19 replies to this topic

[desmo]

Here is an excerpt from an article written by the excellent technical writer Kevin Cameron on detonation that got me to thinking:

"Books will tell you what Harry Ricardo learned back in 1918; that detonation is not the same as preignition. Preignition is lighting of the charge before the spark , by some hot object in the combustion chamber usually the overheated center wire of a spark plug whose heat range was too hot for the application. Preignition soon provokes detonation, so the confusion is understandable.

Detonation, by contrast, is self ignition of some of the last parts of the charge to burn the so-called "end gas" out at the edges of the combustion chamber after the spark has already ignited and mostly burned the charge. This self igniting endgas does not then burn normally, as a flame front spread by turbulence at the usual speed of a few tens of feet per second. This gas burns at the local speed of sound, which is very high because the temperature is high. This form of combustion, called detonation, forms a shock front, a sudden jump in pressure that propagates at thousands of feet per second.

When it hits parts, it hits hard. If we hear it al all, it is as a high, dry, irregular clicking, not unlike the reverberating sound of rocks struck under water. Detonation's pressure front can damage bearings by its hammering shock, but the real problem is what it does to an engine's natural, internal insulation.

In any situation in which gases move next to solid surfaces, there is a layer of significant thickness that remains largely stagnant because it is attached to the surface. In internal combustion engines, this boundary layer quite effectively shields the engine's metal internal surfaces from direct contact with combustion gas, keeping them cooler than they would otherwise be.

When detonation begins even light deto this boundary layer is scoured off by the impacting shock waves, and heat transfer from hot gas to cool metal accelerates. In only a very few detonating cycles, piston temperatures rise dramatically, and the rest of the parts exposed to combustion gas aren't far behind.

What is strange to many people is that as this happens, exhaust gas temperature falls. This seems odd because people associate detonation with heat, and heat with failure. But the fact is that as you jet an engine down, its exhaust temperatures peak, not when the mixture becomes lean (that is, too little fuel to react with all the oxygen in the air charge), but when the mixture is chemically perfect. Exhaust gas temperature falls when detonation begins because the engine's internal insulation is destroyed; that being so, some heat that would otherwise go out the exhaust is now being diverted into the piston, head, and cylinder walls. Because those parts are getting hotter, the exhaust gas becomes colder.

Those of us who began racing before water cooling arrived tend to think that engines get hotter the more we jet them down. With air cooling, this seems to be true, but isn't. The engine runs cool when it's rich because the extra fuel reduces peak flame temperature, and as we jet down towards chemically correct mixture, the engine runs hotter and hotter. Often, in a modified engine with high compression, this detonation begins even before we reach correct mixture and peak flame temperature. Then the engine really heats up. This leaves us with the idea that leaning down the mixture raises engine temperature, in a straightline relationship.

Now we know, from our experiences with water-cooled engines, that power, engine temperature, and exhaust gas temperature all rise as we jet down until we go beyond chemically correct mixture. When we do, power, engine temperature, and exhaust gas temperature all begin to fall again. We couldn't see this before, with air-cooling, because the power we were making was overwhelming the engines the engine's cooling ability. But it makes perfect sense because heat release in combustion depends upon finding enough oxygen so that each and every hydrogen and carbon in the fuel is completely reacted to form water and carbon dioxide. Any fuel left over is potential chemical energy unreleased which is why running lean makes less power. On lay well cooled engine that is not detonating, you can jet down until it starts to slow down.

Now back to detonation. The above explanation is the common one, but it leaves important questions unanswered. For example, why does detonating combustion travel at the local speed of sound, and not at normal burning speed? Why does the endgas auto ignite, rather than simply wait there like a stand of trees in the path of a forest fire? Understanding how this comes about helps to understand how the variables that affect detonation generate their effects and it helps to fend off the phenomenon that sets the upper limits on performance.

There are two basic forms of combustion, deflagration and detonation. In deflagration, the propagation of combustion is carried out by simple convection; the hot combustion gas heats what is ahead of it, raising its temperature to the ignition point. Because this process of heating what lies ahead takes time, it is relatively slow. The burning of a quiescent gasoline air vapor is in fact slow only a foot or so per second. Combustion in an engine cylinder is much faster than this because of turbulence, which so wrinkles the flame front that its area becomes hugely enlarged. This area, multiplied times the slow quiescent combustion speed, computes out to a very large volume combustion rate.

Detonation is a different animal, and not all gaseous mixtures will support detonation. It is a form of combustion in which the unburned material is heated to ignition at least partly by shock compression, as the detonation wave moves a the local speed of sound through the medium. This has to happen very quickly, so fuels with simple molecules or those with low stability lend themselves to this form of combustion. Now how does the endgas ignite by itself? It does so when its temperature is raised by any combination of effects to some critical value in the range of 900-1000 degrees F.

In a running engine, air is drawn in at some ambient temperature, and this air then begins to pick up heat from the hot internal engine surfaces it contacts. Finally it enters the actual cylinder, where is it heated by even hotter surfaces. Trapped there, it is heated again by the process of compression.

In this heating process, some little discussed chemical reactions begin to occur in the fuel. Called preflame reactions, these take the form of slow, partial breakdown of the least durable types of fuel molecule. Fuel hydrocarbons have three basic forms; straight carbon chains, branched chains, and ring structures. Temperature is a measure of average molecular activity, but there are always some gas molecules moving significantly faster than the others. These faster moving molecules hit and break some of the less durable fuel molecules, dislodging some of their more weakly bonded hydrogen atoms. This released hydrogen is very reactive (normally hydrogenous travel in bonded pairs, but his is atomic hydrogen) and instantly pairs with an oxygen from the air to form what is called a radical, a chemical fragment that is highly reactive because if contains and unpaired electron. Its attraction for the missing electron is so great that it can snap one out of other chemical species it happens to collide with, thereby breaking it down as well.

At some point in the compression stroke, the spark ignites the mixture and combustion begins. The burned gases, being very hot, expand against the still unburned charge, compressing it outward into the squish band. This compression rapidly heats the unburned charge even more, accelerating the preflame reactions in it. As a rule of thumb, the rate of chemical reaction doubles every seventeen degrees F. All this while, the population of reactive molecular fragments radicals is increasing in the unburned endgas. If this process of heating takes long enough, and reaches a temperature high enough, this population of radicals becomes great enough that its own reaction rate one radical creating more and more through further reactions accelerates into outright combustion. This is autoignition.

Now why does this heated, chemically changed endgas detonate instead of simply burning? The fuel in the endgas is no longer ordinary gasoline. The preflame reaction that have taken place in it have changed it into a violent explosive much like a mixture of hydrogen and air, or acetylene and oxygen. It is in a hair-trigger state, being filled with reactive fragments from preflame reactions. When it autoignites spontaneously, combustion accelerates almost instantly because the material is so easily ignited now. The combustion front accelerates to the local speed of sound, igniting the material it passes through simply by suddenly raising its temperature, through the shock wave it has now become."

Detonation doesn't occur in F1 engines, Honda discovered long ago that once past 14Krpm or so, there simply isn't time for the necessary preflame reactions to occur using racing fuel, but if flame speeds are indeed limiting max rpms in F1 might there be answer here somewhere? Some kind of quasi-Diesel means of initiating combustion?

In any case it's a good read and causes one to think about what is going on in that combustion chamber!

[Ray Bell]

Interesting stuff... makes one wonder how old Harry found his answers (don't you love that line... "Those of us who began racing before water cooling arrived .."?) in those dark days of post-WW1...

Always wondered myself about the noise of detonation, how it could be heard while all else was noiseless...

[MacFan]

If flame speed is indeed limiting RPM, why not use 2 or more spark plugs per cylinder, as Alfa did with its F3 engines in the late 80s?

[desmo]

Not a bad idea, but real estate in a modern racing 4-valve head is pretty much spoken for although I understand Aprilia's Superbike engine has both 4-valve heads and twin plugs. 2-valve heads actually flow nearly as well as 4-valvers and they have the places to put double plugs. Maybe even 3? There's just the problems of less circumfrence at low lift and valve mass to control. Back to the future?

[Zoe]

Don't some top fuel dragster engines have two spark plugs per cylinder?

Zoe

[desmo]

Yes, last I knew they all did. But those are 2-valve heads. I tried to find a photo or drawing of the Aprilia RSV head to see how they fitted the TSI(Twin Spark Ignition) into their 4-valve head, but I couldn't.

[Halfwitt]

Didn't at least one of the old Rotax single cylinder racers have three plugs? It was a few years ago, so I can't remember exactly.

I suppose it come down to where to put the plugs, and where this twin-plug combustion leaves this dangerous end gas. Is the twin-plug set-up used on the factory-Aprilia race bikes?

[desmo]

Yes I believe it is.

[MacFan]

Is it actually flame speed that limits RPM? I was led to believe it was the problem of filling and emptying a 300cc cylinder quickly enough without the gases going supersonic, hence the renewed interest in V12s.

If flame speed is the bottleneck at the moment, then wouldn't the engine designers use a slightly longer stroke and smaller bore, thus reducing the distance the flame has to travel?

[Dents]

I'm sitting here looking at my piezoelectric cigarette lighter, and wondering... why cant they develop some sort of plug that is no more than an annode? a teeny filament that would arc against a portion of the cylinder head... or something like that...

hell, the wire inside this lighter is about 1mm across... you could have a ring of these (insulated properly of course) around the cylinder...

this is why I come here... I dont know jack ;)

[Zoe]

MacFan,

what little understaning I have of that matter, the limit of rpm is actually a product of many factors: material stress (from the accelerations etc), internal friction, valve speeds, combustion speed (including flame travel speed), and of course, the ability to have the engine breathe fast enough.

Concerning the speed in which the flame front travels, I understood that of the piston is actually moving faster than the flame front, then there is no more "push" left and the engine basically doesn't produce more power. Maybe I'm confusing the terms flame front speed and combustion time, I dunno.

Another thought: an F1 engine has 300cc displacement per cylinder and revs up to, say 20000 rpm (just a number here) and is supposed to be the forefront of todays technology. Now, the 425cid (7.1 l) V-8 engines of my Cadillacs have about 880cc to 890cc per cylinder and revs up to, being conservative, max. 5000 rpm. Now this engine was designed in 1977, has two valves per cylinder and pushrod valve actuation. The filling ratio (or speed: ccs per time) per cylinder of an F1 engine is about a factor of 1.4 better than that of my old ladies engines.

Or calculated differently by looking at the flow rate: Assuming you have 10 ms time to get 300 cc in and out of a cylinder (20000 rpm and a four stroke engine) than is a rate of 30000 cc/sec.
The Cad needs to get 880 cc in 40 ms, resulting in 22000 cc/sec.. Equally a factor of 1.36, or about 1.4

This improvement seems to be quite small to me.

Am I missing something plain obvious? Am I performing a stupid calculation? Please enlighten me!

Zoe

[desmo]

The difference is that the aerodynamic resistance to filling the cylinder is a function of the square of the airflow velocity. This is related to piston speed, which isn't terribly high in an old Caddy. Thus, easy to fill even big cylinders @ 5Krpm, tough to fill even little cylinders @ 20Krpm. Engineers, please excuse my massive oversimplification of a complex topic!

[Cole-B]

>
>..."Those of us who began racing before water cooling arrived .."?)
>in those dark days of post-WW1...
>
Ummmm... I think Kevin Cameron is a motorcycle racer. I think I used to read his articles in one of the cycle magazines. Desmo's right - an excellent writer. Someone who not only knows his stuff, but can explain it, too.

Outstanding article, Desmo. Once again, many thanks!

[desmo]

Thanks Cole-B, you are correct about Cameron coming from bikes, I should have pointed that out when Ray mentioned it.

[Zoe]

Originally posted by desmo
The difference is that the aerodynamic resistance to filling the cylinder is a function of the square of the airflow velocity. This is related to piston speed, which isn't terribly high in an old Caddy. Thus, easy to fill even big cylinders @ 5Krpm, tough to fill even little cylinders @ 20Krpm.

Agreed (at least the part with the piston speed in my Caddys ;) ) However, isn't the difference in airflow speed too small to make such a huge difference? I mean, a racing engine is much more optimised in every aspect of the air channel, from the intake to valves and the large and short exhaust? Somehow I feel I'm missing a part of the equation here, but then it is pretty late already ;)

I also think that the airflow velocity must be a product of the piston speed and the piston diameter (the bore). When the piston moves "down" it creates a vacuum at a rate depending on the piston diameter as well?

Another aspect that came to my mind is, that maybe those figures I quoted are theoretical maximum values, and it is probably true that an F1 engine is achieving a much higher degree of efficiency than a >20 year old big block.

Engineers, please excuse my massive oversimplification of a complex topic!

Well, actually I am an engineer but definitely not in the field of combustion engines. However, it is never too late to learn and maybe putting things so easy is the only way for me to grasp the most basic things ...

Zoe, confused and heading for the bedroom....

[bugeye]

Desmo, very interesting article, thanks for posting it...


Dents: I believe one of the major man. has just such an engine under development..


Dave

[kike]

Very good article about detonation. Thanks Desmo.
I am very repetitive, but I have to say again there´s no limit because combustion time (CT). Mcfan is right, limit is on 0,6 Mach index inlet valve number. It is incredible but turbulence increase with revs, and flame speed increase with turbulence linearly (aprox) since mid-revs, and little less than linearly on high revs. One more thing, when I am talking about CT interval, I am referring to turbulent CT. (Laminar CT increase with revs, in which interval only 5-10% mass into is burned, and little pressure increase is got.
Let´s make some aprox. numbers:
CT=(2·pi·n/60)·tb
tb=lcc/Vf in seconds
Vf=constant·n
CT=(2·pi·n/60)·(lcc/constant·n)=(2·pi/60)·(lcc/constant)
so as can be noticed it doesn´t depend on revs (suposed: linear increase for Vf with revs in turbulent phase, and Vf a medium flame speed during the interval).
CT:combustion time in radians; tb:time in seconds; Vf: flame speed; n:rpm; lcc: combustion chamber characteristic dimension.
Some aeromodels run about 30000 rpm´s, as can be guessed lcc may be 1/2 of F1 cylinder, but rpm are double too. What I think is that combustion time problem is solved making better inlet pipe design as Vf increase with microturbulence (is not a big swirl) born in it when admision stroke. I know is a pretty delicate theme, and it is difficult to accept that chemical burning time is about 0,0001-0,0005 seconds and so do in premixed turbulent flames into these variable volumes.

[malbeare]

Originally posted by MacFan
If flame speed is indeed limiting RPM, why not use 2 or more spark plugs per cylinder, as Alfa did with its F3 engines in the late 80s?

Well folks there is plenty of room in the sixstroke head for 6 plugs per cylinder, and several direct injectors protected from flame as the upper upside down piston has the intake ports fully open at TDC. The ideal has been realised the intake valve cannot be opened any faster.
A head that produces power:p
malbeare http:/www.sixstroke.com

[desmo]

Another article on combustion:

http://www.llnl.gov/str/Westbrook.html

Of particular interest to me was the section on homogeneous-charge-compression-ignition technology (quick burn rate) although it may have little or no motorsport applications.

[Top Fuel F1]

Originally posted by Zoe
Don't some top fuel dragster engines have two spark plugs per cylinder?

Zoe

The Pro Top Fuel Dragsters and Funny Cars have two spark plugs per cylinder. Each plug is supplied from a separate magneto. The magneto's timing changes during a run. The change is brought about by the settings in some of the pneumatic timers and in some cases by RPM detected by the crank speed.

In any event the main purpose of the two plugs per cylinder is to try to make sure the mass of the hard to ignite nitro methane fuel ignites. This is such a problem that on many many runs you will see one or more exhaust pipes piping white smoke instead of flames. When any given cylinder goes out the car is losing about 800 HP at that point.

Rgds;