| High performance
racing engines are typically low torque, high revving and
Schneurle-ported. With RPM in the 30000 to 42000 range,
propellers must be made of very strong materials. Incredibly, up
to 5 kW can be absorbed in propellers of only 173mm diameter.
Such props are typically glass fibre or carbon fibre set in an
epoxy resin matrix: such propellers are normally hand laid in
moulds, as nobody has yet found a way of mass producing these
props. Otherwise I would be out of business, along with many
other backyard engineers.
The materials and techniques involved in making these propellers
are well known: the method of making moulds, also from composite
materials, has been common knowledge to aeromodellers for at
least 50 years. The design process has also been well known,
since Larrabee and others published suitable algorithms at the
time of the man-powered flight challenges.
Despite this, most props are designed by “cut and try” methods,
which are in fact most effective. This website has tried to
guide this process with various technical articles: the most
productive of these guides has related to the choice of
effective airfoils for use at high Mach numbers and low Reynolds
numbers.
Now it is time to look at one mans experience in moulding props,
and see just how much knowledge has been gained from all his
mistakes.
My need to mould props first arose in 1983, just prior to the
Goulburn World F/F champs at the TSR field, on the Breadalbane
plains. I had imported a stock of F1C props and put them to one
side, just for use at the W/C. Imagine my horror, when with only
days to go, I found that these props were quite useless on my
very agricultural models! I turned to Andy Kerr in desperation.
He had turned out a fine bunch of engines for me, and now I had
no props to grace them. Andy was up to speed right away, and
that same day we had a mould made and props in production. Going
into the last round, I was in 5th place, but disaster struck in
the seventh and final round.
I had a slow start with my OPS15, and launched into a squall,
with the rain tattooing my corneas. The model went off line.
Rolled horizontally down the flight line, bunted vertically
downward, then zoomed up again as the VIT cut in.
Luckily my wing wingler was set to give a great transition off
the top of the zoom, and I still did 2 minutes, but every lost
second was a place, and I ended up 25th. Ah well, thems’ the
breaks!
The method Andy used was to make the mould from aluminium filled
epoxy, with a horizontal split line at LE and TE. This was the
classic method, and has its advantages. I found that people were
prepared to pay $5 each for F1C props: that seemed like easy
money to me. However, series production was not to be relished.
With this type of mould, the carbon/resin mix is pretty much
balanced in space on the lower mould: I found it difficult to
get consistent results, with respect to voids, fill and flash.
So I developed what I call a bathtub mould. The split line are
near vertical, so that the lower mould is a female. The resin is
poured into this tub, then the carbon dropped in and wetted out
in the mould. Hygenic and simple. Or so I thought.
Soon problems became apparent. The mould was not stable. After a
few cycles of heat cures, it became apparent that the epoxy was
shrinking. The props came out thicker and thicker, until the
beautiful hand-formed airfoil became fit for a paint stirrer.
Then, the moulds started to crack, from the pressure of closing
down the male mould.
Even worse, I could not eliminate bubbles in the prop rear
surface. When mixing the resin and hardener, air became
entrapped in the liquid. This plus air already in the carbon tow
was the source of these bubbles. I still remember the moment
that I first observed these pernicious little devils. I had been
trying to get rid of them by letting them sit for 10 minutes,
and waiting for their buoyancy to bring them to the surface. I
leaned forward to get my eye close to the surface, in the hope
of seeing them burst. Nothing. In frustration, I sighed:
instantly, as my warm, moist breath hit the surface, the bubbles
broke surface and positively fizzed out!
I tried getting rid of the bubbles by pumping a vacuum over the
mixture. This did not work. Heating the mould helped, as it
reduced the surface tension. So far, so good. Then I ran into
problems with the release agent, QZ11b silicon. After many
pulls, the wax built up and the finish on the prop surfaces
deteriorated.
Also at this time, I became away that the epoxy I was using
would soften in the direct heat of the sun, despite having a
full heat cure. As an example, a commonly used resin used in
ultra-light aircraft would soften and be useless at only 50C. No
wonder those aircraft were all painted white!
OK, so I was not having much fun!!
Once again, Andy came to the rescue. A CNC milling machine was
coming up for auction. The mill was just a good size for milling
propeller moulds from solid aluminium, with an X travel of
450mm, Y travel of 300mm, and Z of 150mm. What is more, it was a
contouring machine, meaning that the firmware could move the
tool (x,y,z) to (x1,y1,z1) directly. No more hopeless epoxy
moulds! With the generous support of Steve Rothwell, I won the
bids and had it shipped to my home here in Perth. But that is
another story!
After 6 months, I had written my own CADD/CAM software and was
able to make metal moulds. What a revelation! No more mould
shrinkage, no more mould cracking, no more need to use PVA
release agent on top of wax.
Also, no more bubbles! Where did they go? Well, in order to get
the resin to penetrate well into the fibre bundles, I was
preheating the moulds. With epoxy moulds, I still got some
bubbles, even.with this preheating. Now, with the metal moulds,
there were none! I put this down to the high thermal
conductivity of the aluminium. There was simply more heat
getting into the mixture and driving out the bubbles.
Further, with the metal mould, the polished aluminium surfaces
only need a light coat of wax, which was polished anyway. Great
surface finish and no bubbles!
But one more devil was lurking to test my resolve. I was getting
curious shallow depressions on the rear prop surface. Where did
they come from? Well, folks tell you that epoxy does not shrink.
Well it does: not as much as polyester, but enough to wreck a
closed-shell moulding.
My final lesson was this. The epoxy shrinkage takes place mainly
in the liquid phase, which is to say, while the resin is
gelling. That means that the mould cannot be closed finally
until this shrinkage is done. And not too much later, either,
because the resin then is too thick to allow the mould surfaces
to come together.
Deflashing was the next problem. It’s no joke trying to remove
the flash from a 32” diameter propeller, or even from a 7” prop!
The best I could find was carbide coated files: these were good
but soon went blunt. Abrasives didn’t last long: I was getting
desperate. Then I went to the Perth Royal show, and there was a
hawker there with some pretty remarkable tools.
One of these was Diamonal discs. These thin discs appear to have
a fibreglass substrate, coated in a mixture of carbide and
diamond. They mount on a bench grinder, which I run at 3000 RPM,
and which has a dust extractor system . The wonderful thing is
this: they don’t get clogged by the epoxy, no matter how hard
you drive them. They do wear away slowly, but it takes years to
wear them out.
Regrettably, they no longer seem to be available here in
Australia, so I have to get them in from the USA
That is most of the story. Many other problems had to be solved,
but these were the most intractable. Hopefully someone else can
benefit from this wasted life of mine! Although, judging from
the energy and happiness on my 4 beautiful grand-daughters, I
must have done something right!
|
