Ignition Advance Curves.

I found this interesting :

http://www.autospeed.com/cms/article.ht ... t&A=109132

I'm still of the opinion it is possible to program the curve to suit the interaction of the internal engine geometry with the fixed time combustion event and jet the rest.

I spoke to my friend with the 750ss Ducati yesterday. He has fitted an igniton system from a Honda VT750 Shadow, which advances quite well. He said the major effect of the change was the bike is much smoother and quieter to ride. To my mind 'quieter' is good because loud indicates wasted energy. The system cost him $90. It must be programmable, however I doubt that Mr Honda would provide the relevant software. I am still trying to get my head around how to derive the theoretical optimum advance curve. I won't buy a programmable ignition system before I know what I want to program into it. It is too easy to end up in a situation where the dog chases it's tail.

All any ignition system has to go on on an old brit twin is where the crank is and the revs, the ignition map will simple and crude, best you could do is to store a few different curves for different conditions. To be able to use a modern ignition system you need to provide the additional info it needs via throttle position sensor, knock sensor, engine temp and emission data eg Lambda probe. Then when you then work out the carb is holding you back you then need to fit fuel injection, then you will be looking for water cooling etc etc.

A good start would he here

http://ignitech.cz/en/

In my own experience, these old, crude engines that we play with are not very sensitive to fine tuning of a few degree here or there and even less fussy about finely tuned advance curves.

Far better to focus on getting the biggest, fastest spark you can get.

This was typified by my playing around with a highly tuned 988cc BSA triple. I fitted a Tri-spark and 3 high output 12volt Dyna coils in place of a Boyer with 3 x 4 volt coils all on a wasted spark. The spark was very powerful and would jump a big distance, I could probably have welded with it!

With no other changes, the fatter spark yielded a 5bhp increase, more mid range and easier starting.

The Tri-Spark had a changeable advance curve, so I spent an afternoon playing with a myriad of different combinations of settings. No fiddling of ign timing or advance curve rate achieved any meaningful differences either way!

That's only in my own experience though, I'm happy to observe the fruits of others labour. And expense.

The expense is a major factor in my thinking. I'd like to buy a programmable ignition system, however I think that using curves at random from other engines might be a futile exercise. If you read that article I've posted, I think it says that the combustion event takes about 2 milliseconds. To my thinking that time would not change much regardless of the revs. What does change however is the rock-over time of the piston as the revs rise. So to have the highest pressure occurring at the same piston position you need to advance the starting point of the combustion process proportionately to the revs. In Commandos we use two different rod lengths and these affect the rock-over time, also many of the systems which are available have pre-programmed curves. What I am suggesting is idealistic and possibly unnecessary.
That article mentioned that it is impossible to calculate the theoretically ideal advance curve. I don't accept that - it might be a mind-bender, however NOT impossible. I suggest the algorithm required to calculate the added advance as the revs rise, involves the same calculation we use to convert degrees crankshaft rotation to piston distance from TDC. My approach is always to set the ignition advance in a fixed regime, then jet to it. Ignition advance and jetting are not independent variables.
Eddie, I take your point that getting a fatter spark probably has more effect than using an advance curve. However in my experience, sometimes what you get is not what you might expect.
I'd point out that my friend's experience where the noise level dropped and the motor ran more smoothly when he used the Honda VT750 ignition system on his Ducati, is an indication that the might be a torque benefit in using an advance curve. If you change the exhaust cam timing on a four stroke race bike without changing the exhaust pipe, usually what happens is - as the valve opens earlier beyond a certain point the noise gets louder and the bike goes slower. The problem is also this - if you manage to improve the torque output of your motor, it often doesn't do anything until you raise the overall gearing. It is very deceptive.

Kommando, In your post you mentioned using an engine management system with a knock sensor. In the article I posted it mentioned that in a certain Porsche engine that maximum torque occurs well before knocking occurs. If you knew how much before, it wouldn't be difficult to programme for it. I take your point about throttle position. On some Harley systems they use a vacuum switch to change the advance curves. On some Yamaha two strokes the exhaust port timing is altered using an analogue system and I think the throttle is coupled to that. I don't know whether they have ever used a full EMS or even an ignition system with programmed advance. I always ride my 850 as though it is a two stroke as far as snapping the throttle open is concerned. I always feed the throttle on in one in one firm deliberate motion, that way you don't lose vacuum.

A knock is a either yes or no, no sensor will tell you its nearly knocking, thats why they have ECU's that learn and adjust fueling and timing to allow use of low octane fuels and then adapt to take advantage of a tank of high octane fuel. I had a Saab turbo, I tuned it by adding an intercooler and increasing turbo boost, pulled and replaced a fuse to force ECU learning and then 20 miles later it had retuned the fuel and timing map.

If you read the article, it mentions that in one of the Porsches as the advance is increased the maximum torque occurs a considerable number of degrees before the knock occurs. Surely it must be possible to programme the unit using the knock to create the curve, then move the static advance back a bit ? Finding the amount to retard the whole curve would be similar to the situation when the timing is fixed.
What I am saying is that running the motor against the knock sensor doesn't necessarily give you maximum torque, and with a Commando you necessarily don't know the torque has increased unless you are able to pull higher gearing and the bike is faster at the ends of the straights.

No, as no sensor exists to know just before the knock occurs then you are stuck with a knock sensor. So you have to advance to the knock to find the knocking point during the learning phase. As soon as you map this then yes you can advance short of it but like my SAAB if you use low octane for a tankful then unless you clear the memory then the next time you go hi octane the ECU will not utilise the higher octane as the map will not go near the knock point. So your performance is stuck because of one bad tankful or petrol unless every now and again the timing is advanced to the knock point to find the new reference.

So have you got any idea how far the ignition is retarded after the knock occurs in the SAAB, to get maximum torque ? Or is that info only available to the factory franchised service people ? With a Commando, I would not even hazard a guess as to how far maximum torque occurs away from knocking. Many years ago I used to ride a WW2 Indian Scout which had the advance retard twist grip on the right and the throttle on the left. Form memory it was possible to get it to knock. However if you advanced it beyond a certain point it would slow down and run with a bit of distress before you ever got the knock.

Why not use a dyno first on a Norton engine and play around with the timing, what happens in a Porsche combustion chamber may not happen in a Norton combustion chamber, just look at the difference between the advance used on a Norton compared to a Triumph with their differences.