From: "Jon Carlson" <jon@carlsonhome.com>
To: "LML" <lancair.list@olsusa.com>
Subject: FW: why does roughness happen while LOP  and not ROP?
Date: Sun, 6 May 2001 16:37:09 -0700
Message-ID: <EGEKJCEHGGMPNGLKDEOKKEIACKAA.jon@carlsonhome.com>
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Content-Transfer-Encoding: 7bit
Importance: Normal
Reply-To: lancair.list@olsusa.com

          <<<<<<<<<<<<<<<<--->>>>>>>>>>>>>>>>
          <<  Lancair Builders' Mail List  >>
          <<<<<<<<<<<<<<<<--->>>>>>>>>>>>>>>> 
>>
I received this excellent (and interesting) reply from George Braly to my
last question regarding the mechanism causing engine roughness, but it for
some reason didn't seem to make it to the list (even though properly CC'd,
as far as I can tell).

Anyway, thought I'd take the liberty of duly forwarding it in its entirety.

-Jon C.


-----Original Message-----
From: George Braly [mailto:gwbraly@gami.com]
Sent: Saturday, May 05, 2001 7:08 AM
To: 'Jon Carlson'; lancair.list@olsusa.com
Subject: RE: why does roughness happen while LOP and not ROP?


Jon Carlson asks:

I figured this, more or less, but by what mechanism does roughness come in?

Is it simply due to the imbalanced forces from the combustion events
generating less power? I wouldn't think this to be the case because each
individual combustion event's forces (equal and opposite and all that Newton
stuff) are mostly contained within the engine.

Or, does it have more to do with changing the timing of the combustion
pressure peaks such that the cylinders are now beating against one another
as timed on the crankshaft's rotation rather than being well timed and all
pulling together?

				*************************

It is the imbalance in the net forces from the combustion events.  But it is
more complicated.

Note, if you take a "perfect" cylinder and record the combustion pressure
traces for 100 sequential events, you will be amazed at how much "cycle to
cycle" variation there is in a "perfect" cylinder.

If you plot the mean pressures for each cycle and do a Gausian distribution
you can calculate a standard deviation, etc.

When you operate lean of peak,  the "width" of the distribution gets
"wider".   And wider as you get leaner.

Thus, operating rich of peak at say 75%,  cylinder 1, on average, may be
making 40 hp.  But on 10 of the 100 sequential strokes, it may make more
than 42Hp and on another ten it may make less than 38.

Same cylinder at 75% power, and lean of peak, and on average, it will be
making 40 HP.  BUT, on 10 of the 100 sequential strokes, it may make 44Hp
and on another 10 it may make 36.  +- 4, rather than +-2Hp as rich of peak.
Normally, you would not perceive any roughness from this.  The leaner you
go, the more variation there will be, and you will eventually perceive
roughness, which we have all experienced.

So, now... in a real engine, rather than an ideal engine,  while lean of
peak, if another identical cylinder, say cylinder 6 (next in firing order)
but with 5% less fuel, is operating, then it will be making an average of
38Hp (rather than 40HP), and on 10 of 100 strokes it will be making less
than 33Hp and on ten strokes it will be making > 42 Hp.

Now... this is a statistical deal. MOST of the time,  Cyl 1 and Cyl 6 are
making +- two to four Hp of the same horsepower, and the engine runs
smoothly.  BUT... if the statistics conspire against you, and you get lean
of peak, you can get an event where Cyl 1 makes 44HP and cylinder 6, firing
next, makes only 34 Hp.. and when that happens, you feel the engine "dog
walk" and your wife says, "What was that, honey?"

Thus, the effect of operating lean of peak is that in order to keep the two
Gausian distribution curves of power pulse events sitting on top of each
other, rather than being offset, one has to get the cylinder to cylinder
variation in fuel/air ratios down to a minimum.  We find if the spread from
first cylinder to peak to last cylinder to reach peak EGT is less than 0.5
on engines with the traditional TCM runner log-branch induction systems and
less than about 0.3gph on Lycoming engines and TCM engines with the top down
spider induction systems (TSIO-550 and 520BE and TSIO-360MB and IO-550G) the
engines will operate smoothly lean of peak.

In general, the problem gets worse as you get leaner and leaner, because the
natural variation of a single cylinder gets wider and wider and so the
opportunity to get ever wider sequential variations in power pulses gets
more and more likely.

The exact location of the peak of the pressure pulse has an influence on the
horsepower of the pulse, but it is a combination of the location of the peak
of the pulse and its magnitude that determine the mean effective pressure
(and thus, horsepower) for each combustion event.

Note, also, that it is usually the "couple" of one event happening on one
side of the engine and the next "different" event happening on the opposite
side that produces the perception of roughness.

If you want to experience an exaggeration of this effect, then just put one
bad spark plug in the engine and go fly.  Lean the engine and as the mixture
gets leaner and leaner,  the cylinder with only one functioning spark plug
will cause the engine to operate rough because the cycle to cycle variation
is very sensitive to good spark functioning... and more so at leaner
mixtures, and one plug inoperative will cause that cylinder to operate at
3-8% lower horsepower.

By the way,  I have never seen this explained in this manner in any text
book.

Regards,  George


>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
LML website:   http://www.olsusa.com/Users/Mkaye/maillist.html
LML Builders' Bookstore:   http://www.buildersbooks.com/lancair

Please send your photos and drawings to marvkaye@olsusa.com.
>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>