[Diy_efi] Responses from John Deakin (long!)

[Adam Wade]

In case anyone is terribly interested, I emailed links
to the thread via archives, as well as forwarding him
my last email to the list on the subject.  Here is his
reply, my reply to that, and his final reply (edited
for extraneous/personal content).  Looking forward to
receiving a copy of his book.

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

Date: Sat, 9 Apr 2005 21:07:12 -0700 (PDT) 
From: "Adam Wade" <espresso_doppio at yahoo.com>
Subject: Fwd: Re: [Diy_efi] false myths 
To: jdeakin at advancedpilot.com 

    
Well, I forwarded this to you on the AvWeb site, but
thought I'd send it this way, too, as I prefer real
email anyhow.  :)  Hope you get a chance to peek at my
book (you're in the acknowledgements; maybe we can
trade books!).  And don't take my criticisms too
personally; Gary here was asserting some things about
the physics of combustion, and made reference to your
Pelican's Perch stuff as his justifications for his
beliefs.  You can find his original post here, where
he claims there's no such thing as lean misfire, that
lean running without more spark advance raises EGT
*and* CHT (IME, it should be the former but not the
latter) and that quickest flame propagation = most
power (quickest flame propagation would be detonation,
which ain't best power!): 
<http://www.diy-efi.org/pipermail/diy_efi/2005-March/000100.html>

He was replied to by a professional combustion
researcher and engineer, which started a flame war
between him and the engineer: 
<http://www.diy-efi.org/pipermail/diy_efi/2005-March/000103.html>
and
<http://www.diy-efi.org/pipermail/diy_efi/2005-March/000104.html>

Here's where he referenced you: 
<http://www.diy-efi.org/pipermail/diy_efi/2005-April/000118.html>

My response: 
<http://www.diy-efi.org/pipermail/diy_efi/2005-April/000121.html>

And his specific references: 
<http://www.diy-efi.org/pipermail/diy_efi/2005-April/000122.html>

Many of my complaints are his extrapolating your
comments on the specific environment found in a
four-cycle engine's combustion chamber during normal
operation into rules about how fuel and air behave as
absolutes, in any setting.  That doesn't work.  But
there are a few things that could be honed a bit, if
you cared to (the end point is the same from a pilot's
perspective, whether he/she knows the physics behind
it or not!).  Thanks for keeping us all thinking and
learning!

*snip repost of my last post to the list on this
subject*

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

This is his initial reply...  Here he generally seems
to get confused, thinking that my writing is not mine,
and him criticizing my writing as if I had reposted
someone else's writing to him.  We clear that up later
on.


Date: Sun, 10 Apr 2005 06:41:31 -0700 
To: espresso_doppio at yahoo.com 
From: "John Deakin" <jdeakin at advancedpilot.com>
Subject: Re: Thought you'd like this thread from the
DIY_EFI list 

    
Hi,

I'm at a disadvantage here, as nothing in your message
indicates who anyone involved might be!

Nevertheless...

The whole treatise appears to be from someone with
very strong opinions, and no data in evidence.  Some
is wrong, some is right.

At 17:56 4/9/2005, you wrote:

>>> I don't recall him having addressed anything about
>>> the actual combustion process, and I know I
haven't
>>> read anything about there being no such thing as
>>> lean misfire, or your description of why there is
>>> incomplete combustion from lean operation.

You have to realize, ours has been a journey of
discovery, shared with my readers as we progress.  One
example is the "lean misfire."

When we started out talking about balancing fuel and
air to the combustion chambers, people would talk
about leaning to roughness and say "That's lean
misfire."

No, it isn't.  It is simply one or more combustion
chambers going past the point of "peak power" (about
80 ROP), and starting down the lean side, while some
of the other combustion chambers were still increasing

power, still approaching peak power from the rich
side.  If you get cylinders producing different HP,
you get shaking.  NOT lean misfire.  At that point in
time, we may have even said "There's no such thing as
"lean misfire!"  That's not true, as further research
proved.  It might be better called just "misfire,"
because the exact same thing occurs when ROP, and LOP.
 This is where the mixture is not quite homogenous,
and there are little pockets of "richer" or "leaner"
fuel.  If the spark happens to fire in a "too rich" or
"too lean" pocket of fuel, the combustion event may be
delayed, or it may not take place at all, on that one
power stroke.

As a side note, aircraft engines have a HUGE cycle to
cycle variability, AND a huge cylinder to cylinder
variability, too.


> Here he describes, in layman's terms, flame
> propagation in an infinite volume container with a
> totally non-homogenous mixture (the earth's 
> atmosphere), with the burner of a gas stove.  If we
> were talking about a stratified-charge aircraft
> engine, then this might be generally applicable;
> however, to my knowledge, there are no stratified-
> charge aircraft engines, and the principle cannot be
> applied to an enclosed combustion chamber with a
> homogenous mixture in any way.

Oh, give me a break.  I was trying to illustrate a
point with an everyday event!  Anyone can SEE a "too
rich" or a "too lean" mixture in a variety of ways,
from the gas stove, to a blowtorch, to a welder's
torch.


> Further, his claim of "the fire leaping from
> molecule to molecule" is completely inaccurate, and
> even if it were, it could not be applied from an
> infinite volume container (the atmosphere) to an
> enclosed container (a combustion chamber), due to
> the basics of the Perfect Gas laws.  I'd have 
> thought that was patently obvious to you.

Well, gee, just how does the flame front travel?  Atom
to atom?  A buncha stuff here to a buncha stuff there?
 If my half-facetious use of "fire" offends, in place
of "flame front," point taken.  Go read an engineering
text, I'm trying to understand this stuff on a much
lower level, and to explain it to others.

> His talk of flame propagation in the movies might
> hold true for the common range of atmospheric
> pressures and temperatures, and can be analogized 
> somewhat for any other narrow range of pressure and
> temperatures, but neither pressure or temperature in
> an ICE during a complete operational cycle are
> anywhere close to so constant as those found in the
> earth's atmosphere.  This is where the analogy
fails.

By definition, any analogy fails, eventually.  In
order to begin to understand these events, we need to
start simple, and build.  The alternative is to pick
up a pure engineering textbook, and start at the
beginning, math and all.  I'm trying to make it simple
enough so that I can understand it, and then write
about it in a way that others can grasp it.


> Even under the fairly consistent conditions from
> cycle to cycle in steady-state operation of a
> typical ICE, you can end up with detonation,

What?  You mean you can be running steady-state, and
suddenly detonation occurs for no reason at all? 
Poppycock.


> or a failure to run reliably, with an identical
> mixture to that used when the engine is running
> strongly.  So the idea that flame propagation 
> and speed of combustion vary based on AFR is
> fallacious.

No, it isn't.  We have the data to prove it.


> It's whether there is enough heat generated at the
> flame front to propagate the flame steadily without
> the temperature of the end gas rising to the point
> of combustion (and therefore detonation) that is
> our goal in operating an ICE; mixture is, at best, a
> single part of that very complex reaction.

There are many things that work together to produce
combustion, detonation, preignition, etc.  Mixture
ratio is a major player, perhaps the most important
one.


> Aside from that, he has one mistaken reference to
> the "fuel molecules being close enough together" to
> continue the combustion process (which is totally
> wrong; it's the localized heat where there is fuel
> and oxygen that causes combustion, not proximity to
> anything else.

"Localized heat" can't do anything, unless there is
fuel and air present, and in the proper quantities
relative to each other.


> While there may be very little additional heat
> created before the combustion chamber volume
> increases from sparking a lean mixture, it's 
> the lack of enough heat to continue the combustion
> chain reaction that prevents complete combustion of
> the available mixture, not proximity.

I'm sorry, I guess I'm not smart enough to understand
the logic (?) in that sentence.


> He gets most of it right in this one.  His bits on
> flame front propagation are right on.

> His final comments on detonation are incorrect,
> though.  Detonation quality changes based on how
> quickly the CC volume is increasing, the combustion
> temperature of the various bits of remaining mixture
> (which may or may not be homogenous), and how much
> mixture remains at the point of detonation.  If the
> volume increases rapidly enough, detonation can 
> even be stopped once initiated; the rate of increase
> can control the quality and severity of detonation,
> but what happens during detonation is the same; the
> remaining mixture reaches the combustion point
> before the flame kernel reaches the mixture in
> question.  The detonation may be "spread out" in
> time slightly by variations in the mixture of the
> end gas, and therefore the auto-ignition
> temperature.  But again, it's the same exact 
> principle, with some slight variations in action.

Detonation (when present) starts well before the
combustion chamber volume increases significantly. 
That volume increase will probably act to halt the
detonation, but whatever damage might happen is
already done.  I wonder how he accounts for WIDELY
varying degrees of detonation?  The FAA calls them
"Light," "Moderate," and "Heavy" detonation, with
formulae to calculate index numbers.  We demonstrate
all three on the test stand.

>> and #43.

> Oddly, he gets more of it right in this one (another
> inconsistency!)

Or possibly I'm learning more?  Unlike some, I wasn't
born with the knowledge I have today.  I blush at some
of the crap I wrote ten years ago, in informal
messages in the old AVSIG.


> Again, though, in his discussion of mixture, he
> talks only about conditions during the limited time
> available for combustion, which cannot be
> extrapolated into general theory.  There is not
> enough time for more complete combustion in
> stoichiometric mixtures before pressure and
> temperature in the combustion chamber drop to well
> below ignition temperature (and he doesn't mention
> that there will be a boundary layer of mixture
> against the cylinder wall that won't ignite due to
> the cooling of the mixture via contact with the
> cylinder wall, either, which is typically also a
> large contributor to unburned O2 and HC in the end
> gas; oddly, he mentions it later when explaining why
> the engine doesn't melt, but forgets about it when
> talking about unburned fuel and air!).

Just how much does he expect in a column that covers
just one specific subject?  Does he want a full
8-volume engineering treatise in every column?  My
editors were always fussing at me for "too big," they
wanted 2500 words, and I rarely did less than 10,000
in a column.


> He also gets it wrong in saying that the cylinder
> walls donate heat to the mixture.  It's the
> compression that heats the mixture to where the
> spark ignites it and creates a stable flame front;
> the cylinder walls tend to ROB heat from the mixture

I don't think I said cylinder walls "donate" heat, but
if I did, I apologize.  Very hot cylinder walls will
"Rob" LESS HEAT from the combustion event, which makes
it hotter.  That can contribute to detonation.  We
know that if we hold all factors constant on the test
stand, and simply run the CHT up (close off some
cooling air), we can bring on detonation, and end it,
simply with a change in cooling air.


> Also, there is only one flame front, unless there is
> another point of ignition.

My columns are specifically for aircraft engines,
which do have two spark plugs.

You know, we get guys like this in our seminars, which
generally start on Friday evening, and run through 3
pm Sunday.  They sit there with their arms folded, a
glare on their face, obviously not willing to be
convinced.  About 2pm Saturday afternoon, the
compelling data they see suddenly sinks in, and they
realize they don't know quite as much about this stuff
as they thought they did.  Neither to we, for that
matter, but we're learning, slowly.


Best...
John Deakin
Fly-Bye-Knight Press http://www.flybyeknightpress.com
Advanced Pilot Seminars http://www.advancedpilot.com
Index to all columns: 
http://www.avweb.com/news/columns/182146-1.html

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


Date: Mon, 11 Apr 2005 14:52:57 -0700 (PDT) 
From: "Adam Wade" <espresso_doppio at yahoo.com>
Subject: Re: Thought you'd like this thread from the
DIY_EFI list 
To: "John Deakin" <jdeakin at advancedpilot.com> 

    
John, I have to apologize.  :)  The DIY_EFI mailing
list talks a lot about the actual physics inside the
combustion chamber, and how those effects are divided
up between the various operating parts; spark,
mixture, combustion chamber, and how those things
affect combustion.  There are a number of folks who do
research on combustion science professionally (Phil
LaMovie is one of them, the guy who dismissed 'gary').
 Gary admits to being a layperson, and says in his
posts that he got his three key ideas that we disagree
with from GAMI's and your writings.

My post which I forwarded you was about separating out
the actual physics of burning things from the
operating conditions inside the engine, since Gary was
extrapolating backwards; "This is how it works in an
engine, with cool walls and a changing volume over
time; therefore, this is how it works everywhere,
regardless of physical constraints".  As I noted in
another post to the list, you are trying to give
pilots enough information so they can understand a bit
of the process that underlies your operating
instructions; turning skeptics into believers.  I'm
100% in favor of that, and think you're doing a truly
excellent job.  However, you don't separate out the
environmental variables (combustion chamber shape,
size, material, changing volume, cooling factor, etc.)
from the mixture, spark, and combustion event itself. 
You have no reason to!  But because they are not
separated out, you can't extrapolate backwards to
conditions where more than a few of the variables have
changed.  That's all.  :)

--- John Deakin <jdeakin at advancedpilot.com> wrote:

> The whole treatise appears to be from someone with
> very strong opinions, and no data in evidence.  Some
> is wrong, some is right.

Agreed.  :)

<'Lean Misfire'>

> You have to realize, ours has been a journey of
> discovery, shared with my readers as we progress. 
> One example is the "lean misfire."

> When we started out talking about balancing fuel and
> air to the combustion chambers, people would talk
> about leaning to roughness and say "That's lean 
> misfire."

> No, it isn't.  It is simply one or more combustion
> chambers going past the point of "peak power" (about
> 80 ROP), and starting down the lean side, while some
> of the other combustion chambers were still
> increasing power, still approaching peak power from
> the rich side.  If you get cylinders producing
> different HP, you get shaking.  NOT lean misfire.

I 100% agree with you.  I pointed out that GAMI
*never* says lean misfire does not exist; just that
lean misfire is not the culprit in this case.

> At that point in time, we may have even said
> "There's no such thing as "lean misfire!"  That's
> not true, as further research proved.  It might be
> better called just "misfire," because the exact same
> thing occurs when ROP, and LOP.

Absolutely, and I pointed this out to Gary in another
post.  Given the range of operating conditions within
a combustion chamber, there are mixtures that are too
rich or too lean to create a flame kernel and/or to
propagate the flame kernel if it comes into being. 
Even with perfectly homogenous mixtures, this can be
the case (although as we know, there is never a
'perfectly' homogenous mixture inside a combustion
chamber).

> This is where the mixture is not quite homogenous,
> and there are little pockets of "richer" or "leaner"
> fuel.  If the spark happens to fire in a "too rich"
> or "too lean" pocket of fuel, the combustion event
> may be delayed, or it may not take place at all, on
> that one power stroke.

At the borderline, where one or more cylinders are not
making power consistently from cycle to cycle, I
agree.  At that same point, it is often the case that
the flame kernel will originate in a pocket of
'acceptable' mixture, but some or all of the remaining
mixture will not ignite, due to a combination of
pressure, temperature, and mixture in a given
'pocket'.  I'd call that 'incomplete combustion', when
compared to a typical cycle in a 'normally-operating'
engine.

Go past the borderline mixture and you'll end up with
no cylinders making any power at all, and that's where
we're into true "misfire"; no flame kernel is ever
generated, no power is produced.  Everything is
negative power going to friction and pumping losses.

> As a side note, aircraft engines have a HUGE cycle
> to cycle variability, AND a huge cylinder to
> cylinder variability, too.

Same with all the four-stroke engines I've
investigated, even with flow-matched fuel injectors. 
Amazing how well things go considering how chaotic
these systems really are!

>> Here he describes, in layman's terms, flame
>> propagation in an infinite volume container with a
>> totally non-homogenous mixture (the earth's
>> atmosphere), with the burner of a gas stove.

> Oh, give me a break.  I was trying to illustrate a
> point with an everyday event!

I know.  But Gary also used your gas burner analogy to
validate his physics, and I was pointing out that the
physics were fairly different inside a combustion
chamber.  Changing volume and pressure, changing
temperature, and the fuel much more distributed around
the enclosure than with a gas burner (which is close
to a point source).  I wasn't trying to nitpick,
honest!  Just showing Gary that is was a lay analogy
which couldn't be used to validate physics.

>> Further, his claim of "the fire leaping from
>> molecule to molecule" is completely inaccurate, and
>> even if it were, it could not be applied from 
>> an infinite volume container (the atmosphere) to an
>> enclosed container (a combustion chamber),

> Well, gee, just how does the flame front travel? 
> Atom to atom?  A buncha stuff here to a buncha stuff
> there?  If my half-facetious use of "fire" offends,
> in place of "flame front," point taken.  Go read an
> engineering text, I'm trying to understand this
> stuff on a much lower level, and to explain it to
> others.

So am I.  Give me a shipping address, I'll send you a
copy of my motorcycle EFI book.  The first full third
of the book is about the physics of four-stroke
engines, and the combustion event.  And it's written
for laypeople, so they can really understand what goes
on.  So far, I've gotten good reviews from laypeople
and technical experts alike, and I hope the work
stands up over time.

In any case, what causes combustion is heat.  The
flame front is much hotter than what it touches, and
if it adds enough heat to cause combustion in the
adjacent mixture, the flame front grows.  Since the
volume of the combustion chamber is increasing as the
flame front propagates, the pressure (and therefore
temperature) drops; normal combustion is a tricky
balance of engine and combustion speed.  Should
combustion progress too slowly, the pressure in the
combustion chamber will drop, such that the combined
temperature of the flame front and the ambient
temperature of the unburned mixture cannot sustain
continued combustion, and power is lost (or absent). 
If the flame propagates TOO quickly, pressure (and
heat) rise in the remainder of the mixture to the
point where the remaining mixture auto-combusts; that
is detonation.  We want to keep things in the
combustion chamber such that the increase in temp and
pressure from the burn match the increase in
combustion chamber volume.  With gasoline engines, we
use a spark plug to create plasma, which forms the
heat core of the flame kernel and starts the whole
process.  Diesel engines have the same physics going
on, but they have no plug; instead they use phenomenal
compression ratios to CAUSE detonation, and that
detonation is what drives the engine's power output. 
Diesels can run extremely lean mixtures and still run
fine (this is why they are economical for large boats
and tractor-trailers); same physics, but different
operating conditions to some degree.

Your analogy is plenty good enough for a pilot
managing his controls in his piston-engined aircraft. 
He doesn't need to know the nitty-gritty of flame
front propagation (although he might benefit from the
details, if he had interest).  Again, I was only
pointing out to Gary where his claims did not hold up,
despite having claimed that he used your articles for
his beliefs.

> By definition, any analogy fails, eventually.

Point.  :)

> In order to begin to understand these events, we
> need to start simple, and build.

This is an argument I've long had with myself, being a
technical author and columnist.  Should I use
analogies that are not technically accurate, but get
the gist across?  Should I use a more complicated
analogy that might take more processing to understand
but is more technically accurate?  I like to think
that I spend a lot of time trying to get both
simplicity and technical accuracy in the same breath. 
I do know how challenging that is.

> I'm trying to make it simple enough so that I can 
> understand it, and then write about it in a way that
> others can grasp it.

And I in no way fault you for that.  Although I hope
you can find some interesting things in my book that
will further your understanding of combustion events
in a four-stroke engine.

>> Even under the fairly consistent conditions from
>> cycle to cycle in steady-state operation of a
>> typical ICE, you can end up with detonation

> What?  You mean you can be running steady-state, and
> suddenly detonation occurs for no reason at all?
> Poppycock.

Of course not.  Let me re-read the context from my
original email.

Ah, okay.  You cut off the important part!  I was
saying that MIXTURE ALONE does not produce or prevent
detonation.  I can use the same mixture to run fine at
a steady state, or to create horrible detonation,
depending on other operating parameters (spark timing,
engine speed, throttle position, engine load, and so
on).  I wasn't saying that engines just spontaneously
begin to detonate.  Is that more agreeable?  :)

>> So the idea that flame propagation and speed of
>> combustion vary based on AFR is fallacious.

> No, it isn't.  We have the data to prove it.

That should have read "*SOLELY* based on AFR".  If the
same mixture can deliver strong running and
detonation, dependent upon other variables, then
clearly the rate of combustion is different, while the
mixture is the same.  You can even burn aluminum if
the temp and pressure are high enough; that's what
creates holes in pistons from ongoing heavy
detonation.  But that aluminum is always available to
be burned in the combustion chamber; the "mixture"
there is constant, but only under certain operating
conditions does the aluminum actually burn.  Exact
same principle.

I will agree that *under normally-seen operating
conditions in four-stroke engines that are running*, a
large part of the burn rate is due to the mixture in
question (with the remainder being spark timing, CC
wall temp, combustion chamber shape and size, and
throttle position, by and large).

> There are many things that work together to produce
> combustion, detonation, preignition, etc.  Mixture
> ratio is a major player, perhaps the most important
> one.

For detonation and pre-ignition, especially at or near
full throttle, AFR is perhaps the most significant
player, yes.  And since most piston-engined aircraft
don't have pilot-operated spark advance, that's out of
our hands for all practical purposes (although not for
those who are tinkering with their own EFI systems,
thus my increased focus on it in the mailing list
posting).

> "Localized heat" can't do anything, unless there is
> fuel and air present, and in the proper quantities
> relative to each other.

Well, yes and no.  There must be oxidizer present, and
some form of fuel present (which is anything that can
be oxidized).  How MUCH heat is needed to burn a
certain "bunch of stuff" in a finite area is something
else entirely.  Any AFR will burn, as long as there is
fuel and air, PROVIDED you introduce enough heat.  As
I noted above, you can burn aluminum in a combustion
chamber if the temperature is high enough.  Different
AFRs need different amounts of heat to cause
combustion, and even then there can sometimes be more
complete combustion (releasing more energy) with a
higher local temperature.  Diesels are again a prime
example; without ANY spark, they can reliably ignite
mixtures as lean as 30:1.  Now, can we expect certain
AFRs not to combust 'normally' given the changing
variables found in an internal combustion engine? 
Absolutely.  Normal operating conditions are very
particular, and are in a fairly narrow range of
possible physical scenarios for combustion in general.
 It may sounds like I'm splitting hairs here again,
but it is only to show Gary that the "rules" for a
fairly narrow set of specifics cannot be backwards
extrapolated to the general case.

>> While there may be very little additional heat
>> created before the combustion chamber volume
>> increases from sparking a lean mixture, it's the 
>> lack of enough heat to continue the combustion
>> chain reaction that prevents complete combustion
>> of the available mixture, not proximity.

> I'm sorry, I guess I'm not smart enough to
> understand the logic (?) in that sentence.

*chuckles*  Okay.  With less fuel to burn in the above
case, as compared with 'correct mixture', less energy
is released within the flame kernel.  This means the
flame kernel is not as hot, and does not grow as
quickly.  As the cylinder travels downward, volume
increases, lowering the temperature still further.  If
this continues, the remaining unburned mixture will be
at a low enough temp that contact with the flame front
will not be enough to cause combustion in that
mixture.  The remaining mixture does not burn, and you
only get the energy from the amount of mixture that
was burned initially.  HOWEVER, a lean mixture means
that more of the energy available from totally
complete combustion of the available fuel is
liberated, albeit at a somewhat slower rate. Sometimes
this is enough to balance the slower rate of
combustion; sometimes it is enough to cause localized
temperatures that are hot enough to damage the piston
crown WITHOUT causing detonation in the remaining
mixture!  This is rare, but entirely possible with the
right combustion chamber design.

> Detonation (when present) starts well before the
> combustion chamber volume increases significantly.

Exactly.  If it started later, it would be much less
severe, since the pressure drop (and thus temp drop)
from the increasing volume would drop the remaining
mixture below the auto-ignition temp.

> That volume increase will probably act to halt 
> the detonation, but whatever damage might happen is
> already done.

While not always damaging the piston itself, the big
issue I have seen is that it tends to collapse the
film of pressurized oil in the plain bearings on the
rod big end (primarily) and the crank main bearings
(less so).  This can be a truly ugly thing if not
caught and corrected.

> I wonder how he accounts for WIDELY varying degrees
> of detonation?  The FAA calls them "Light,"
> "Moderate," and "Heavy" detonation, with formulae to

> calculate index numbers.  We demonstrate all three
> on the test stand.

Yep.  It all depends on how much mixture is unburned,
and how fast the combustion chamber volume is changing
as detonation occurs.  The later it starts, and the
less unburned mixture there is, the less damaging the
detonation should become.  The speed of flame front
propagation is an issue here; since lean mixtures burn
more slowly, by and large, but hotter, lean mixtures
are most prone to the more serious levels of
detonation, as they provide A) higher temps and B)
much more available oxidizer for auto-ignition
combustion.

>> Oddly, he gets more of it right in this one
>> (another inconsistency!)

> Or possibly I'm learning more?

Indeed.  :)  I was more pointing out inconsistencies
to poke holes in Gary.  :)  I, too, am always learning
more.  I learned more about spark and flame front
propagation while researching my book that I even
dreamed existed.

> Unlike some, I wasn't born with the knowledge I have
> today.  I blush at some of the crap I wrote ten
> years ago, in informal messages in the old AVSIG.

If things are going well for you and you have some
spare time, you might want to add addenda to the older
articles, tweaking them a bit.  Or you could hire,
say, a technical writer to help with that.  *nudge
nudge*  :)

> Just how much does he expect in a column that covers
> just one specific subject?  Does he want a full 8-
> volume engineering treatise in every column?

Well, *I* don't.  Again, you're not trying to teach
the exact physics; you're trying to teach pilots how
to better operate their recips.  Mostly this was for
Gary, again, to show him that he should think twice
before basing his physics of combustion on lay
articles.  I think you do a damned good job with your
articles, and you're introducing important information
to people who have thus far not had access to it.  I
truly commend you, and I consider you one of my
inspirations in writing my own book, and writing
technical articles for magazines as well.

> My editors were always fussing at me for "too big,"
> they wanted 2500 words, and I rarely did less than
> 10,000 in a column.

*laughs*  I have the same problem.  I was contracted
for 40k words in my book, and I delivered 118k!  :D  I
hope you have good reading glasses, because the
publisher made the print pretty small to get most of
those extra words in there.  ;)

>> He also gets it wrong in saying that the cylinder
>> walls donate heat to the mixture.

> I don't think I said cylinder walls "donate" heat,
> but if I did, I apologize.  Very hot cylinder walls
> will "Rob" LESS HEAT from the combustion event,
> which makes it hotter.

Yep, that's bang-on with what I have seen.  Let me see
if I can find the reference I was using...

I think I may have been referring to the bit about the
combustion chamber walls donating heat during the
intake stroke, as I found no other reference.  I was
probably over-zealous there; certainly the combustion
chamber walls (and piston crown) will heat the
incoming mixture at least somewhat.  The intake valve
is a place where a lot of the work is done of creating
and maintaining vaporized fuel; this is one reason why
intake valves run cooler.  They donate a lot of their
heat to the droplets of liquid fuel, helping vaporize
them.  Swirl inside the combustion chamber does "the
rest", hopefully giving us a fairly homogenous mixture
(depending on combustion chamber and intake tract
design, as well as method of fuel delivery).

Oh, and we were talking about detonation here, aye?
Any heat added will move you toward detonation,
strictly speaking.  I guess I was speaking in overall
terms, where taking the net heat exchange between
combustible mixture and combustion chamber walls, the
walls take more heat than they donate, causing a lower
likelihood of light detonation.  Heavier detonation
has such huge increases in pressure and temp that the
cylinder walls cannot carry it away fast enough, and
therefore the net cooling is of no help in moderate to
severe detonation.

> That can contribute to detonation.  We know that if
> we hold all factors constant on the test stand, and
> simply run the CHT up (close off some cooling air),
> we can bring on detonation, and end it, simply with
> a change in cooling air.

Oh, yes, agreed.  I would say that you need to be
running pretty close to detonation already, or would
have to add a lot more heat (less cooling air) to a
typical normally-operating engine.  Would you say
you've found that to be the case?

>> Also, there is only one flame front, unless there
>> is another point of ignition.

> My columns are specifically for aircraft engines,
> which do have two spark plugs.

Ah, see, I didn't know they were all twin-plugged.  MY
mistake, for sure.  :)  I wonder what we would see if
we used two Colortune plugs in a test engine (I'll
fill you in if you're not familiar with Colortune).  I
wonder if the flame kernels join relatively soon after
ignition, or if they remain separate for a longer
period of time.  If the latter, that would speed
combustion significantly, as well as causing a sharper
rise in both pressure and temp during a normal cycle. 
You'd want less spark advance for sure under such
conditions, and you'd move yourself closer to the
margin for operating in detonation, especially if
timing was not retarded compared to a single-plug
engine.

> You know, we get guys like this in our seminars,
> which generally start on Friday evening, and run
> through 3 pm Sunday.  They sit there with their 
> arms folded, a glare on their face, obviously not
> willing to be convinced.  About 2 pm Saturday
> afternoon, the compelling data they see suddenly
> sinks in, and they realize they don't know quite as
> much about this stuff as they thought they did.

Heh!  I gather you're talking about me there, as it
was all my points you were answering.

I am in full agreement with all you've written about
control operation and the observable impact of that
operation (through instruments and engine overhauls). 
I was simply showing Gary where the intermediate bits
(between the input and output of the engine as a
'black box') was not always strictly accurate, since
Gary was making some claims (that lean misfire is a
myth, that peak EGT is at the stoichiometric ratio,
etc.) that couldn't be based on your or GAMI's
writings as justification for the physics behind them.
 None of the things I pointed out are essential for
properly operating a recip safely, with better
economy, and longer life.  Whether you or your readers
consider the details outlined above useful or
necessary is another question entirely.  Me, I went to
great lengths in my book to try and be as technically
accurate as possible in the physics underlying it all,
even though 99% of the people out there would never
know the difference.  I just hate teaching things that
are "off the beam" and leaving someone having to
un-learn them later, or worse, making an expensive
mistake based on what they'd learned from a book *I*
wrote.  But I'm anal retentive that way, too.  :)  I'd
love to attend one of your seminars, actually, as I
feel the "hands-on" of it would give me a much more
intuitive feel about a lot of things I know as "book
learning".

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


Date: Mon, 11 Apr 2005 21:45:55 -0700 
To: "Adam Wade" <espresso_doppio at yahoo.com> 
From: "John Deakin" <jdeakin at advancedpilot.com>
Subject: Re: Thought you'd like this thread from the
DIY_EFI list 
CC: "gary" <gas- at charter.net> 

    
Adam,

Nice to put a name to all this!

And I owe you the apology, I probably came across as
testy and cranky.  All through that, I kept thinking,
"He's right, but not in this context."

I've looked very hard, and I just can't find much to
quibble about in your text.  I thank you for the time
it took to give such a comprehensive reply.  In
context, too!  <grin>

Will the rest of the list see your response, and this
one of mine?  I hope so, it clears up a lot.

I do have a couple further comments, hopefully a bit
of light without heat, this time.


> I 100% agree with you.  I pointed out that GAMI
> *never* says lean misfire does not exist; just that
> lean misfire is not the culprit in this case.

True, but I'm afraid I'm guilty of saying "No such
thing as lean misfire" in the broader sense.  Pretty
sure I'm guilty of that, but too lazy to go look. 
<sigh>


> At the borderline, where one or more cylinders *are
> not making power consistently from cycle to cycle*,
I
> agree.  At that same point, it is often the case
that
> the flame kernel will originate in a pocket of
> 'acceptable' mixture, but some or all of the
> remaining mixture will not ignite, due to a
> combination of pressure, temperature, and mixture in
> a given 'pocket'.  I'd call that 'incomplete
> combustion', when compared to a typical cycle in
> a 'normally-operating' engine.

I'm twisting a little in my chair at the text
highlighted between the stars.

I guess it depends on how we define
"normally-operating," probably the reason you put
quotes around it.  We see a LOT of this on the test
stand, looking at cycle-to-cycle variability with
pressure transducers capable of sampling rates of up
to a million times per second.


> Same with all the four-stroke engines I've
> investigated, even with flow-matched fuel injectors.
> Amazing how well things go considering how chaotic
> these systems really are!

Yes, it is amazing.  But you may find these aircraft
engines have more cycle-to-cycle variability, AND a
broader area of  "borderline."  I don't know for sure,
because I have no experience with your engines.


> I know.  But Gary also used your gas burner analogy
> to validate his physics, and I was pointing out that
> the physics were fairly different inside a
combustion
> chamber.

Ok, let's both beat Gary up!  <grin>  (Just joking,
Gary, we're all exploring, here).


> In any case, what causes combustion is heat.

Awright already, so the heat jumps from molecule to
molecule, then.  <grin>


> The flame front is much hotter than what it touches,
> and if it adds enough heat to cause combustion in
the
> adjacent mixture, the flame front grows.  Since the
> volume of the combustion chamber is increasing as
the
> flame front propagates, the pressure (and therefore
> temperature) drops

Ok, here's maybe one quibble.  You talk almost
exclusively (in these messages) of the effects on the
combustion process as if there is only combustion
during the power stroke after TDC.  I KNOW you know
better.  We cover the entire OTTO cycle, with strong
emphasis on the compression and power strokes, and we
treat the combustion that occurs before TDC as
"anti-work," including temps and pressure changes. 
Good engineering term, that!  <g>
  
By the way, aircraft engines almost always operate
with fixed-timing spark at 20 to 22 degrees before
TDC.  The Malibu engine (no other physical
differences) have 25 BTDC timing, but only because
that engine is REQUIRED to operate LOP.  To compensate
for the advanced spark at full takeoff power, GOBS of
extra fuel are shoved in, slowing the combustion of
the obscenely rich mixture back to where it would be
with spark timing in the 20-22 range, and less fuel.

(Years ago, the big Curtiss-Wright 3350 DID have
switch-selectable timing to advance the spark several
degrees during cruise at LOP mixtures.  Wonderful
system, wish we had it today, but FAA certification is
a bitch.

PRISM is a new electronic ignition system that
measures the "theta PP" (angle past TDC of the peak
combustion pressure) on every power stroke, and
automatically corrects the next spark to move theta PP
to the ideal 18 ATDC (varies a little, depending on
other factors).  This will revolutionize
gasoline-powered aircraft engines, and allow us to get
rid of all lead in fuels.  Works on a lawn-mower, too.


>> By definition, any analogy fails, eventually.

> Point.  :)

Thenkyew!  <g>  You're very gracious.


>> In order to begin to understand these events, we
>> need to start simple, and build.

> This is an argument I've long had with myself, being
> a technical author and columnist.  Should I use
> analogies that are not technically accurate, but get
> the gist across?  Should I use a more complicated
> analogy that might take more processing to
understand
> but is more technically accurate?  I like to think
> that I spend a lot of time trying to get both
> simplicity and technical accuracy in the same
> breath.  I do know how challenging that is.

I believe you do.  I look forward to seeing how much
of your book I understand.  <g>  You and George would
be right at home with each other, I'm constantly
telling him, "Ya gotta dumb it down, George."  He
wants to put some incredible engineering data in the
presentation, and he would cheerfully spend hours
going over it loving detail.  I have literally walked
up to him in front of 75 people, clamped a hand right
over his mouth, and hauled him off the stage.  This
took more than a little courage, for he is 6' 2", and
300 pounds, a "formidable presence."  Fortunately, he
took it good humor -- I think.  <g>


> Of course not.  Let me re-read the context from my
> original email.

> Ah, okay.  You cut off the important part!  I was
> saying that MIXTURE ALONE does not produce or
prevent
> detonation.  I can use the same mixture to run fine
> at a steady state, or to create horrible detonation,
> depending on other operating parameters (spark
> timing, engine speed, throttle position, engine
> load, and so on).  I wasn't saying that engines just
> spontaneously begin to detonate.  Is that more
> agreeable?  :)


Much!

As a point of interest, it is just about impossible to
make a normally-aspirated aircraft engine detonate, if
the engine and fuel are "conforming."  Just can't do
it.  Turbo'd engines are somewhat easier to get there.
 Almost all the failures we see probably start with
preignition, which is far more devastating to an
engine than detonation.  If fact, we believe that most
of these engines would make TBO, running 100% of the
time in "light" detonation.



>>> So the idea that flame propagation and speed of
>>> combustion vary based on AFR is fallacious.

>> No, it isn't.  We have the data to prove it.

> That should have read "*SOLELY* based on AFR".

Ok.  Better.  It's a little simpler in aircraft
engines, with constant-speed props.  We can control
the fuel flow directly from the cockpit, and run the
temps and pressures all over the place with that
control alone.  We normally speak of "WOT" (Wide Open
Throttle), a specific RPM that remains the same unless
the pilot selects a different RPM, and roughly the
same cooling (airspeed).  We can also control the MP
with the throttle, though other factors certainly
enter into that, too.


> You can even burn aluminum if the temp and pressure
> are high enough; that's what creates holes in
> pistons from ongoing heavy detonation.

I had always thought of those holes (I've created a
few in my time) as the aluminum simply melting.  Not
too comfortable with the concept of it actually
combusting, but I know that given enough heat, almost
anything will "burn."  I guess I'm not clear on the
difference between "melting" and "burning" from a
physics point of view.


> I will agree that *under normally-seen operating
> conditions in four-stroke engines that are running*,
> a large part of the burn rate is due to the mixture
> in question (with the remainder being spark timing,
> CC wall temp, combustion chamber shape and size, and
> throttle position, by and large).

Exactly what we teach.


> For detonation and pre-ignition, especially at or
> near full throttle, AFR is perhaps the most
> significant player, yes.  And since most piston-
> engined aircraft don't have pilot-operated spark
> advance, that's out of our hands for all practical
> purposes (although not for those who are tinkering
> with their own EFI systems, thus my increased focus
> on it in the mailing list posting).

Yes.  The predominant difference between aircraft
engines and auto engines is that aircraft engines run
at very high power settings for virtually their entire
lives, with frequent and prolonged use of 100% rated
power (takeoff and climb) while auto engines loaf
along at 10% or 20% rated power, and probably never,
ever, not once seeing 100% power.  The good news is
that the aircraft engines runs for very, very long
periods of time at some constant power, even if high. 
I run my Bonanza at 85% to 90% of rated power all the
time in cruise.   Wide open throttle (sea level MP due
to the turbo), 2500 to 2700 RPM (redline is 2700), but
ONLY about 90F LOP.  Engine loves it.


> While not always damaging the piston itself, the big
> issue I have seen is that it tends to collapse the
> film of pressurized oil in the plain bearings on the
> rod big end (primarily) and the crank main bearings
> (less so).  This can be a truly ugly thing if not
> caught and corrected.

Sounds reasonable.  We have very little data on that,
aircraft engines tend to be well overbuilt for that,
and most damage from abuse occurs in the cylinder, not
the "lower end."


> I, too, am always learning more.  I learned more
> about spark and flame front propagation while
> researching my book that I even dreamed existed.

Ain't it a bitch?  The more we learn, the less we seem
to know.


> If things are going well for you and you have some
> spare time, you might want to add addenda to the
> older articles, tweaking them a bit.  Or you could
> hire, say, a technical writer to help with that.
> *nudge nudge*  :)

I've done that from time to time.  I try to never
change the text, for I would consider that
"revisionist history," but I do insert comments like:

         [Deakin note inserted 2005:  "I was AFU here,
what I should have said was...."]


> I think you do a damned good job with your articles,
> and you're introducing important information to
> people who have thus far not had access to it.  I
> truly commend you, and I consider you one of my
> inspirations in writing my own book, and writing
> technical articles for magazines as well.

Whooee, you do know how to charm, don't you!  <grin>


> Oh, yes, agreed.  I would say that you need to be
> running pretty close to detonation already, or would
> have to add a lot more heat (less cooling air) to a
> typical normally-operating engine.  Would you say
> you've found that to be the case?

The problem in aircraft is that it's all cumulative. 
Misset the mixture a tiny bit, and the CHT will tick
up, keep right on going, and eventually go right into
preignition and detonation (not always in that order).
 It takes a large change in cooling to correct that. 
But this is tough without actual numbers, this is all
very subjective without them.


>> My columns are specifically for aircraft engines,
>> which do have two spark plugs.

> Ah, see, I didn't know they were all twin-plugged.

Well, a very few have had single plugs, but I'd guess
that nowadays all gasoline-powered, certificated
airplanes have completely separate and dual ignition
systems, fixed timing, about 20 to 22 degrees BTDC
(except as noted above), and with the two plugs as
widely separated as the construction allows.  I can
send you pictures, if you wish.  You'll find them
primitive by your standards, but they're not, when you
consider they are designed for continuous operation at
60% to 75% of rated power and above, and long periods
at full rated power.  For that, the design is superb. 
There is much comment about how superior automotive
fuel injection systems are, but what most folks miss
is that when an auto engine is at its highest power
and comparable to the aircraft engine, the system
BECOMES a continuous-flow fuel injections system,
exactly like the aircraft engine.

Spark timing is a major issue, and many have tried to
improve that, with mixed success.  Any aircraft
ignition system has to be dual-redundant, protected
from transmitting or receiving EMI and RFI to roughly
the level of a nuke going off 3 feet away, or a
lightning bolt.  (I exaggerate, but not much!  The FAA
is NEVER satisfied with "better," they insist on
"perfect.")


> I wonder what we would see if we used two Colortune
> plugs in a test engine (I'll fill you in if you're
> not familiar with Colortune).

Please do?  I think I grok it, but not sure.  My
partner George Braly has wired up spark plug tips to
measure temperature under various conditions.  All
else being equal, that temperature peaks at about 25F
ROP.  Agrees with NACA data from 1943, and Lycoming in
1966.


> I wonder if the flame kernels join relatively soon
> after ignition [No], or if they remain separate for
> a longer period of time.  If the latter, that would
> speed combustion significantly, as well as causing a
> sharper rise in both pressure and temp during a
> normal cycle. 

Yes, exactly.  


> You'd want less spark advance for sure under such
> conditions, and you'd move yourself closer to the
> margin for operating in detonation, especially if
> timing was not retarded compared to a single-plug
> engine.

Yes, again.  Aircraft ignition systems can be tested
by switching either plug off (the famed "mag check"). 
On one plug, power will drop slightly, CHT will drop
slightly, EGT will rise dramatically.


> Heh!  I gather you're talking about me there, as it
> was all my points you were answering.

Yup, and I apologize.  You ain't nearly as bad as
everyone says!  <big grin>


> Gary was making some claims (that lean misfire is a
> myth

That's what he gets for believing some nut on the
Internet.  <g>


> that peak EGT is at the stoichiometric ratio

We do teach that peak EGT is "close enough" to
Stoichiometric for all practical purposes.  Can you
expand on that?  Using peak EGT (measured a few inches
outside the exhaust port, in the exhaust pipe) as a
reference, peak power occurs at 80 ROP, peak CHT at 40
ROP, and BSFC at about 20 LOP.  BSFC actually shifts
depending on power, it's about 20 LOP at low power
settings (60%), and 90 LOP at high (85% and more).


Best...
John Deakin
Fly-Bye-Knight Press http://www.flybyeknightpress.com
Advanced Pilot Seminars http://www.advancedpilot.com
Index to all columns:  http://www.avweb.com/news/columns/182146-1.html

| 82 Honda CX500 Turbo (Cassandra)  90 Kwak Zephyr 550 (Daphne) |
| "It was like an emergency ward after a great catastrophe; it  |
|   didn't matter what race or class the victims belonged to.   |
|  They were all given the same miracle drug, which was coffee. |
|   The catastrophe in this case, of course, was that the sun   |
|     had come up again."                    -Kurt Vonnegut     |
| M/C Fuel Inj. Hndbk. @ Amazon.com -  http://tinyurl.com/6o3ze |


		
__________________________________ 
Do you Yahoo!? 
Plan great trips with Yahoo! Travel: Now over 17,000 guides!
http://travel.yahoo.com/p-travelguide