The outer slot has an area of .13 by .070 or .0091
square inches.
so the radius squared would be .0028966.... working
backwards.
The square root of that is .0538.
So the largest section diameter to fill the outer slot
would be double
that or 0.107640. That lends credence to Bill's theory
as a 3/32
section diameter ring is not .096 section diameter but
is about
0.11 section diameter or about 14 thou larger (.014).
And then you have the O-ring tolerance on top of that.
Some might be 0.005
under and some might be 0.005 over. Also any crude in
the slot will not
help matters. Make sure the slots are spotlessly
clean.
Perhaps Mazda themselves ran into Bill's contention
back around 88 and
that is why some housings cracked in this area.
Perhaps they reduced
the section dimensions on the stock O-rings slightly.
This is why we call them experimental airplanes.
Bill, do you happen to know what the coefficient of
thermal expansion
is for red silicone?
The bottm line is measure your water O-rings and make
sure they are
less than .093 +/- .005 section diameter. Particularly
the outer
one.
Paul Lamar
Hi Guys;
I'm in full agreement with Bill's synopsis, and
can personally attest to it by identical experiences in
designing Pharmaceutical equipment.
Paul your numbers are good ones, you are close, but for
practicality this is what we practice when retrofitting
exotic seals (Rings) in general...................
I'll use the 13B existing gland design:
Brand new groove is.............130 x .070 = .0091 area
Lapped groove is................130 x .067 = .0087 "
Ideal o-ring (.094d)+ max tol (.005)...... = .0077 "
This will give you a net area of .0010 or 11.5% for
MAXIMUM thermal expansion of the seal.
So the question remaining is, What is the thermal
expansion of these composite seals?
The manufacturer NOT the Rep is supposed to have this
information thru the Lab(s) who performed the tests.
The other downside is that this type of seals vary in
diameter throughout the same seal.
Butch
Yes there is some variation around the ring and they also appear to be oval
in some areas. Hang in there I will heat one up tonight.
So the one to use then is the .093 as I guessed.
Paul Lamar
Will this really give you useful info if you don't contain
the linear (circumference) growth?
Charlie England
Charlie,
This is a good point, but we have calculated that even with the
high end of the estimated CTE the volumetric increase is far less
than the normal empty space in a properly sized O-ring groove,
as it should be, since this stuff does normally work just fine.
The volumetric change is what needs to be compared for thermal
expansion to cause a fully trapped (hydraulic) high stress effect.
We are just comparing the 2D effects, but the third dimension is
the same for either a skinny O-ring or a fat one, so it isn't a variable
between the two cases of different O-ring sections in the same groove.
If you are trying to decide if the whole groove fills up, you need to
look at the total groove volume and the total O-ring volume. Each
dimension of the rubber increases perhaps 100-200ppm/F, so you
can calculate the volumetric increase. The expanded volume of the
rubber had better not get greater than the volume of the metal groove
with lid.
For example, if we assume the CTE is 200ppm/F and the temp change
is 150F, then each dimension will increase by a factor of 200E-6 * 150
= .03 or 3%. If we have a volume of 1 cubic inch, each dimension is
now increased to 1.03 inches, so the new volume is 1.0927 cubic inches.
So, the volumetric increase for a piece of rubber with a CTE of 200ppm/F
over 150F is about 9%. This is far less than the normal free volume in
an O-ring groove. OTOH, if you tightly fit in an O-ring, and manage to
get less than 9% free space, watch out! .
Apparently Tom's problem was a ring so oversized that it was LARGER
in volume than the groove at room temperature. The force to drive the
hydraulic effect and break the edge of the groove came from torquing
the thru-bolts holding the stack together rather than thermal expansion.
Hope this clarifies the problem
Bill Freeman
Long EZ builder & pilot
BSME, MSME
EAA Technical Counselor
No disagreement with your numbers; I was commenting on the
technique of just measuring the cross-section cold, then
boiling & measuring the cross-section hot, without
considering the growth in overall diameter, in addition to
cross-section growth. If it expands xxx ppm/F & can't grow
in overall diameter, it will get significantly larger in
cross-section than the unrestrained case (in the closed
groove within the engine).
Charlie England
I agree.
We were mainly looking for the coif. of thermal expansion. Since it is apparently
down around 100/ppm per degree F that is not a problem at the temps we are
operating these rings. I did not want to cut a perfectly good O-ring so I could
measure the change in length. No doubt it would be a better way of determining
the coef. of thermal expansion.
Paul Lamar
The AirCraft Rotary Engine NewsLetter. Powered by Linux.
ACRE NL web site.
http://home.earthlink.net/~rotaryeng/ http://www.linux.org