Hard Disk Generator...
Part # 2 (Dec 28/05)...
The following section outlines 2 (two) separate
components for the Hard Disk Magnet based wind generator.
I have to say that this particular project has
kept my attention and renewed my spirit of curiosity higher than it has been for
quite some time. It is approximated (in the crudest of terms) that the energy
used by the (average) human brain is 25watts... What this represents, be it
either neural activity, heat radiation through the surface of the skin or a
combination of the above and other factors, I don't know.
But if that is true than I feel confident that
I've been burning 28 to 29 watts over the last week or so. This project has had
me consumed, posting questions on boards, reading replies to similar questions,
digging out old text books, checking feverishly every hour or so if there were
answers to my questions...
The last time that I'd so closely approximated an
Apollo or Manhattan Project type intensity was when I was building the original
CNC mill.
As it turned out, being Christmas, traffic was
slower than usual, and I did read quite a bit that had already been posted that
was of tremendous help. As well I did get some very thoughtful replies that were
based on solid experience.
Enhanced Rotor Configuration
The
image to the right is the finalized rotor configuration. This is the third
layout that I've designed beyond the original that is on the preceding
page.
12 Magnets (totaling 24 poles) with 8 coils
arranged as noted.
The initial design problems were related to arbitrarily
laying out magnets in a ring formation based on the size of rotor I wanted as an
end point.
This resulted in odd numbers of magnets and in
magnet counts that left a gap in the coil spacing.
As stated earlier, I did some reading on the
web...
Of the numerous factors that I wanted to address,
or follow in terms of "Best Practices" the orientation of the Magnets,
spacing and the properties of the coil(s) relative to the magnets were a
priority.
Of the numerous postings & generator
"Builds" that I reviewed there was a consistent theme in that the
inner diameter of the coil should fully straddle an entire pole. This was by
chance pretty close with the coil winder already, so not a concern.
The next issue was the wiring of coils in a 3
phase Star or Delta configuration to gain a significant efficiency of the
generator. This was quickly ruled out, as it would spoil the fun in making the
next generator. This one will stay single phase for now.
The point of greatest concern was the spacing of
the magnets, and this is an issue that is still unresolved to this point. The
issue at hand is that "legs" or sides of the coils should be exposed
to opposing magnetic poles in the course of the rotor's movement. This is a
given that I verified in my dated college text, but the issue of using Hard Disk
magnets that have 2 poles per face in very close proximity to each other poses a
problem if you are not willing to break the magnets.
The illustration above is a very close
approximation of what I've implemented below, and it does work contrary to the
logic that the similar poles passing the legs of the coils should cancel each
other out. The next step (not in this section) will be to break the magnets and
reduce the pole count by half.
Effectively this will introduce the dis-similar
pole orientation relative to the poles and either yield the same or a higher
output with less magnet material. (yet to be determined)
The
other pointer that was unquestionable and makes perfect sense is the idea of
using a steel (Ferrous) backing for the magnets to sit on. The idea is that this
would terminate the flux that was previously emanating from the back face,
creating a return path for the flux to the opposite adjacent pole.
In
the test rotor originally build the backing was wood, and the sides and face
were covered with a resin and black sand mixture... I believe that such an
arrangement most probably hinders the potential flux availability to the coils
as a more direct path is established along the sides and face.
The
unique mix of materials on hand and the order in which I was learning new
information resulted in the images above. The left image is actually 2 old
blades mounted in the saw backwards with my trusty angle grinder and a cut off
disk. The point of reversing the blades is so that if things really go bad I may
lose skin to the bone, but there won't be a pronounced ripping and tearing
action at play.
The
image to the right is the teeth being cut from the saw blade as the saw spins
counter clockwise and the cut-off disc spins in opposition. You have to look
close to see the sparks as the digital camera must have a fast shutter speed (or
equivalent), but I can attest to being showered with sparks hot dust and the odd
tooth.
This
really is the limit to how many magnets I can squeeze into place on this disk,
and underscores reason I ended up actually drafting out the placement of mags
& coils.
Although
the blade is relatively thin and even somewhat flexible at this point, it will
become part of the blade hub assembly (cast as 1/2" thick disk).
Again
a second layer of magnets are positions to bump the output of the coils with a
higher density of flux.
The
magnets and face side of the rotor will be filled with resin (no black sand!)...
The
backside of the rotor/saw blade does still allow a small amount of flux to pass
through but barely enough to hold a 3/8" nut in place.
Given
the above improvements over the previous configuration, and a new test coil of
200 winding using AWG#24 wire this is what I get.
The
coil under test is cored as the previous test coils to reach this value.
Eight
of these coils in a single phase configuration should sum to 16VAC, from what
I'm seeing here.
But
there are considerations when converting from AC voltage to DC voltage.
This
meter reads an average or more accurately an RMS (Root Mean Square) AC voltage.
The basic math (as I recently was enlightened) is that the RMS AC voltage is
multiplied by the square root of 2 to determine the peak AC Voltage.
16Vac
X 1.414 = 22.6 Vpeak
This
would be close to the effective DC voltage, but an allowance has to be made for
the voltage drop across the diodes that rectify the AC to DC. The typical
Voltage drop is 0.6V across a diode, and full wave rectification is typically
through a Bridge rectifier which imposes a voltage drop across 2 diodes for
either the positive or the negative half of the AC wave, or 1.2V, leaving an
effective DC output of 21.4Vdc.
Even
if this is too optimistic, there could be up to 36.9% losses, and the output
would still be 13.5Vdc.
This
is an image of the coils output on the scope. I was expecting to see a rather
lumpy malformed sine in support of the magnetic orientation issues as stated
above...
The
peak voltage is lower than would be consistent with the RMS value read on the
meter as I had increase the air gap between the coil and the magnets to get the
scope to trigger on a stable waveform for the camera.
The
appearance of 3 waves closely spaced may be a result of the magnet placement or
an anomaly of how the camera records the screen of the scope, as I definitely
didn't see 3 waves when I took the picture.
The
timescale is set to 5ms and a full wave period is approx 12ms... the reciprocal
of the period duration should resolve to a frequency of 83hz.
Given
that the rotor has 12 (pole pairs) passing the coil at 280RPM (not 220 as
stated on the previous page).
12 X
280 rpm = 3360 cycles per minute
3360cpm/60sec=56hz...
32% deviation doesn't thrill me but the accuracy of the drill press and even the
scope could combine to that margin of error.
Physical
Generator Frame & Spindle
This
front wheel will be sacrificed to the cause.
Bicycles
are an absolute wealth of parts that have solid engineering and reasonable
mechanical tolerances behind them.
This
section is inserted here as I really need a break from the math, electrical
theory and fancy book learning that has given me a headache.
This
time I'm after the hub of the front wheel. the rear wheel has already suffered
the same fate.
Below
to the left is the reassembled front hub, but with the axel and bearing races
from the rear wheel.
The
rear axel is 3/8ths" in dia. as opposed to the 5/16th" of the
front, as well as being a full inch longer in length. The axel is a higher grade
of steel than typical "All-thread" which is made from mild steel. My
basic premise is that if the axel can withstand miles of 50 to 100lb radial load
over a variety of terrain it should withstand the axial load of the blades and
rotor assembly for the targeted 100Watts of output.
I can
appreciate that the load on the axel will be far greater than the 100Watts of
output, as the blades will be identical to the 4ft diameter 3 blade unit already
tested to 200 plus RPM in winds typical for this location.
The
weak point in my estimation is the inner cradle assembly that the bearing races
run in inside the hub.
The
above concern is addressed by casting the generator frame around the hub
assembly.
This
will provide 1/2" thick solid casting on both the front and back of the
spindle for support and mechanical strength.
This
innovation has been a month in the making or more accurately 29 days in the
visualization stage and a day in the making.
The
foam pattern with the spindle embedded in it was CNC'd and hot glued together.
The
idea of integrating as many components into a single casting was a good exercise.
The longish bars that extend outward around the main bearing support will be
drilled and tapped to suspend the coil stator and allow a margin of stator/rotor
gap adjustment.
Mild
steel typically has a shrinkage factor of 0.010" per inch, while Aluminum
alloys are closer to 0.021" per inch. This variance in dimension change may
pose a problem in that as the casting solidifies the steel insert will set-up
internal stresses within the aluminum and lead to a failure of the hub under
stress. My inclination is that the steel will essentially grow in length, definitely
not expand to the full 30 thou as it only will be heated to the temperature of
the molten Alum Alloy (1400F) and cool far more slowly than the casting as steel
has a higher density than the alloy.
The
inside faces of the hub supports (in foam) will be masked off to ensure that the
entire bearing cradles are supported by metal.
The
tail for this generator will also be fixed as in the previous mill.
The
next generation will incorporate the appropriate Tail -boom assembly to allow it
to swing upward on an angle as the blades are pushed aside by excess winds. (in
actual fact I still don't have a clear idea how to accomplish this, but it's not
a burning issue for now.
So
in closing, I'm well pleased with progress to this point. The Blade Hub and above
generator frame will be cast in the new year along with a number of other
castings that have been stalled due to this project.
The
coil properties may still be tweaked a little bit before I wind the final set
that will be set-up in resin as the stator. The idea of jumping one or two AWG
sizes larger to increase current and drop the voltage closer to the 14 - 16VDC
range is appealing, but quite speculative as all the data I have on hand is for
single coils, and they may not behave as I expect when all wired together.