Knock/advance

[Robert_A_Brady@rohmhaas.com (Robert A Brady)]

Some comments on ignition timing and knock (sorry for the length... it 
inadvertently became a treatise):

To meet NOx regs in the US, auto manufacturers retarded the timing in cars 
in the early '70s.  NOx forms at higher combustion temperatures, and also 
has a relation to pressure as well (scales with P^0.5 as I recall).  In 
spark ignition engines, as you retard timing from TDC, both peak combustion 
temperatures and peak cylinder pressures go down (as does NOx).  On a 
simple level, , this is because the gases are expanding at the same time as 
the combustion chamber.  On the other hand, if you spark as you are 
compressing (ignition BTDC), the expanding gases (due to combustion) are 
acting against a compressing combustion chamber.  Thus, peak cylinder 
pressures and temperatures are higher.  

In terms of efficiency, if combustion were instantaneous, you would expect 
the engine to be at its peak power AND efficiency while sparking at TDC. 
However, combustion does take some time.  So, in practice, peak power 
normally occurs at timing somewhere BTDC.

A quick look at tune-up specifications for Euro and US alfas will reveal 
that the rest of the world tunes their cars for more advance than in the 
US.  Expect your power to increase as you advance your timing from US 
specs.... to a point.

At some point of advance, power goes down.  Simplistically, you can imagine 
the expanding gases working against the piston moving up.  At some point, 
you'll be firing too early  and the pressure produced from combustion works 
against the piston (on the up stroke) more than it works with the piston 
(on the down stroke).  Again, if combustion were instantaneous, this effect 
would not be there.  If this were the only issue with spark advance, the 
method of maximizing acceleration would be a fine way of determining proper 
spark advance.  Of course, life is not so simple due to knock.

Knock is synonymous with detonation.  Strictly speaking, a detonation is 
different from a deflagration (typical "burn") in that the flame front is 
supersonic.  Detonations typically produce much higher pressures than 
deflagrations.  Spark ignition engines are designed to operate with 
deflagrations.  As the flame front proceeds, the gases behind cool and 
contract a bit.  When a detonation occurs, combustion occurs much more 
rapidly, and peak pressures increase due to there being less time for the 
combustion gases to cool (and contract) behind the flame front.  Thus, when 
you detonate the fuel-air mixture in an SI engine, you produce cylinder 
pressures (and temperatures) much higher than the design conditions.  This 
leads to melted/damaged pistons and other "bad" things.  Detonation is 
something to avoid in SI engines.  Diesels, on the other hand, are designed 
to operate in this mode.

For a given fuel air mixture, the tendency for knock increases with spark 
advance.  For a given spark advance, the tendency for knock increases with 
fuel-air mixture initial temperature, and reduces with octane number of the 
fuel (empirically determined, by the way, using a standard engine whose 
compression ratio is increased to the point of knock.... not much "science" 
in that number).  The tendency for knock is also a function of fuel-air 
mixture stoichiometry (actually Lewis number.... but for this discussion, 
the practical quantity is fuel/air ratio).  I don't know on which side of 
stoichiometric gasoline falls with respect to minimum ignition energy, but 
I'd expect it to be on the rich side.  That is, the point of minimum 
ignition energy is with a slightly rich mixture.  Of course, in the real 
world, the fuel-air mixture is not truly premixed, and the theory is a bit 
hard to apply....

So, in practice, to make your car knock, advance the timing, drive it on a 
hot day, and use low octane fuel.  I've also observed that the tendency for 
knock is supressed with richening the fuel-air mixture.  (When I bought my 
'74 gtv, the owner had the timing advanced to >30 degrees BTDC!, but the TA 
was shot and the car did not knock at all).  We've also all experienced the 
phenomenon of knock occurring under low rpm with wide open throttle.  I'll 
have to think about that one a bit to explain why it occurs... but not 
today.

The bottom line is that I'd recommend advancing timing until either the car 
knocks or acceleration goes down.  In practice, I find that on American 
iron, knock will occur before acceleration diminishes.  On my spica gtvs 
with stock cams, I've found best performance to be about 7 degrees BTDC at 
idle to be the best advance.  At this position, I experience no signs of 
knock (either audible or visual... when the engine is apart).  Further 
advances don't seem to increase acceleration.

Hope this helps.

Bob Brady, DVAROC
'74 gtv
'88 verde