Light Electric Vehicle #1...
May 2007 Update (Part #5)
1 step
forward 2 steps back...
This page is brought to you by The House of
Morgenstern - I.T Services, Rick B. of Peterborough and TNC Scooters of
Tennessee.
As this installment is crazy long and photo-rich,
I thought it best to include a shot of the finished product for any attention
deficit readers that might frequent theworkshop.ca.
To
the left is a good portion of a $500 capital expenditure that was made possible
by the fine folks at The House of Morgenstern - I.T. Services...
Custom
drive electronics, some cosmetics accessories (lights, turn signals and brake
lights) all from www.tncscooters.com,
and heavy duty awg #8 wire and connectors from NAPA of Renfrew.
The
basic wiring of the head-light & turn signals is fairly straight forward and
doesn't warrant much space on this page, all items are rated and wired at 24Vdc.
I did
want to note though that the turn signal relay is not a relay, but rather is a
solid state oscillator of sorts that has a rather annoying piezo speaker that I
quickly filled with hot glue to dampen it's output.
I
honestly thought that this would be a quick upgrade, perhaps just under a day of
shop time and 5 or 6 photos... And it all started out that way.
This
is a 48V 50Amp controller from Yi-Yun Model LB-37.
It's
entire housing is aluminum and provides that heat sink function that such a
small device would require.
As I
had concerns about frying the controller, I opted to fabricate a mounting plate
of 1/4" Alum and a second plate that mates to the controller's underside.
Each
contacting surface is sanded with 800 grit sand paper to achieve a smooth
surface for optimal thermal transfer.
Ideally
if I had some heat-sink compound on hand I would have smeared a layer between
the surfaces to further enhance the heat dissipation properties of the now monstrous
heat sink.
The
power electronics look quite smart when mounted, and free up some valuable real
estate for the addition of a 4th battery to bring the system up to the rated
48V.
The
silver switch is wired to the key lock, and the black plug-in connector is for
the 48V battery charger.
I do
have to note that the 48V charger line uses a plug configuration that is also
found on computer power supplies and monitors at 110Vac. This poses the
potential of a well intentioned but un-informed user plugging in the LEV-1
directly to 110Vac and destroying the controller and running the risk of both
Fire AND an Explosion if the AC current were to get past the controller. Oh My
Gawd!!! Don't plug the LEV-1 Directly into an AC power source. (after just
writing this, I think that I'll drop that connector and wire the physical
controller to the trike later this week)
Above
is a shot of the original wiring (to the left) that was a mish-mash of AWG#12,
#10 and scraps of wire that I found along the side of the road. Some ends are
crimped, some are crimped and soldered... To the right is a much cleaner
configuration of Prime AWG#8 wire all crimped and soldered to ensure that every
watt of the precious energy makes it's way to the rear wheels.
As
AWG #8 wire and the appropriate crimp connectors are quite large, I opted to use
a small butane torch to ensure that the solder would flow properly and create a
good electrical connection.
It's
minor details like these that start to draw out a 1/2 day project into the next
day...
One
would think that I'm trying to show-off the snappy Sigma digital speedometer,
but in actual fact the purpose of this image is to high-light the Brake lever
that has a switch that is wired into the controller.
As
well as the $3 or $4 headlight/turn signal control that is integrated into the
hand-grip.
I
can't help but feel a measure of unease purchasing (and essentially advertising)
such cheap off-shore components, but I know that I simply could not afford to
purchase equivalent hardware (if I could find it) that was made any where in
North America.
As an
aside, the Much Hyped AWG#8 wire was made in the USA and was the highest capital
cost item @ $74/100ft roll and it didn't have to travel several thousand
kilometers to get to me.
With
everything wired and tested without incident, collectively (the kids, my wife,
and I) logged about 20Km on the re-vamped LEV-1... It was wild, 27Km/h on flat
ground, blazing acceleration (to the point that the rear tires actually would
spin if the throttle was cracked open on wet grass or loose gravel).
By
all accounts it was a winner! The Controller exhibited zero rise in temperature
above ambient, though I did check the motor periodically and it was getting up
to 60C according to the touchless thermometer I keep handy for such tasks.
This
went on for a couple of days, but our bliss filled existence with the reworked
LEV-1 achieved it's zenith when my wife opted to go into town (5Km in the valley
below) to pick-up one of the kids. As an aside to this current aside, in the
mind of a 16 year-old, as cool as a home-built electric vehicle is to run around
the farm, it is a transport medium that apparently carries a social stigmata
akin to leaving a public bathroom with a 3ft streamer of soiled toilet paper
dangling from the back of ones pants.
But
my wife in her free-spirited embrace of life simply kept driving circles around
the now outraged (and enraged) spawn and taunted her that either she could hop
on the back for the ride home or be buzzed at like the swarm of black-flies that
make May in rural Canada the joy that it is.
Realizing
that she could be whisked out of this psychological trauma @ 27Km/h or endure it
6.5Km/h (walking speed) she opted to hop on the back... They made it home
without any further incident that I heard about, though I wonder about the dead
groundhog that we found in our bed later that night...
Once
they returned I took the LEV-1 out and found that it was running at approx half
speed, so I returned to the shop... I didn't need the thermometer to realize
that the motor had over heated as I burnt my hand when I touched it to see if it
was hot.
Here
is the motor as I start the process of removing it to see whutz-up...
This
shot is off the beefy internal gear reduction unit that is integrated into the
motor.
It
had to be removed to allow the motor to the split for inspection.
And
here it is...
Above
is a shot of the commutator (copper part that delivers electricity to the coils
of the armature). It certainly looks to be in rough shape, to clean it up, it
was set in the lathe and sanded down to the finish shown to the right. I had
thought about actually turning it down with a cutter but had concerns that I may
damage it rather than improve it.
Also
it should be noted that the copper dust that is evident between the individual
segments was washed out and blown clean with 100psi of compressed air.
The
cause of the power drop and over-heating was that a complete set of brushes has
let loose inside the motor, above are the before & after images of that
event. Unfortunately I had to use generic lead/tin solder to perform the repair,
at my earliest convenience I will have to procure a qty of solder that has a 2
to 3% silver content as it would offer a greater measure of thermal robustness.
Relieving
the symptom does not address the cause, so I opted to embark on a massive
"active-cooling" campaign.
A
series of 1/4" holes are swiss-cheesed into strategic locations. Two
adjacent to each brush, with an equal number of access holes on the end of the
motor (Brush-side).
After
all the drilling is done on all parts they are brushed, banged and blown clean
of any shavings or grit. Every thing was then washed with WD-40 and cleaned
again prior to re-assembly.
I
can't stress enough the potential destruction that would result from leaving
loose metallic shavings in a contained space that is transferring two thousand
four hundred watts of energy into rotary motion.
This
is a view of the completely ported motor, notice that the active cooling exhaust
ports extend between the stationery magnet poles housed in the center of the
motor.
The
idea I have is that the 1/4" holes add to the surface to volume ratio of
the components that enjoy their presence.
That's
cool stuff!!!
By
adding a positive air pressure internally to the motor, two additional issues
are addressed, a) the motor will not become contaminated by environmental foreign
matter and b) the exhausting air will aid in removing more excessive heat than
would typically be dissipated otherwise.
...shown
with the fan that will provide the required air-flow to cool the motor.
The
fan is of superior quality to what is typically recovered from a used computer
power-supply. This particular unit is one of a 3 X 3 panel of 9 fans that were
recovered from a Sun Solaris server that I gutted at the turn of the millennium.
Although
the fan is rated for 12Vdc it has been tested at 24V to increase the air-flow
and simplify wiring by tapping into the existing secondary 24V circuit that was
established to run the lighting accessories.
The
fan just screams but has run for a 15 minute span with no issues, so it is
mounted.
A
simple air filter is fabricated by sandwiching a square of filtering felt
between two plates of steel grill-work.
The
felt was liberated from the numerous filter pads that we use for maple syrup
production.
The
felt is washable and has a tight durable weave yet is still breathable
enough that I can actually feel the air flow out of the motor when engaged.
So
Life goes on, the LEV-1 is hauling jerry cans of gas to the tractor or truck
around the farm, and I'm just waillin' between the machine-shop and the foundry
or any where else that I used to walk like a sap...
One
cool and drizzly evening after dinner I thought I'd head back into the bush to
look for poplar saplings or trees that have gone into seed... (the poplar
seedling project is for a 5 acre re-forestation initiative that I hope can
provide 2 to 3 years of firewood for either myself or the next stewards of this
farm.)
Along
side of these circumstances one of my fishing and hunting chums has procured a
shiny new Suzuki 450cc 4wd ATV... As I had a few chances to run the trails with
the 450, my aggressive driving habits were transferred to the LEV-1 in kind.
As is
typically the case with any sort of catastrophic component failure, the LEV-1
blew-out the gear reduction unit internally as far back in the bush as one can
get on our farm with out actually crossing a fence into neighboring property.
This
is a nasty business, and I feel slightly nauseous just glancing at the image to
the right.
It
would seem that the weak point was the helical gear that couples to the output
shaft of the motor itself.
The
spiral drive on the shaft appears to be completely intact.
To
the left is a complete replacement motor, drive train, differential and axels.
These
are courtesy of Rick B. in Peterborough.
We
did a swap for a set of custom hubs and the 13.5" low rolling resistance
tires (on-rims) as posted earlier in this series.
Obviously
this is one of Rick's numerous spares, as he has greater access to this sort of
product than I.
Over
the last year I've found that there is a significant Fortress Scientific FS2000
community in the world that are keeping this workhorse alive. At least that's
the impression that I get from the number of emails I get about sourcing parts
(I don't know) and how to fix (it would seem that so far I've established a few
proven methods on how to break the FS2000).
Since
the motor has a solid day of work already invested in it, I opted to just swap
the gear reduction unit.
This
unit is in pristine condition and likely still has the original factory grease
applied.
As
with the burnt brush wiring, to simply replace the reduction unit without
resolving the cause of the problem would result in a similar outcome.
After
a few hours of careful reflection I felt that the drive train was obviously
stronger then the failure point and was being loaded by jack-rabbit starts and
sharp load variations when the LEV-1's rear wheels would leave the ground
momentarily at full speed and re-establish contact as when encountering
wash-boards ruts or generally rough terrain.
By
reducing the final drive gearing between the gear-reduced motor and the tires at
2 (two) points I could reduce the total drive speed (essentially slowing the
entire vehicle down) and impose more sensible driving habits upon myself ie;
smoother more controlled acceleration from a standing start.
As I
mentioned earlier my previous wheels were swapped for the spare motor, so I'd
been running an old set of 16" tires that have at least 4 to 5 times the
contact area (as well as their larger diameter) as a load factor on the drive
train.
Almost
$100 later I've reduced the load factor on the drive train by 9.375% (16"
divided by 14.5" (these are slightly larger than the originals))...
...
the top speed should have dropped from 27Km/h down to 24.5Km/h, this loss
translates into additional hill climbing and or load carrying capabilities.
Since
the motor was out and I had the properly sized tires on the unit I figured that
I may as well remove the batteries and deal with a few structural problems that
I wanted to resolve.
Firstly
the 26" wheel had such a sever rake relative to the frame that steering was
poor at these higher speeds and the head-tube fit that I had to make was
becoming a problem.
So I
re-installed the original 20" tire and applied the proper rake, and boom
placement to establish adequate ground clearance and better steering and braking
functionality.
As
the battery area, seat placement and the cargo box were starting to infringe
upon each other's territory I figured I may as well remove and rebuild the box
assembly to accommodate a solution.
The
framing shown to the right is all 1/8th" walled tube construction that were
delivered as surplus by Rob & Bob in town (this is pure bonus as the price
of steel is climbing and would have been another hundred bucks or more...)
Gentlemen, I salute you!
The
side supports are more bed framing that for now is still free from dumpsters if
you are so inclined.
Did I
not fuckin' say that I thought that this would be a single day project a week
and a half ago at this point in time!!! I get freaked out be these sorts of
extended play, remixed, drawn-out drains on my time as I have real work that is
supposed to be getting done, beyond letting this damned LEV-1 suck the very life
out of my body.
After
a couple of false starts this is what I ended-up with and it is entirely
consistent with the idea that I can increase the load on the LEV-1,
unfortunately I'm sure that with 200 to 300lbs more load plus the actual dumping
box hardware just added The helical gear that lasted just a few days will now
strip clean on the first run.
But
it looks cool...
Since
the entire day has been shot any way I worked into the dusk and spray-down the
frame modifications and the box assembly with a rich coat of Lustre Noir ...
With
that done I had no choice but to either remove the 48V 50Amp controller and
return to the original 36V/40Amp controller or find another point of mechanical
advantage to levy on the final drive train.
The
only post gear-reduction point of modification that was obvious to me was the
gearing between the motor and the differential assembly. So I calculated that a
1/3 (or 33%) reduction in the ratio between these chain driven components would
be the easiest option as they were already aligned and in a 1:1 ratio of 27
teeth each...
If I
could fabricate an 18 tooth drive gear on the motor output shaft to drive the
unaltered differential gear I'd have my 1/3rd speed reduction and proportionate
load carrying gain (with hopefully a margin of safety to spare the helical gear
that started this whole mess earlier on)
The
image above is the template that I thought about for the better half of a day
and took an entire day to make on the computer.
Don't
be fooled by the simplicity of the image above the criteria of accuracy to
0.001" and angular precision to 2 decimal points was critical as I had to
resort to graphical successive approximation to find the tooth and hob placement
of the gear geometry...
The
template was printed and glued to the gear face, and punched as accurately as
possible.
The
chain hobs (or area to drive the links) is drilled out to a diameter that is
1/64th" larger than the physical chain link cylinders...
The
1/8th" outer holes establish the top of each tooth of the gear and set the
angle that they will be cut at.
The
Lathe was used to set the concentric outer reference for what will ultimately be
the finished gear.
The
gear took about 2 hours to shape and test and reshape the teeth against the
physical chain and the spare gear that was a pre-requisite to me even
considering this course of action.
At
the end of the day I did end-up with a very passable gear and would not hesitate
to repeat the above in future projects now that I have a better idea of what is
required.
And
here we are revisiting the more mundane tasks that typically fill my day...
With
a 1/3rd drive reduction the now (Calculated) top speed of 24.5Km/h should drop
to 16.3Km/h...
In
actual fact it has dropped to 13Km/h top speed on the flats... So between the
tire size and gearing I've lost slightly more than 1/2 the top speed and doubled
the entire load capacity of the trike.
The
big variable is the load capacity and user aggressiveness on the twist throttle.
My
primary aversion to down grading the controller to 36V/40Amps was the low speed
power that is head and shoulders above at 48V... This is the first 1/4 to 1/2 of
the throttle movement on the handle bar...
The
function of the LEV-1 has to be more clearly defined as a materials hauler over
steep un-even terrain, with at least a 200lb load capacity for soil, rocks,
fuel, tools, scrap etc... Speed realistically is not a concern as I know that I
couldn't carry 100 to 150lbs of anything up hill at even 10km/h on my back for
the distances that the LEV-1 is routinely doing...
In
conclusion, I will count on exercising better judgment in how I drive the LEV-1
but before long will start to look at how to modify the controller to limit it's
output to initially 40Amps and possibly as low as 30Amps to offer the piece of
mind that anybody can run the LEV-1 without a Miranda-Rights reading on what to
do and not do to keep from destroying the drive train.
And
lastly, the 24V motor blower eventually died in a puff of smoke after 4 or 5
hours of run-time since it was installed (this unfortunately has to be dealt
with immediately as it had kept the motor running in the 40C range even after
being heavily loaded for extended periods of time.)