Wasted spark conundrum - NOW SOLVED

[Steve Kirk]

My engine runs wasted spark ignition. In other words there is no distributor but two coils, and each coil fires two spark plugs at once (one off each end). Plugs 1+4 fire together off one coil, and do 2+3 fire together off the other.

 

So am I right in thinking that it should run equally well if I swap over HT lead 2 with HT lead 3 at the coilpack connection? (Or HT lead 1 with HT lead 4)?

 

Edited by - Steve Kirk on 5 Aug 2005 01:03:39

[F355GTS]

Steve

 

Yes i think so, try it, it won't do any harm if it doesn't!

 

Mark

[Peter T]

It depends on the trigger signal from the ECU, because yours runs an amplifier ignition module, it does not generate true pulse on the non firing cylinder, otherwise it tends to overheat the amplifier.

Most ecu have an internal channel that does pulse twice on ouput signals and being processor controlled it does not generate any excess heat or current.

Also tends to keep spark plugs cleaner by burning of any excess fuel particles.

 

Carbon Components from K9 Composites

 

[Steve Kirk]

Mark, thanks. My logic was the same as yours . The reason I asked is that I did try it. My #2 is misfiring even after the plug and lead have been swapped (as Peter has read on other threads) but swapping 2 and 3 caused a worse misfire. This is why I was a bit confused. It's almost as if this is not a wasted spark system.

 

Peter, what you say is interesting, I'd like to understand more. Here's some more info, not sure if this helps or confuses:

 

There are two outputs from the ECU, these both feed the IGN1 and IGN2 inputs of the "amplifier"/"ignition module". These are pins 2 and 7 in this diagram here. Pins 1 and 6 in that diagram feed pins 1 and 3 of the coilpack, with pin 2 fed from ignition (white). This all matches the Caterham VXi post-1993 wiring diagram, and that diagram is actually how my car is wired.

 

Anyway, I had assumed that this is a simple setup in which each of the ECU signals is a sqaure(ish) wave to one coil, the "ignition module" is simply a pair of amplifiers packaged in a single component, one amp per coil, which increases the LT signal to each coil, and the coils fire both of their plugs with equal energy. I mainly assumed this because symmetry seems logical/appealing. This of course means wasted spark. However it's possible I suppose that the 2 ECU signals are not one to each coil, but instead are crank phase and cam phase, and the ignition module/coilpack do in fact work out more cleverly which single plug to fire. Is this what you are getting at or have I got the wrong end of the stick?

 

I think I'm going to take all 4 plugs out and earth them to block and crank the engine, this should reveal for sure if they fire in pairs (at least for 1 and 4 at the moment).

 

Edited by - Steve Kirk on 4 Aug 2005 10:19:06 to correct pin numbers referred to in diagram

 

Edited by - Steve Kirk on 4 Aug 2005 10:19:44

[Rob Walker]

A dirty or blocked injector nozzle on number 2 will also give you a misfire I had this problem recently. Got the injectors cleaned all was hunky dory.

[Steve Kirk]

Thanks Rob, that is true. However, I'm confident that fuel is going in from other tests I've done, and I know that there is definitely no spark to #2 hence my focus on the ignition side for now.

[Tom White]

Could be that you have a HT lead/plug break down to ground or a Coil Pack problem.

 

The Coil pack secondary effectively generates a high voltage pulse which passes (via plug lead) from centre electrode to engine block ground on the first spark plug and then from engine block ground to centre electrode on the second spark plug then back (via second plug lead) to the coil pack to complete the circuit. This method can only operate in wasted spark mode as to get a return to the coil pack, both plugs must be in series and therefore fire together. I did read somewhere that if not enough high voltage is generated, one plug will give problems before the other ie. centre pole to engine block is less efficient (I think).

 

Have you tried swapping the coil packs?

 

Hope this helps.

 

Tom

[Peter T]

Is the ECU signal earth or positive fired? to the coil??

 

Carbon Components from K9 Composites

 

[Steve Kirk]

Tom, thanks for your suggestions. What you say is very interesting, I've been thinking along those lines myself but got confused with the theory. In particular, I thought initially, as you say below, that in wasted spark mode, that if one plug/lead fails, then they both must fail together, because the circuit that you describe is then broken. So based on this I tried it on the good pair of cylinders (1+4) on my engine. I removed plug lead 1, but 4 kept firing strongly... so much for that theory!! To be honest that left me totally confused - how can one plug fire alone when there is no electrical circuit? Unless perhaps each coil is "virtual earthed" via the LT terminals from halfway along the HT coil (?) - but I'm in pure guesswork mode now.

 

Peter commented that my system might not be true wasted spark, although I'm not sure exactly what that means. Peter, if you could elaborate I'd be very interested please

 

Tom, having done some more reading on the web last night, I think the reason that one plug tends to stop sparking before the other might be that apparently the negatively fired plug requires less voltage than the positively fired one, so I would assume that as the coil HT voltage declines, the +ve firing plug will stop first.

 

Overall though Tom, I still do think though that in general you are right, it is probably some kind of coilpack or HT problem. I've tried swapping for new HT leads but no joy. So coilpack is my pet theory. I'm going to Steve Foster's tonight as he has a spare I can try, hopefull that will shed some light.

 

I haven't tried swapping the coils yet, but I agree that is a good test. I was going to do it, but after my test above with 1+4 failed, I realised that I didn't understand what was happening, so kind of gave up. As I have a single coilpack that contains 2 coils (each double-ended), I think that to swap over the coils, I would need to both reverse the HT leads on the coilpack (e.g. swap 1&3 and 4&2), and also swap over the LT signals that trigger each coil, does that sound right...?

 

Here's a bit more info that might shed some more light. It is always #2 that misfires, even if I swap over plugs and leads (I have a spare new set of leads). If I connect an HT telltale light in series with #2, if flickers dimly/erratically, so it appears that there is at least a feeble signal there (the teltale flashes regularly and brightly when on 1/3/4). However, exhaust pipe #2 is stone cold, suggesting no firing at all. Also, removing #2 from the block and earthing it to the chassis reveals that it never fires at all. So, it appears that there is a weak HT signal on #2, but not enough for a spark.

 

Edited by - Steve Kirk on 4 Aug 2005 16:25:34

 

Edited by - Steve Kirk on 4 Aug 2005 16:52:12

[Steve Kirk]

Peter, not sure about polarity of ECU signalling. Here's what I measured so far, is what you were asking for?

 

The coil has 3 LT feeds: centre pin (2) is definitely constant +12V from ignition. So I assume that 0V is applied at pins 1 or 3 to energise the LT coil, and then 12V is applied again to trigger the spark.

 

More info: I backprobed coilpack LT pins 1 and 3 with the engine running with a DMM and measured what appears to be an AC signal. Measuring 4-cyl RPM across 2-3 or 1-2 gives half engine revs on each side. Peak voltage is similar on each side, although I didn't note it down. Measuring 4-cyl RPM across 1-3 gives full engine revs. Measuring dwell angle on DMM's 4-cyl range reads 79 deg (=12% duty cycle) at idle across 1-2 or 2-3, falling as revs increase. Just re-read the relevant section in my Haynes engine management manual and found that my readings are consistent with those listed for the "constant-energy" (variable dwell) setup which this Motronic system apparently uses.

 

I'm becoming more convinced that the fault is within the coilpack. Hopefully we shall see for sure tonight...

[Tom White]

I am thinking from what you describe that this is not a wasted spark system. On my VVC (wasted spark for sure) each coil pack gets only one drive signal from the ECU once per rev. You say that you have one +12V and two drive signals at your coil pack input (I am assuming here that we are talking one coil pack for two cylinders!) and you get half engine revs when measuring between +12V and one input but once per rev across both drive inputs.This would be consitent with a system which is driving each sparking plug individually The secondary winding of the coil pack has a centre tap and is therefore able to drive each plug individually once per combustion cycle this would explain why one side works on it's own , on the other hand I might be talking (or writing) B******ks.

 

Still you should find out tonight

 

Good Luck

 

Tom

[Steve Kirk]

Peter, I took some more measurements on the LT side of the coilpack.

 

The AC voltage amplitude between LT pins 1-2 or 2-3 seems to be 2V. I don't have an AC ammeter that can take enough amps to measure the LT current.

 

Also I checked the battery cutoff, this seemed OK, appears to have ballast resistor. Jiggled switch around while measuring ignition volts, seemed steady, varied by max 0.1V.

[Steve Kirk]

To be honest Tom I still don't really know 100% for sure how it's working. I'm pretty sure it's wasted spark but I'm not sure how to test/prove it. But the good news is that we put on Steve Foster's spare coilpack tonight and it worked perfectly immediately.

 

Now that it's working, things seem to have changed. I can now swap the HT leads for 2/3 or 1/4 and it works no problem. I can only guess that the old coilpack was tracking to earth, or across other HT leads somehow, and this was completely messing everything up so that "random" things were happening that wouldn't normally be possible... as you can tell I'm guessing.

 

Re the 2 LT RPM signals reading half rpm each, I think I confused you by explaining badly, here's another try. My DMM has various RPM range settings, you set it to the number of cylinders and it tells you how many RPM that AC signal would represent if the DMM were connected to a conventional dizzy/single coil LT terminal. It does this by counting pulses per second and dividing by 2 (a dizzy pulses 4 times per dizzy revolution, which is twice per engine revolution). So when connected to the LT side of wasted spark coilpack, it divides the pulse frequency it sees by 2, and displays that value, and as a result only reads half the true rpm. Or, connected across both sides of the LT together, it sees the 2 LT signals superimposed, which is a signal of twice the frequency of either side of the coilpack LT alone, hence giving the correct rpm.

 

So.... because of this logic I thought that the readings I got were consistent with wasted spark. I took this as confirmation of my suspicion that it was wasted spark. Hence this is why I was confused when I could not swap the HT leads over that shared the same coil. This was my original conundrum.

 

But now I have a good coilpack fitted, I can swap the pairs of HT leads over and the engine still runs smoothly. So the conundrum has been solved.

 

Also I re-read my Haynes engine management tech book and this states clearly that Bosch Motronic (my engine management system) either has a dizzy in early versions, or is wasted spark in later models.

 

So to summarise, the conundrum is solved - the system IS wasted spark, and the HT leads can be swapped over, just like you'd expect. It seems that because the coilpack had died, it was causing bizarre effects on the HT side of ignition, maybe to more than one cylinder at the same time.

 

Thanks to everyone who helped out

 

 

Edited by - Steve Kirk on 5 Aug 2005 01:09:01

[Tom White]

Hi Steve

 

Good to hear that your not wasting too many sparks any more 😬

 

Tom