respectively...back then).

Since then I regularly check the resistance through the coils to help determine their condition. Measuring from the spark plug pin through either of the two connector pins I would always get 11.4 kOhm with new coils. Bad coils yield lower resistance, and completely burnt coils had no resistance. I find measuring hot coils had around 10.0 kOhm but letting them cool would return to 11.4 kOhm.

I read on here that many teams run between 1.8 and 2.2ms with the F4i coils. I've tried different dwell times on the dyno and compared power output. I did not notice any power changes using dwell times as low as 1.8ms. Below that I had problems running the engine – especially at low RPM (sub 4000). With the lower dwell times the idle suffered. I now run 2.2ms and there is still a noticeable difference between idle with the 3.0ms. I just retuned it for 2.2ms. . . .and keep checking the coil resistance. No problems yet though (unfortunately the cooling system on the car hasn't improved much).

Although this gets me running I would prefer to monitor the amperage through the coils and see how long it actually takes the coils to charge. . . And try this with different input voltages. I suppose at low engine speeds the voltage is lower and therefore the coils require more time to charge. This could explain the idle differences. But with a good alternator, battery and voltage regulator the voltage shouldn't change too much. . .

[Poe]

Gus, do you vary dwell based on voltage, or do you have it set as a constant?

[The Gus]

My ECU can only hold dwell time constant regardless voltage.

Unfortunately I don't have voltage compensation for fuel injectors either. I've considered making a circuit to supply regulated voltage to the injectors and coils rather than just using the battery (say 12V, or 15, 17,...whatever. Even if battery drops to 8V), thereby eliminating the need for battery compensation. Anyone else do this?

[Tim.Wright]

Originally posted by Ian M:

Thanks Timo, I suspect that you may have hit the nail on the head. I probably should have thrown a gauge in the line as soon as we started experiencing problems.


It only took us about... 6 months to figure that one out.

Tim Wright
Suspension & Vehicle Dynamics
Curtin Motorsport

[Ian M]

Yep, our M4 has the voltage compensation, but I haven't set up a dwell vs. voltage compensation table. I have just utilized Motec's default compensation.
So let me make sure I am thinking straight on this one...
The ECU can only control "discharge time" by cutting ground therefore sparking the plug. So, a longer dwell time means less charge time for the coils? The longer you charge the coils, the hotter they get and the more likely you are to overheat them (due to charging, not engine heat).
Is my thinking on the dwell time backwards?
So on idle you would need to charge the coils longer due to the lower battery voltage below 3000 rpm?

P.S. Our engine got hot once (101 C) and I bet that is what killed our coils originally. From then on the only time our engine got hot is when I tried to lean the thing way out to keep from fouling the plugs, apparently due to weak coils, which probably just made the coils worse!!!

[Ian M]

Gus,

Thanks for the coil resistance numbers. We were going to check resistance ourselves, but the RR shop manual did not provide resistance data. It said to check the coils using a peak voltage meter, which looked fairly expensive.
We will give our coils a quick resistance check.
We put a backup in and it sparked noticibly better and we did have one coil that seemed to be decent still on the engine, so it should yield at least a little resistance. I am anxious to check them now that I have a number to compare to. RR coils shouldn't be much different.

[The Gus]

The analogy I've been given by an EE (I'm ME) is to think of the coils like a compression spring. You apply a charge for a certain length of time (aka dwell time) to compress the spring, and the energy stored in the spring is the energy used for the spark. Problem is, eventually too much charging will "bottom out"� the spring. Charging (dwell time) beyond this will only overheat the coil. Not good. So it is important to find a dwell time that fully charges the coils, but does not overheat them. With 3ms you're probably overheating them (due to charging, not engine).

I wouldn't expect that reaching 101C once is going to seriously harm your coils. . .not on it's own anyways. But in parallel with a dwell time that is too large – that could be the catalyst that the coils needed for failure. But failure was likely nearby anyways.

The best way to find out the proper dwell time is to use an oscilloscope and monitor the current going into the coils. See how long it takes for current to drop to zero(ish). Charging beyond this amount of time is bad.

[Kirk Feldkamp]

Ian M,

Surging can be a common problem unless you have some sort of an 'idle pocket' built into your ignition map. Basically, you need to create a small depression (of relatively retarded timing) around the map sites on the map you want the engine to idle at. The RPM you get the engine to idle at will be a balancing act between your throttle plate low stop position and the ignition map. Retarding the timing relative to the adjacent cells will cause the engine to 'fall' into the pocket and keep the RPMs from taking off without any throttle input. There should be a 'ridge' of advanced timing below the pocket so that when the motor tries to drop below your desired idle RPM, the advanced timing will bring it up again. Also, at higher throttle positions and higher RPMs directly adjacent to the pocket, you should have a 'ridge' that helps to 'wake up' the engine up with changes in low throttle position. Be aware, however, that if your pocket isn't large enough, the motor could start to catch these higher adjacent cells and initiate surging. The motor will pick up RPM, then run out of air and the RPM drops... then get bumped up again, then run out of air again, etc.. Too large an 'idle pocket' will make low throttle position changes sluggish, so shoot for as small a pocket as you can, within reason. All of this idling business will require that your low throttle stop is very REPEATABLE.

If there are any factors that will make your low throttle setting change, then eliminate them. I built a TB a few years ago that was fine when I was using it, but the idle stop wasn't easily repeatable unless someone knew exactly how to adjust it. Once I graduated, it was no longer an easily servicable piece of the system. I've heard that in the ancient times of carbs, tuners used to drill a small hole through the face of the throttle plates for the idle 'setting'. This originally was to help stabilize the airflow into the carb when the throttle was suddenly snapped shut. I guess the normal 'slightly cracked open' low throttle stop was highly unrepeatable and messed with low throttle fueling enough to warrant the modification. I really like the idea because it's essentially impossible to screw up. Start small with your drilled hole, and work up until you get the idle setting you want. Then, all anyone has to do is 'zero' the throttle plate shut and you're ready to run.

-Kirk

[Poe]

The analogy I've been given by an EE (I'm ME) is to think of the coils like a compression spring. You apply a charge for a certain length of time (aka dwell time) to compress the spring, and the energy stored in the spring is the energy used for the spark. Problem is, eventually too much charging will "bottom out"� the spring. Charging (dwell time) beyond this will only overheat the coil. Not good. So it is important to find a dwell time that fully charges the coils, but does not overheat them. With 3ms you're probably overheating them (due to charging, not engine).

That's a very good analogy. I hadn't thought of it that way, but that might help the more mechanically inclined people understand.

[Ian M]

Yeah, I originally thought dwell time = charging time, but after thinking about how the ECU controls the spark I got to thinkng it was the other was the discharge time. I guess the ECU just does the math.

Turbotwig,
I have tried the idle hole with timing, but I was only trying 2-3 degrees and it probably wasn't quite big enough, which caused it to catch the top end of the hole and then drop back like you said. I didn't mess with it a whole lot. It like to sit at 3000, so I just let that be the idle, although I would like to see 2000.
We have an adjustable idle stop and it is pretty consistent, but we are working on a new throttle body for this year, and I am considering the hole in the butterfly idea. Just don't like the lack of adjustability. Guess you could always make another butterfly if for some reason you wanted to make your hole smaller to idle lower.
Thanks for all of the advice guys, this has been a very beneficial thread.

[TMichaels]

The dwell time is the charging time. The Motec switches the coils to ground for the dwell time. The best is, if you don't have any data, to measure, how long the dwell time is while changing the voltage. You can do this with a strong 5 to 20V, 30Amps powersupply. Monitor the current to the coil with an amperemeter and look at which point, after the voltage raised up, the current gradient begins to sink. The current will drop exponentially over the time. The time from the beginning of the coilcharge until the current gradient begins to fall is your dwell time. Usually this is between 3ms at 9V and 1.5ms at 16V. I think I have a graph somewehere showing this. I will try to find it.

[Kirk Feldkamp]

Ian M,

2-3 degrees isn't quite enough. On the Hayabusa I did recently, the idle cell was 20 (degrees of advance). The ridge to the left was 25. The adjacent cells above and to the right were 24. Then above and to the right of that is basically normal timing (28 and beyond). So you can see the pocket is small, but deep. Don't forget the balancing act between the idle adjustment screw and the ignition map.

Now that I have some high production design experience, I am hesitant to add adjustable features to components that don't necessarily require adjustment. From a production standpoint, adjustability invariably costs more, both in calibration and in parts. I believe if you do the proper development work with your hole sizing and tuning you can come up with a configuration that won't ever need adjustment. I'm starting to believe that some adjustable systems are only adjustable because the 'standard' adjustment method isn't reliable or repeatable enough to keep the desired function steady. In a sense, the adjustment is unnecessary because if the system were developed properly there would be little need for adjustment. Another way to think about it is a system can never be "out of adjustment" if there is nothing to adjust! Furthermore, it can never be adjusted incorrectly (think stupid noobs). Just something to think about.

-Kirk