Bent Genny
Part # 2 - Recumbent Generator (Dec
26/06)...
Where
the last section finished off, a basic drive train,
frame and 16 pole rotor were assembled...
Given
that the original magnet template for the rotor was created on the computer, I
used it as a reference for sizing a "Core Mold" that the coils would
be wound onto.
Again,
since the rotor was laid out as a 16 Pole configuration, and this is to be a
3-Phase generator, the coil count is 12.
Each
mold has a 0.25" center depth and 0.100" backing.
The
mold is painted with 2 (two) coats of a latex based paint to seal the EPS board
from being melted by the polyester resin that will make the cores.
Once
completely dried the mold is liberally brushed down with either petroleum jelly
or generic bearing grease. This acts as a releasing agent when the cores are
cured.
The
cores contain 4 to 5 parts Magnetite or "Black Sand" to one part
resin. (This is sifted from regular silica sand with a pan lined with
magnets, as described elsewhere on this site)
This
dense mixture has a high permeability and aids in drawing the energy producing
"Flux-Lines" from the magnetic rotor through the windings of the
coils.
A
measure of care was employed to ensure that the mold is lying flat so that the
backing has a uniform thickness.
Since
these parts have no dye or paint added to colour them the cure time is
reasonably quick.
I
added twice as much catalyst to the resin/Magnetite mixture as recommended to
accelerate the cure time further.
Even
with 2 coats of paint over the molds and a skin of grease, the vapours start to
soften the EPS mold within 15 to 20 minutes.
A
simple fixture is made from 2 (two) Hard Disk Mags to ensure that every Core is
center drilled (for the Coil-Winder).
As an
aside, the drill shown is a generic HSS (High Speed Steel) bit that needs to be
sharpened after about 8 Cores due to the abrasive nature of the black sand.
Below
is a quick & dirty coil winder that is simply a short length of 1/8th"
band steel, clamped in a vice with a hole drilled, 1 1/2" machine screw,
nut & washers.
The
core slips onto the screw, a wooden crank/handle that has a 1/16th" pocket
machined that mates to the core with a hole drilled to align as well.
The
whole assembly is lightly oiled and fastened with a wing-nut.
The
above accounting is noted as it ensures that the coils are wound in the
same direction... If you were to wind one backwards it would literally un-wind
the assembly and the wiring would ball-up about the machine screw.
I
wound a number of test coils with various gauges of wire ranging from 250 turns
of AWG #24 to 75 turns of AWG #19 and a few in between.
The
objective in this exercise is to determine the maximum amount of copper that can
be shoe-horned into a coil while hitting a predefined voltage and current
relative to the load that will be applied.
I can
appreciate the groans echoing across the web as you start shouting at your
screen that I've omitted any mention of "Gauss, Tesla or Weber" with
regard to the mass of magnets and their velocity relative to the coils that they
ultimately influence. But there are a number of solid references to that math
just a few short key-strokes away... and I don't feel very comfortable simple
regurgitating it and possibly quoting a few equations out of context, or worse
in out-right error.
Essentially
by successive approximation I arrived at 125 turns of AWG #24 (times 2, or 2
in-hand as the real genny builders would say).
Coils
like life itself, are devised upon a series of compromises. It's a balancing act
between high output voltage and low coil resistance, which translates into high
current. As well I still have about 3/4's of a spool of #24 from the last
attempt at building a generator.
I
can't stress enough that #24 is VERY thin and has too high a resistance for
these types of Axial Flux Generators. By winding 2 (two) strands in parallel on
the coil the resistance drops to a quarter of a single strand... a) there are
half as many turns, and b) there is twice as much copper along the path of the
coil... I could have also tried winding 70 or so turns of #24 X 3 strands
in-hand, but I think it would have been too low a Voltage, the same as the AWG
#19...
Keep
in mind that these coils have to fit into a confined space forming a tight ring
concentric to the Magnets on the rotor.
In
the course of the coil testing it quickly became obvious that the 1:2.5 gearing
was inadequate to achieve any meaningful output for an extended period of time.
Yes, I could pedal like mad for 2 or 3 minutes but certainly not for 30 minutes
much less an hour to actually generate some reasonable power.
So...
I stripped the gearing off the frame and laid down a 2 X 6 and fastened it with
lots of 3" wood screws.
The
additional width & length would be needed to accommodate a simple Jack-Shaft
to increase the gearing...
I sat
out in the barn for a few minutes while I waited for some inspiration...
I
could have used another crank assembly, but felt that their far too utile for
such a mundane application.
Ultimately
I settled for the hub/axle assembly including the free-wheel cassette off a
rather twisted rim.
The
axle & bearings are removed for a cleaning prior to welding a second
sprocket onto the hub.
The
sprocket is held with 4 or 5 beads applied at as low a current as possible...
The
axle is returned and bolted to a pair of uprights and welded as shown.
The
new gearing 1 : 6.5...
The
rotor just wails now and with 2 free-wheel drives in the drive chain it spins
freely for 10 to 15 sec after the pedals come to a rest.
If
the rider, cranks too sharply on the pedals it is possible to twist or flex the
jack-shaft uprights and throw the chain.
In
later images note the additional bracing to minimize this problem
Knowing
ones weaknesses is a strength... For my self, assembling odd-shaped objects with
any measure of accuracy by eye would fall into the camp of a
"weakness".
With
this in mind I'd learned from the previous generator that I had to develop some
sort of fixture that would ensure that I didn't fuck-up the alignment of the 3
coil phases relative to the rotor.
The
mill has routed a 0.100" pocket sized to snuggly orient the coils relative
to each other...
I can
appreciate that this would appear to be a minor detail but was worked out before
I'd made the first cut on either bikes or the lumber of the frame.
As it
turned out, I was glad to have made the effort as everything from the coil
winding through to the assembly of the stator went quite smoothly.
Each
Coil is labeled A, B or C in sequence from an arbitrary starting point. Given
that the individual coils are all wound in phase (the same direction) it is
impossible to wire them wrong provided that each start wire (closest to the
core) connects to the next coils ending wire (furthest from the core).
All
of the A's are connected in series, as are the B's & C's which all end at
the terminals off to the left of the stator.
The 3
Phase stator can be wired for either Delta or Star configuration depending on
the application. In this case I have opted to bring the phases out complete
rather than hard wiring so that I have a measure of flexibility after the stator
is poured with resin.
One
of the reasons 3-phase is so popular is the potential efficiency that it offers
over a single phase generator. Given that each phase is identical (ie;
resistance, V/rpm etc) they can offer either a low cut-in or start-up in a star
configuration though at a lesser current, or higher current and ultimately
greater power at higher RPM's if wired as Delta.
Each
leg is out of phase relative to it's neighbors by 120 degrees of their
respective sine waves, this has a smoothing effect on the mechanical loading of
the generator as well, compared to a single phase machine that has a far more
pronounced cogging effect.
For
more detailed info on 3-Phase http://www.windstuffnow.com/main/3_phase_basics.htm
, of the various sites this seemed to lay it out most clearly and included lots of
nice formulas to play with.
With
the stator wired, I cut out a section of 1/4" plywood to form the outer
shape of the stator...
Again,
the cores are 1/4" thick plus 0.100" for the coil backings, though the
backings are recessed into the stator jig.
The
actual coils of wire inevitably loosen and bulge even with a minimum of handling
and great care...
So
once the mold is greased down and poured, a piece of 1" MDF was clamped
over the stator to ensure that the face was completely flat.
The
resin was tinted with about an ounce of green paint, although it may not look
very sturdy, it is quite rigid. The latex paint does add a very slight
plasticity to the resin which may make it more tolerant of sharp blows, as the
pure resin is prone to shatter with little give.
The
stator is positioned by 3 lengths of 3/8" all-thread with jacking nuts to
allow easy adjustment.
The
gap between the stator and the rotor is set to 1/8".
Each
phase measures 3.2ohms DC resistance... at a moderate cadence of 60 pedal RPM
(just under 400RPM at the rotor) a single phase has an open circuit Voltage of
14Vac.
Given
the data from the windstuffnow.com site,
14Vac
=14 X 1.414 =19.8Vdc (1.414 is the constant that converts Vac RMS to Vdc)
19.8Vdc
- 1.2Vdc = 18.6Vdc (the 1.2Vdc is the allowance for the voltage drop across the
diodes that convert the AC to DC voltage).
In a
star configuration 18.6Vdc X 1.723 = 32Vdc (1.723 is the cube root of 3, forming
the constant or multiplier of a single phase voltage).
This
is where things need some clarification...
If
the Star 3-phase output is loaded by a 12V load (a single battery) then;
32Vdc
- 12Vdc = 30Vdc / 3.2 ohm phase resistance = 9.4Amps
So
12Vdc X 9.4 Amps = 112Watts power output...
But
the reality is that with a star configuration, I read 4 Amps into the 12V load
or just under 50Watts. The missing factor is the cumulative resistance of the 3
phases... As well I have a sense that the stator is essentially a massive
serially wound inductor. And as inductors behave differently under different
frequencies, is the impedance of the stator varying with the speed of the rotor?
Before
wrapping-up this installment I hooked-up the classic 55Watt Halogen automotive
headlight.
As
much as I wanted to pedal up to about 100RPM at the pedals to pop the bulb I was
able to resist the temptation (for now) as I wanted to apply a known load
to see if the math or my readings were closer to the truth...
It
would appear that the math is flawed and 50 Watts is pretty much right.
The
next section will concentrate on building a PicAxe based data-logger that will
interface to a PC and graph the output of the generator in real-time... and
hopefully I'll have resolved some of the total stator resistance issues...