Index Home About Blog
From: tgl@slee01.srl.ford.com (Tom Leone)
Newsgroups: rec.autos.rod-n-custom,rec.autos.tech
Subject: Re: Water injection
Date: 27 Apr 1994 15:30:30 GMT

EricY wrote an excellent article on water injection.
Just a few nit-picks on details:
 
: Any gains are achieved indirectly...

The biggest gains are indirect, but there is also a small gain in
volumetric efficiency from cooling the intake air (the cool air is denser,
so more of it fits inside the cylinders).  This can be significant on
vehicles where the intake air is very hot, like turbos.  The advantage of
water injection compared to an intercooler for cooling off the air, is you
don't add volume to the intake system (which hurts transient response).
  
: ...  Injecting the water cools the intake charge by evaporating
: the water.  The cooler charge burns slower. The result is is similar to
: using higher octane gasoline...

This is a common misconception about knock and octane.  Higher octane
fuels do not burn slower during the main combustion.  They go through
PRE-flame reactions slower, so they do not spontaneously ignite as
easily.  Attached is an article I posted awhile back about knock, for
those who are interested.

Thanks for a good post!
Tom Leone <tgl@slee01.srl.ford.com>

---------------------------------------------------------------------------

Engine knock is the audible pressure oscillation caused by a form of
abnormal combustion.  Normally, the flame propogates smoothly from the
spark plug to the edges of the combustion chamber.  As it does so, it
compresses the unburned gases ahead of it, which increases their
temperature and pressure.  At these high temperatures, chemical reactions
in the unburned gas can lead to spontaneous, extremely rapid combustion
before the flame front arrives.  This fast combustion can produce very
high (even damaging) pressures and temperatures.

There are many ways to prevent this:
(1) Slow down the pre-flame reactions--use a higher octane fuel.
    (note: this does not slow down the normal flame)
(2) Lower the temperatures:
    (a) Flush out cooling system, make sure thermostat, water pump,
        and other parts are working properly.  You can even install
        a lower-temperature thermostat very cheaply.
    (b) Check the ignition timing (too much spark advance causes
        higher combustion temperatures and knock).
    (c) Make sure that intake air "flapper" valve is not stuck (this
        collects hot air from the exhaust manifold for cold starts).
    (d) Do a general car "check-up" to look for things that might
        be making the engine work harder (and hotter) than normal.
    (e) Eliminate "hot spots" in the combustion chamber, like incorrect
        spark plugs (these come in different "heat ranges") or burrs
        (actually these cause pre-ignition, which sounds like
        knock but is a different phenomenon).
    (f) Remove insulating deposits from the combustion chamber (piston
        crown and cylinder head surface).  This requires removal of the
        cylinder head, which is a pretty big job on most cars.  Some
        people say you can clean them out by spraying water into the
        intake air, but I've never tried this.

You may also be interested to know that engine design affects the tendency
to knock.  An engine which "burns fast" gives less time for pre-flame
reactions, and therefore avoids knock.  Some ways to get faster burning are
central spark plug location (and/or multiple plugs, like the Ford 2.3L
twin-plug), higher turbulence (from swirl or small-diameter valves), and
compact combustion chambers (such as the famous Chrysler "hemi"
hemispherical chamber, or pent-roof four-valve engines).

-Tom Leone <tgl@slee01.srl.ford.com>


Newsgroups: rec.autos.tech
From: John De Armond
Subject: Re: What causes detonation??
Date: Sun, 31 Jul 94 17:32:19 GMT

bx970@cleveland.Freenet.Edu (Donna A. Lilly) writes:

>dheister@magnus.acs.ohio-state.edu (David J Heisterberg) says:

>>  knock - autoignition of unburned charge ahead of the flame front,
>>          this of course occurs after the spark
>>
>>  ping  - a *surface initiated ignition* that may occur after the
>>          spark, or before, in which case it can be really bad
>>

>Bravo!!  I'm still not certain, but I may actually understand something in
>this thread.  :-)  I like this non_"pre"ed explanation.

The only thing wrong with these definitions are that they are at odds
with the conventionally used definitions.  First off, ping is NOT 
pre-ignition.  The sound of pre-ignition is NOT the sharp pinging 
sound of detonation (knock).  It is best described as a rumble that may
not even be audible in a mechanically noisy engine.  The pre-ignition
may be severe enough to trigger detonation and THAT makes the characteristic
pinging noise but the pre-ignition is NOT very audible.  One can see 
pre-ignition on a scope or fourier analyzer using a DC-coupled 
or charge sensitive accelerometer and the signature is very characteristic.

In summary, the terms knock, ping and detonation all refer to the shock-
wave-initiated (this term is showing up in the literature but I don't
wholly agree with its usage) explosive combustion process that occures
away from the flame-front but after the spark.  Pre-ignition is a
highly descriptive term that refers to the onset of combustion due to
hot spots, too hot plug or even simple dieseling BEFORE the spark.

>Well, I'm now wondering is whether pinging and knocking are both
>potentially engine blowing phenonoma.  From your cause based definitions,
>it seems to me that either could be timed to 'make short work of ones
>engine'?  When is the most damaging time for pinging to happen? 

Any time.  If you ever have access to a junk engine with the head off,
experiment.  Get a large hammer and whack a piston crown a good lick.
Notice how the sound is just like that of pinging.  That is because
what pinging does to a piston (and the rest of the bottom end) is
just like whacking it with a hammer.  Pre-ignition is not so mechanically
destructive; it simply melts things.  The old rule of thumb is
that if the hole in your piston is nice and smooth and round, it was
probably melted due to pre-ignition.  If it is jagged and sharp,
it was pounded by detonation.  Either type hole has a detremental impact
on power.

John



Newsgroups: rec.autos.tech
From: John De Armond
Subject: Re: What causes detonation??
Date: Sun, 31 Jul 94 21:55:34 GMT

sorlin@magnus.acs.ohio-state.edu (Steven J Orlin) writes:

>The SAE engineers and university professors you have been consulting don't
>amount to a hill of beans.   

and

>:-)  I am only agreeing with the terminology that I spent a quarter
>learning, and studying.  (and will be back at very soon:)

Mighty powerful words for mighty little experience.

>However, an engine can NOT run in PI for more than a few seconds, because
>the combustion chamber temperature SKYROCKETS.
>None of you people (unless your engine has blown up)  have had pre-ignition
>in  your cars' engine.

Such generalizations are almost always false.  Perhaps when you get a
chance to study the phenomena instead of looking over others' shoulders
(2nd quarter maybe?) you'll understand why this generalization is
wrong.  One common form of pre-ignition that is fairly harmless occurs
at very low engine speed.  If you want to see an example, load an
engine down on the dyno to below about 900 rpm with the throttle wide
open and a piece of platinum wire extending into the cylinder to
serve as a hotspot ignition source.  This is exactly the setup I used
to get scope pictures of of pre-ignition for my upcoming book.  Much
more reliable than trying to induce it at speed.  An engine under
these conditions can operate for quite a long period of time
without damage.

>The extreme heat build up in Pre-igntion is due to the early explosion of
>the gases, and the  extreme heat build up which accompanies the pressurizing
>of this explosion as the piston is on it's way up.  

Nope.  Explosive detonation may or may not (usually does) accompany
pre-ignition but it is not involved in the process.  Pre-ignition is
simply that - the early ignition of the A/F mix before the plug fires.
The rumbling sound that accompanies it stems from the irregular
load imposed on pistons compressing an already buring mix.  If detonation
occurs, it will typically occur around TDC, a LONG time after the onset
of pre-ignition.

John



Newsgroups: rec.autos.tech
From: John De Armond
Subject: Re: What causes detonation??
Date: Mon, 01 Aug 94 07:25:14 GMT

sorlin@magnus.acs.ohio-state.edu (Steven J Orlin) writes:

>SAE spark plug rating engines are one cylinder, variably 'supercharged'
>engines, which are driven daily to the brink of pre-ignition.  That is
>how spark plug heat ranges are 'found', and labeled.
>I am definitely not the best one to describe how the pre-ignition detectors
>work, because I have not worked with those in depth yet.
>However, as a half assed explanation, they are tied in with the ignition
>system.  The detector tries to throw a low voltage across the plug gap.
>This can only happen when the plug gap "atomizes" or something like that.
>I am NOT clear about that.  But, at any rate, at that point, when the
>low voltage can be sent, this is the signal that pre-ignition is about to
>start.   

That is called an ionization gap sensor.  It measures the pressure
in the cylinder by measuring the change in voltage drop across the
spark gap during combustion.  Cambustion of England makes a decent lab-
grade instrument.  There have been some decent papers on
design principles in the last couple of years of SAE Sensors and
Actuators.  This is but one of many techniques and not the favorite 
choice of most people.  Among them:

*	Direct cylinder pressure piezo transducer.  Kistler makes transducers
	small enough to mount inside a sparkplug and sells special plugs already
	fitted with the sensor.  Other sensors mount through the cylinder
	wall head.  The thru-the-head sensor is quite popular with 
	racing mechanics.  SuperFlow offers a set of 8 as an option to their
	dyno.  It is also expected that Honda will introduce a production
	engine using in-cylinder pressure sensors for engine management.

*	Piezo accelerometers.  Not the common knock sensor but the wide-
	band sensors made by Kistler, Endevco, etc.

*	Magnetostrictive crankshaft torsion monitor.  These sensors
	are being considered by Ford for production deployment.  Couple of
	good SAE papers available on this one.

*	Valve displacement sensor, typically a small accelerometer mounted
	against the intake valve stem, it measures the cylinder pressure
	by measuring the linear displacement of the stem caused by pressure
	on the valve head.

*	UV color sensor.  This sensor involves a small sapphire window on
	the combustion chamber backed by a fiber optic light pipe and 
	a "color"-sensitive UV detector, typically a small grating and a 
	linear CCD device.  Combustion onsset and intensity correlates to 
	emissions in the UV band and it turns out that the UV spectrum 
	"hardens" at the onset of detonation.  

*	High speed photography.  Those research engines are nice but it's
	a lot more real world to just slap a sapphire port in the head
	of an engine of interest and take a peek with a high speed camera.

>>Such generalizations are almost always false.  Perhaps when you get a
>>chance to study the phenomena instead of looking over others' shoulders
>>(2nd quarter maybe?) you'll understand why this generalization is

>Or perhaps you don't quite understand it yourself?

perhaps.  But I'd not bet on it.

>>engine down on the dyno to below about 900 rpm with the throttle wide
>>open and a piece of platinum wire extending into the cylinder to
>>serve as a hotspot ignition source.  This is exactly the setup I used
>>to get scope pictures of of pre-ignition for my upcoming book.  Much
>>more reliable than trying to induce it at speed.  An engine under
>>these conditions can operate for quite a long period of time
>>without damage.

>How do you know that the platinum wire was serving as a source of
>pre-ignition rather than detonation?

Sigh.  Connect one channel of a dual channel digital storage scope to 
a piezo sensor and connect the other to a missing tooth crank
encoder.  Send the output of the scope to a PC for analysis.  Run
the engine and confirm no PI or detonation.  Insert the same plug
with a chunk of platinum wire spot-welded to the shell.  Run engine,
observe PI but no detonation.  Plot pressure vs crank angle on the PC
to confirm.  Advance the timing and/or lean the mix until there is PI
accompanied by detonation.  Plot pressure vs crank angle on the PC.
Remove the platinum wire.  Observe no PI or knock.  Advance the timing
until knock returns.  Plot pressure vs crank angle on the PC.
Save all the pretty plots away for future use and declare it Miller Time.
If you already have an engine on a dyno, you can set this up and
run it in a couple of hours.  Pretty trivial, really.

>I would like to know exactly what YOUR credentials are for working with
>pre-ignition, and how you have been able to decipher between it and
>detonation in your testing.

*Chuckle* Why I'm just a good 'ole Tennesse redneck who has built a few
racing engines and won a few races and done a bit of research in my 25
years or so of messing with engines.  You can take or leave what I have
to say; It's ultimately worth what you paid for it.  Hope you're not one
of those clowns who's actually impressed with academic credentials.
Sure way to ruin a budding career and/or make a fool of yourself.

I will comment that you're making the mistake of all engineering puppies.
You're taking the SAE standards as gospel.  Read the preface to any
SAE standard, the part about where they describe the text to follow
as a GUIDE.  Very important.  With the exception of a very few 
instances where the government has unwisely enshrined a particular 
standard into law, they are just that - guides.  Don't use 'em as
blinders to alternative techniques.

John


Newsgroups: rec.autos.tech
From: John De Armond
Subject: Re: What causes detonation??
Date: Wed, 03 Aug 94 05:46:01 GMT

dheister@magnus.acs.ohio-state.edu (David J Heisterberg) writes:

>In article <azx8drq@dixie.com>, John De Armond <jgd@dixie.com> wrote:
>>In summary, the terms knock, ping and detonation all refer to the shock-
>>wave-initiated (this term is showing up in the literature but I don't
>>wholly agree with its usage) explosive combustion process that occures
>
>Ok, since you and Steve Orlin pretty much agree on these terms, lets
>go with it.  Is there necessarily a shock-wave?  Even a well-behaved
>compression wave could create conditions for auto-ignition.  And a
>shock-wave, if sufficient pressure is not maintained behind it, may
>pass too quickly bring about ignition.

True.  It is important to realize that the combustion process
starts long before the spark happens.  These are referred to as
precursor reactions and are the subject of intense investigation.
Normal engine operation is a race between these reactions growing 
exponentially until they constitute pre-ignition or knock, and
the spark happening.  If you were to isothermally compress 
a charge of A/F to 10:1 or so and held it there, it would burn,
just not as fast as with a spark aiding it.  Like any chemical
reaction, the reaction rate increases with temperature.  The mixture
away from the flamefront is heated from the compression of the 
other mixture burning, by radiant heating and shock wave heating
from other detonation nucleating sites.  Detonation usually starts from
more than one place in the chamber.  When an engine is not quite knocking,
it simply means that the mixture is burning just a little bit faster
than the precursors can accelerate.


>>away from the flame- front but after the spark.  Pre-ignition is a
>>highly descriptive term that refers to the onset of combustion due to
>>hot spots, too hot plug or even simple dieseling BEFORE the spark.
>
>What exactly do you mean by simple dieseling?  In the absence of
>inappropriate/fouled plugs or deposits, is non-surface-initiated
>ignition at all common in a modern engine?  At 3000RPM the compression
>stroke is about 10ms, which is still shorter than the auto-ignition
>delay time, right?  

No.  Remember that even fairly large engines can run at fairly high 
RPM using auto-ignition - gas fired model engines.  You're probably
right about diesel fuel; I don't know a whole lot about diesels other
than I know that the auto-ignition velocity is the main limit on RPM.

>So one can have PI without the resulting combustion invariably
>entering the domain of "explosion."  That may be the root of
>Steve's objections to my use (which isn't mine of course, I
>innocently took it out of a book, really) of "ping."

Yes.  As my friend Gary Johns (of Gary Johns Racing Engines in Jacksonville,
Fl) says, a pre-igniting engine sounds like it's grumbling.  Grumbling
cuz it's about to get some holes in its pistons :-)  There's a little
rule of thumb in determining what's happening.  You have to look 
at the engine's balls.  The little balls that collect on the plug's
insulator, of course.  Little black balls indicate detonation.
Little silver balls (incipient piston kit) mean pre-ignition. Both
mean Bluey!  The little silver balls are little chunks of aluminum 
eroded from the piston and melted onto the plug insulation.

John



Index Home About Blog