Sensitivity of an engine to ignition advance or, in this case, retard.

Dear Friends,
sorry, I don't speak French, but hopefully you won't mind if I post this, as in your forum there are great experts as Frits Overmars and others who could help me and make an interesting discussion too for others.

The premise is that I am using on my 2 stroke enduro bikes a very sophisticated CDI which has a lot of great features, one of which is a sort of innovative/original traction control (I'd say more a power control), which doesn't use any wheel sensor but monitors the speed, acceleration and jerk of the RPM and of the TPS alone. You can consider this system as made of two parts: a part that decides how much reduction in power is required cycle by cycle, and the other part that should execute this, by adding a certain amount of retard to the ignition, to achieve that. There are two ignition maps, one is called IFP (IgnitionFullPower) and is the classical ignition map to achieve full power, the other is called ILP (IgnitionLeastPower) and is a map to achieve as little power as possible, i.e. maximum retard without causing misfires or other problems. The system then choses, within the bounds of IFP and ILP, the ignition timing to improve traction on the wheel. It works really good, but..

The juice of the discussion is that on my KTM I had great control of the engine, after developing the ILP map of course. Now I have a TM motorbike (this CDI can be used on many different bikes) and to my great disappointment I realized that it's incredible how insensitive is this engine to the applied retard. You can ignite at PMS-15 degrees (IFP) or at PMS+10 (ILP) and the change in torque will be minimal (on the contrary, on the KTM the difference will be enormous). You can't ignite past PMS+10 (the exact value depends by the RPM, I'm using this figure as an example) because you'll get a misfire.

So I can't exploit this TC-like system (which I love to say the least) on my new bike, unless I find a way to make this engine sensitive to advance/retard.

But how?

The squish doesn't seem to make any substantial difference. Carburetion either. I tried using a prominent spark plug, thinking that if the spark originated in a zone without turbulence, the ignition timing would become more important. But whatever I tried, this TM engine remains deaf to ignition timing. How is it even possible? It's not that a TM cylinder and a KTM one look that different.

One other road I am trying is to increase even more maximum retard (ILP map), but as I said beyond a certain value I get misfire, so I'd like to investigate why it misfires so that I may improve that and be able to use greater retard values which will certainly improve disempowerment. My idea, as the cause of misfires when retarding much, is that with > ~10 degrees after PMS the turbulence has increased to much that the combustion will originate but will be blown away immediately after, hence a misfire. Another possibility would be that the exhaust port opens when combustion hasn't finished yet, and this disrupts the next cycle or scavenging, but I really much doubt this as, with a retard > x, misfires are happening at every cycle, so it's not excess of residual gases or such, but more probably the combustion doesn't develop at all (again, I think because high turbulence blowns away the just-generated combustion).

Anyhow, could you (Frits? anybody else with experience in ignition timing effects?) help me understand what makes a 2 stroke engine more or less sensitive to ignition timing? And, also, how to get stable combustion also with high retards? It's not a field that has been much investigated in literature, for obvious reasons (usually one tries to get maximum power, not maximum disempowerment possible).

Thank you very much.

With kind regards,
Tony

Tony, your story reminds me of a similar experience Jan Thiel and I had a long time ago when we performed the first experiments with programmable ignition at Garelli. Looking back, we probably ran into HCCI (Homogeneous Charge Compression Ignition).
There is a massive amount of information on HCCI available on the internet but for your case I am thinking of two main factors: compression ratio, which you can easily compare to the KTM engine, and exhaust gas retention, which is greatly influenced by the behaviour of the exhaust power valve.
Fixing the power valve in the fully open position will give an unusable powercurve for an MX engine, but it may give an indication whether your problem is caused by too much exhaust gas remaining in the cylinder.
Your "innovative/original traction control" reminds me of the GET system. If you are really using this system, you could contact [Vous devez être inscrit et connecté pour voir ce lien]; they have done a lot of development work on it.

Hello Frits,
no it's not the GET system, which runs on injected 4 stroke engines only, this instead is the PowerCDI and runs on both 4 stroke and 2 stroke engines. When I looked for a programmable CDI for my 2 stroke engines, this was the most impressive available, it has a lot of features beyond mapping (there's a lot of interesting info on [Vous devez être inscrit et connecté pour voir ce lien] but unfortunately it's only in italian language).

Would a higher compression ratio increase sensitivity to ignition timing? I thought that too, but unfortunately the TM engine (the non sensitive one) is more compressed than the KTM one! So it's not this that is causing the "deafness" of the TM engine to ignition timing. It's really weird.

My TM has an electronic powervalve and the PowerCDI is in charge of that too, so I will map the powervalve more open to achieve more blowdown time * area and see if this improves sensitivity to retard. I actually had thought of it too, thanks for your suggestion, this supports the need to experiment with the electronic powervalve too, which is available on this bike and demands its attentions.

I don't think I ever ran into HCCI, because from time to time I try to cut ignition (just to see if there's self-ignition in extreme cases like part throttle at high RPM, much retard, etc..) and combustions always stops completely.
Just out of curiosity, did you ran on HCCI on your experiments because of excessive retard, or because of not enough scavenging on part throttle? i.e. too much hot gases remained in the cylinder and there was not enough pressure from the crankcase so the fresh gases were heated and then self-ignited? This wouldn't happen on full throttle though, I guess, nor would even happen on part throttle if the powervalve was completely open (of course you can do this only at low-mid RPM's, not at high RPM's where the exhaust valve is fully open anyway and it can't be opened more).
Also, I'm using it on extreme enduro, where the low and mid RPM's are commonly used, while on road bikes like in your case the powervalve is rarely found closed, so I reckon that "playing" with the powervalve was not possible/really_useful for you, but it could for me.

On a side note, have you or Jan experimented with prominent spark plugs or anyway with placing the spark plug higher or lower in the combustion chamber to meet a more or less turbulent zone? That should change the ignition delay. Of course in your application it would be more useful to place the electrodes in a high turbulence zone, while in my case it would probably be more useful to do the opposite (as to increase dependence on ignition advance).

Thanks a lot for sharing your know-how and suggestions, it's always a pleasure to read you or Jan.

With kind regards,
Tony

Tony, I think that a higher compression ratio will make the engine less sensitive to ignition timing variation.
Running into HCCI with an MX bike is not very likely because the specific power is lower; it was just a hunch from me.

In our experiments (Garelli, 1987) we ran only full throttle and the engine did not have a power valve. Jan was operating the dyno, safely behind double glazing, while I was manually varying the ignition timing at the engine. But no matter what I did, nothing changed.  HCCI was unheard of in those days, so we did not investigate any further in that direction.

You want to keep both the ignition delay and the combustion duration as short as possible at all times, and bringing the spark more to the center of the combustion space definitely helps.
We modified a cylinder head insert such that part of the plug thread intruded into the combustion volume.
That part was then ground off, so that only the mass electrode remained,which gave 500 rpm more maximum revs.
[Vous devez être inscrit et connecté pour voir cette image]

Thanks for opening this topic and for sharing your skills!

Thank you Frits,
what you did with the spark plug is equivalent to using a prominent spark plug (e.g. using a BPR8ES instead of a BR8ES). It also improve idle, at least on my engine and according to my experiments.

The underlying effect is that the spark plug electrodes are thus placed into a zone of maximum turbulence. I heard straight from its engine designer (Franco Lambertini) that the Moto Morini Corsaro 1200 has a 107mm bore, the only way he managed to make it run with acceptable advance and only one spark plug, has been to place the spak plug electrodes below the usual point (about 4mm if I recall well), i.e. into a zone of maximum turbulence, to reduce ignition delay as much as possible (he said that 60%/70% of the whole combustion time was otherwise in the ignition delay!).

There's also a con's when one places the spark plug electrodes into a zone of high turbolence, and it is that the combustion may be blown off (like a candle in the wind, so to speak) just after it starts, if the turbulence is excessive, so this is not usually done on engines, also because the ignition delay becomes a problem especially on big bore engines, not on normal engines.

Thanks for sharing your suggestions, I had thought that a higher compression, making combustion quicker, would have made the engine more sensitive to ignition timing because the peak of the pressure would be more concentrated into a certain crankcase angle. Can you confirm please that I was wrong?

On a side note, I actually saw a MX bike ran into HCCI once, and it was certainly HCCI (as we cut the ignition, let away that it went past the limiter revs): it happened on part throttle at very high RPM (beyond the limiter), at neutral (on the side stand), but that engine was heavily choked on the exhaust (exhaust boosters always closed and a 30mm reduction in the exhaust collector) on a YZ250 again with the programmable CDI I mentioned (which has also a limiter and other nice features usually not seen on 2 strokes). It was very impressive, but it ran smooth. We closed the throttle and it stopped doing that.

Have a nice sunday.
Tony

Maverock a écrit:

Thank you Frits, what you did with the spark plug is equivalent to using a prominent spark plug (e.g. using a BPR8ES instead of a BR8ES).

That is correct Tony. But I would not use either of those plugs; I'd prefer the NGK R7376.

Citation :

There's also a con's when one places the spark plug electrodes into a zone of high turbulence, and it is that the combustion may be blown off (like a candle in the wind, so to speak)

Indeed. The high tension ionizes the gas molecules between the plug electrodes, changing those molecules from non-conductive to conductive, and as soon as the ions form a complete path between the electrodes, a current will start to flow.
But while the ionisation takes place, ions may be blown away by turbulence, so the conductive path may become longer or it may not be completed at all. That is why a two-stroke ignition needs a longer period of high tension than a four-stroke ignition.

Citation :

I had thought that a higher compression, making combustion quicker, would have made the engine more sensitive to ignition timing because the peak of the pressure would be more concentrated into a certain crankcase angle. Can you confirm please that I was wrong?

Everything you thought is correct, but a higher compression ratio implies a higher expansion ratio which means that there will be less exhaust gas energy left for the exhaust pipe to work with. And this is decisive. If the fluctuation between good an bad pipe effects is smaller, everything else will also react less sensitive.

So the exhaust pipe is a key factor also in my attempts to "reduce power on command", yes this is very interesting and I will have to devote a lot of future brainstorming to it, thank you.

One last question please: as it's very expensive, what advantages would a NGK R7376-8 give versus a BR8EIX (which is iridium too, and thermal grade 8 like the other)? The NGK R7376 is like 5 times more expensive, is it worth on a 300cc 2T MX/enduro bike?

With kind regards,
Tony

Maverock a écrit:

what advantages would a NGK R7376-8 give versus a BR8EIX (which is iridium too, and thermal grade 8 like the other)? The NGK R7376 is like 5 times more expensive, is it worth on a 300cc 2T MX/enduro bike?

Here is a short test summary of the NGK R7379 versus the EIX:

The NGK R7376 has a fine wire iridium center electrode, same as the cheap EIX plugs. But the similarity ends there.
It has a special ceramic composition that won't crack and drop lumps into the cylinder at the first sign of deto shock, like an EIX does.
The ground strap is also a fine platinum wire, laser welded onto the shell, so that it won't fall off like those of an EIX or EGV do all the time.
The fine wire combination needs way less voltage to ionize the spark gap and the resultant flame kernel is way bigger than a normal fat wire plug. And the resistor element is of a high quality.
They have the exact same nose configuration as the short race plugs for a HRC RS125/250, but are 1/4 the price, and use a normal plug cap, not the special NGK caps that cost more than the race plugs do.
But the best part: in a 125 cc race engine this plug makes more power than any other plug, with around +2 Hp better than an expensive Denso equivalent or an NGK BR10EGV.

Warning: There are 2 different NGK R7376-10 plugs.
The good one with the iridium electrodes is the 4457 R7376-10 which are 28.47 dollars on ebay.
The cheaper 4374 R7376-10 are 10.38 dollars on ebay but their electrodes are not iridium and they may wear out in 3 to 4 hours.

Maverock a écrit:

............ is it worth on a 300cc 2T MX/enduro bike?

With kind regards,
Tony

it's absolutly not , in your case   " I'm using it on extreme enduro, where the low and mid RPM's are commonly used".

you may have some improvement full load/rev's on the dyno , for sure ,but you dont ride that way  

those kind a plug are for TRACK use  , only  ....

if you want to understand why your POWER-CDI has no effect on your new engine , you may first try to see :

1/ if its works well on that bike

2/ if you have made the right program for it

3/ and finaly , you must know what are the main differences between the two engines ( KTM / TM)  

you'r riding enduro , not MotoGp , so , stock plug is almost very good

it will help us to understand , having the data you put in the POWER-CDI , the table,  rev/load/ignition advance  and the dwell table



.

I do race the R7376 on my 50cc - it noticeably improved idle and part throttle behavior! Those BRs are the right thing for a lawn mower...

If you wanna be sure the best way is to try    

but in his case it wont solve the problem  

BR's plug works perfectly well when properly used

it will be interresting to see how many 300 cc enduro bike run the R7376  plugs , because here we speak about this kind of bike , dont we  



.

Thank you all for your new replies, I was in holiday and I forgot my passwords so I couldn't reply until now. :) Next thing I will purchase a good R7376 and experiment with it, then I will report here, but I'm afraid it will take a couple of months before I can do the tests. Meanwhile, have a nice summertime!

Dear Frits,
my enduro 300 bike needs a heat range 7 spark plug (I currently use the BR7EIX), could you please give me the NGK code for the R7376-7 spark plug, both in iridium and cheap form, so I can be 100% sure I'm purchasing the right one? (the iridium)

I tried to find this code by myself, but I'm not sure it's correct as it seems that NGK uses the same code for different products (e.g. 4374 is used for both the R7376-10 non-iridium and for the CR8EKB multi-ground plug). Moreover, NGK site says that the 7 doesn't even exist, but on eBay it seems it does:
[Vous devez être inscrit et connecté pour voir ce lien]

Thanks for your time.

With kind regards,
Tony

Maverock a écrit:

Dear Frits, my enduro 300 bike needs a heat range 7 spark plug (I currently use the BR7EIX), could you please give me the NGK code for the R7376-7 spark plug, both in iridium and cheap form, so I can be 100% sure I'm purchasing the right one? (the iridium). I tried to find this code by myself, but I'm not sure it's correct as it seems that NGK uses the same code for different products (e.g. 4374 is used for both the R7376-10 non-iridium and for the CR8EKB multi-ground plug). Moreover, NGK site says that the 7 doesn't even exist, but on eBay it seems it does:
[Vous devez être inscrit et connecté pour voir ce lien]

Yes, using identical codes for different products can be rather annoying. And I could not find the plug you are looking for either.
Judging by the shape of both the center electrode and the ground strap on the Ebay picture they are not offering the real deal there.

Thank you, from another NGK site it seems that heat range 7 of the R7376 did exist, but has been discontinued (stock no. 6991, and probably it wasn't even of the iridium kind).

I guess that the resistance to fouling of the R7376-x may be higher than that of a BRxEIX, but not that much anyway (it depends by the insulator, not by the ground electrode), so maybe it's better if I stick with a BR7EIX, as in trial-like enduro it's common to run with low throttle and low RPM's, and I wouldn't like to foul a plug (never happened, but as I said always using the 7 anyway).

Or do you think that a R7376-8 iridium (thus, colder than what I use) would be anyway more resistant to fouling (at low temps, turtle-like enduro) than my usual BR7EIX?

Cheers,
Tony

Maverock a écrit:

... do you think that a R7376-8 iridium (thus, colder than what I use) would be anyway more resistant to fouling at low temps, turtle-like enduro) than my usual BR7EIX?

I can't say Tony, I never tested the EIX. Not only because the R7376 is better for power, but even more so because the R7376 uses a better insulator material which is far less likely to crack and fall into the engine.

Thank you very much for your time Frits, I'll make the needed experiments.

Maverock you need a Denso IW01-27 or 29. I have tested a progammable power valve/ignition on a CR 250(standard engine) that gave about 5hp more and kept peak power for another 1000rpm. Have you tried testing on the dyno or on the track?

Institute of TwoStrokes a écrit:

Maverock you need a Denso IW01-27 or 29. I have tested a progammable power valve/ignition on a CR 250(standard engine) that gave about 5hp more and kept peak power for another 1000rpm. Have you tried testing on the dyno or on the track?

Hello,
as real dyno is outside of my possibilities, but the PowerCDI I'm using comes with a USB cable and software as standard, one of the functions is a "inertia dynamometer": I go into a straight, flat piece of road near my house and run from very low rpm till overrev in 3rd gear, while recording telemetry. From that (the RPM data, engine cycle by engine cycle) and other parameters (weight, supposed Cx and cross-section, transmission ratios, etc..) the system outputs torque and power curves. It's very repeatable.

Thanks for the hint, I'll look into the Denso IW01-27 or 29. 5 HP is a lot (I'm not much after extra power, but it's clear that if you get it then you have better / more complete combustion, and that's always good).

What I'm really trying to achieve is different though: the dynamic power control of my CDI requires a "max retard" map, and the DPC (Dynamic Power Control) system will use any value from the max advance ignition map (i.e. the classic ignition map, made to get as much power as possible) and all inbetween till the maximum retard map (used to produce as much disempowerment as possible, i.e. for traction-control like situations). However, this TM engine is proving quite insensitive to retard, but as the response to ignition timing resembles a parabola, I know that if I can use more degrees of retard I'll finally get enough disempowerment.

But too much retard causes misfire. Already at PMS+8 I get misfire, especially around 5000 RPM (i.e. with the top revs at powervalve closed).

Why does it misfire at > PMS+~8? Is the turbulence that "blows away" the just begun and weak combustion kernel?

What can I do to extend the retard, without getting misfire?

This is what I really need.. much more than +5 HP.

Maverock a écrit:

Why does it misfire at > PMS+~8? ............What can I do to extend the retard, without getting misfire?

In your case , do you really believe one will give you any answer to these questions !!!!

it's like asking : the numbers of stars in the universe

you have to study your engine first , and you have to do it very very well , you also need to have some very good knowledge of thermodynamics , electronics , and some deep experiences in those matters

a dyno-bench it's NOT an option , it's a tool for those who work on combustion engine 2 and 4 strokes , and many more !!

if you understand why it work with the KTM , you will certainly understand why it' doesn't with the TM.


ask yourself the right questions





.

It's not that it doesn't work on the TM, it doesn't work enough.

Engine's torque response depending on ignition timing can be graphed like a parabola (think about a graph where on the x-axis you have ignition advance, and on the y-axis you have the torque produced by the engine at that particular advance, the graph will look like a parabola, on any engine), you - when looking for best power, i.e. MBT (Maximum Brake Torque) - want to stay on top of this parabola, if you advance less, you'll lose power, if you advance more you'll not only lose power but you'll also overheat your engine or even detonate.

When you are on top of the "parabola" one or two degrees of advance or retard won't be easily noticed, but the degrees of difference increase the parabola gets more and more steepy, and thus ignition timing sensitivity increases, i.e. for each additional degree of retard or advance, more and more torque difference will be there. Hence the "parabola" is so important, it means that near MBT (optimum advance) all engines are relatively insensitive to ignition timing, but if you add retard, you'll lose power more and more significatively. The first degree of retard won't produce much disempowerment, but the more retard you add, the more (exponentially) the disempowerment comes.

If you want to disempower *effectively* your engine on command via ignition retard, you have to reach the right side (i.e. retard) of this hypotetical graph, enough to get a useful amount of disempowerment. But unfortunately you may get misfire before reaching a disempowerment level that may be useful..

The KTM engine has a "narrow" parabola, meaning it's very sensitive to ignition timing. It also allows a fair amount of maximum retard before it misfires (and thus can't retard more than that). I could disempower, in 3rd gear, from 3000 RPM/s to about 600 RPM/s, that's a great torque reduction and a great result.
The TM engine has similar max retard characteristics, but a wider parabola, meaning it's quite insensitive to ignition timing. I can go from the 3000 RPM/s of the full-advance map to 2000 RPM/s of the full-retard one, which is not satisfactory at all.

The traction-control like system is much more effective on the KTM, for this reason.

Now, if I can "disempower on command" the TM more, it would work great also in my new bike. To do that I could either increase its sensitivity to timing (to retard) or the maximum retard it will accept.

As far as I can imagine, the more an engine is compressed the faster the burn rate and thus its sensitivity to ignition timing. The KTM and TM engines don't look too different at all, but the latter is more compressed so I'd expect it to be more sensitive to ignition timing, but I'm experiencing the opposite, so compression is not all. What else can it make so insensitive to ignition timing? Frits mentioned the exhaust, but I can't think there are many differences there too. All 2 stroke 300's look very similar to each other. Same bore, volume, very similar distribution, etc.. but of course some tiny difference must be there, although its effect of ignition timing sensitivity is much more than tiny..

Another front I could fight on is the one to extend the maximum retard the engine will accept without bogging or misfiring, that would be very effective too as a method because the farther we get from the top of the parabola I mentioned, the more disempowerment we get for each additional degree of retard.

Yes, I have good knowledge of thermodynamics, excellent of electronics (I'm an engineer), but given that here on this nice forum there are engine designers, I hoped that - even if nobody has ever tryed to map an engine before for LEAST power (one always looks for the opposite, i.e. maximum power is the goal!) - some of you may have the experience anyway to tell me e.g. why the engine bogs igniting at just 10 degrees after PMS but runs instead very smooth just 2 degrees less retarded.
I don't believe that it's because combustion continues in the exhaust pipe, we're not talking about huge retards here. I rather think that by PMS+10 the squish has caused so much turbulence that the combustion kernel is blown away.. so maybe some of you can suggest me how to make it stronger, if the premise is correct?

If I can find means to extend of few degrees the onset of misfiring, I'll get much more disempowerment, because I'm more and more on the steepy part of the parabola I mentioned before, so each additional degree of retard I can use will give me exponentially more disempowerment.

I hope I explained my problem better now.

Thank you. :)

Tony.

L'augmentation des turbulences ainsi que l'augmentation du régime moteur demandent de revoir le système d'allumage dans sa puissance et dans sa rapidité

Maverock a écrit:

I don't believe that it's because combustion continues in the exhaust pipe, we're not talking about huge retards here. I rather think that by PMS+10 the squish has caused so much turbulence that the combustion kernel is blown away.. so maybe some of you can suggest me how to make it stronger, if the premise is correct?

I think you just gave an excellent analysis of the problem Tony.
I do not believe either that combustion continues in the pipe; that would require an incredibly slow-burning charge (so lean or so rich that it would be spark-unignitable anyway) or another 40° or so of ignition retard.
Making the ignition more effective may not be that simple, but making it less effective is. Lowering the primary coil tension would do that, but it is even simpler to play with the plug gap.
[Vous devez être inscrit et connecté pour voir cette image]

Thanks for encouragement Frits. I also have to experiment with spark plug indexing, perhaps by having the ground electrode "shielding" in one position will eliminate blow-out (assuming my problem is blow-out) at high (but not so high really..) retard values. Also, higher gap will improve quenching but will worsen blow-out, so I guess it's time for me to stop being theoretical and go into empiric mode now.