Subject: 3 springs/ chatter/ vibration
From: Paul
Date: 7/9/2000, 7:20 AM


Bill Freeman wrote:

Paul,

The governing equation for vibration response frequency is:

         f = (K/M)^0.5

Where f = natural frequency
          K = system spring constant
          M = system mass

The stiffer springs (assuming they don't stack up and go solid)
will increase the natural frequency.  Most of the time this is
a good thing.  The 3mm wide tip seals will be about 50% heavier
than a 2 mm seal (if same material) and this will lower the
natural freq, generally a bad thing.  If the 3mm is a less dense
material, the freq could be unchanged or even increased.

Increasing the seal mass by 1.5X will lower the nat freq to 0.816 of
what it was before.  Assuming that three springs are 1.5X the stiffness
(I don't know enough about the config to know if this is true)
of the original 2 (is this right?) it will increase the nat freq to 1.22
times what it was before.

In many applications, a higher nat freq is either neutral or is good
since it moves the response freq above the (often fixed)
driving freq.  In some cases, it may move the response freq up
to coincide with a drive freq which was originally above the response
freq.  This may not apply to chatter since I'm not sure what
the driving freq is, or how it occurs.  Some driving freqs are clear
like once per rev for imbalance.

Designing for vibration response is non-trivial and often requires
substantial field testing with the right equipment to measure the
system driving freqs and do a fast fourier transform to extract all
the individual component freqs.  Then the design must be tested
to verify it has no modes (nat freqs) that are very close or on the
drive freq.  If you hit one, the whole thing may soon come apart
regardless of how strong it is in static calcs.  OTOH, I have seen
two of ten "identical" units fail from a torsional vibration problem,
and the other 8 ran for another year, 24/7.   Go figure.  Probably
tolerances.  Mostly science, but a tiny bit of voodoo.  Thats why
they hire test pilots.  :-)  That's also why the pay test pilots a lot.

Bill



Dale Smith wrote:

Gentlemen ... educate me please.  Except for seals failing, (via this
torsional vibration route), why is the RPM of the rotary engine considered
"high" when in excess of 6000 RPM?

First the apex seal mode of vibration is not torsional. If vibrates
in and out of its slot. This is called chattering.  6000 RPM is not high 
for a Mazda rotary. This number sort of happened because the gear box in 
wide use had a reduction ratio of roughly 2:1 and a practical and efficient
prop is limited to about 3000 RPM more or less. If you go to a 3:1 
gearbox... well then you can run the Mazda at 9000 RPM and extract
a lot more HP while keeping the prop to less than 3000 RPM. The number
of blades or the width of the blades will have to increase of course.

The real limit to RPM in a rotary is the weight of the rotor and the
distance the rotor orbits around the center of the e-shaft. In the case
of the Mazda the rotor weighs about 9 pounds and that distance is on
the order of 1-1/2 inches. This mass rotating on a radius of 1-1/2 inches
causes centrifugal loads on the rotor bearings. These loads become 
excessive around 10,000 RPM. A lighter rotor  would have a much higher 
theoretical RPM limit.



Turbine wheels often spin to 40,000 + but are made of exotic metals.

What is to limit these things spinning happily away at 16000+ RPM?
I have seen Paul's fuel burn charts that seem to drop SFC slightly above
6000.

Do ceramic seals have the same spring arrangement?  What is considered
the maximum continuous RPM with ceramic seals?

Yes.
The chatter is caused by many things and does not limit the RPM directly.
There is an issue however. Heavy seals and light springs cause the seals
to "jump over the humps" in the rotor housing at high RPM (Over 8 or 9K). 
Ceramic seals have lower density than steel or iron so this is less 
of a problem. Aluminum impregnated carbon also works and is widely
used in Mazda racing engines as is ceramic. BTW the housing is not
a true trochoid and the hump can be profiled to minimized this effect,
much as a cam shaft is profiled to minimize valve float.
 
Paul Lamar
 
The Aircraft Rotary Engine Newsletter.              Powered by Linux.

http://home.earthlink.net/~rotaryeng/            http://www.linux.org