/td> Yahoo! is not affiliated with the authors of this page or responsible for its content.

Kinkajou Prototype Development Report 30 March
4 April 2004

LED

The Luxeon Star LXHL-LW6C now costs $24, and is back ordered until
20 April.   Once the team determines the number of spare Kinkajous
to make and the number of spare LEDs to have on hand, we can place the
order.  The forward voltage of this chip varies, in two samples
measured, from 6.1V to 7.0V. 

Driver circuits

Tim McNerny has requested 4 square inches of circuit
board space for the LED driver.  Although the driver wont require
access to the heat sink for cooling, there will be a thermistor to sense
LED base temperature by contacting the heatsink near the LED. 
I will attempt to locate the board adjacent to the heatsink, probably
as shown in the package illustration.  The driver will provide
700mA so long as the base temperature remains below some safe value,
and will reduce the current thereafter, attempting to maintain the safe
temperature.  It will modulate the LED as a signal of low battery
voltage, and power down at some lower threshold.  It will provide
reverse-polarity protection and provide diagnostic information (operating
time and number of operations).

Temperature control

I tested a new heatsink design, better adapted to
manufacture by

extrusion.  Instead of pin fins, this has conventional
extruded fins 2mm thick and 25mm long, on 8mm pitch.  Two lengths
were tested, both 43mm high.  The longer one, having 23 fins, stabilized
at 30C with no fan or chimney (21C ambient).  The thermal resistance
is therefore 1.8 C/W.  The shorter (6 fins) stabilized at 41C,
giving 4.1 C/W.  The required heatsink performance can be calculated
by adding the chip-to-base resistance (11C/W), multiplying by the 5
W power, and adding to ambient to find the expected chip temperature,
and comparing this to an acceptable value of 135C.  Even the 6
fin heatsink will produce a chip temperature of 125C, but light output
is greater at lower temperatures, so I will try to get ~20 fins into
the design.

Illumination intensity

The lab light meter (an EXTECH 401025) was received and put in service. 
It allows 1-Lux resolution for comparing light levels from different
sources, optics, and spacing of optical components.  The previous
comparison between the Kinkajou and the Fisher Price Viewmaster projector
was repeated, with the following results:

With both projectors illuminating a 20cm-wide area,

F-P Viewmaster: 435 lux at center

Kinkajou:             
369 lux at center

Although both projectors had light values fall off
at the edges, the effect was much less noticeable with the Kinkajou,
and the F-P projector had severe color banding.

Condenser lenses

Since the condenser lens in the toy projector (a pair
of plastic plano-convex lenses (f~19mm) is so much less expensive than
that in the Kinkajou, I tested the plastic lens with the LED. 
Although the LED needed to be ~17mm from the lens to match the light
to the projector lens aperture, the new light meter showed that the
maximum light was transmitted when the distance was only 10mm, even
though this caused light to fall outside the projector lens.  This
can be understood as the lens intercepting a larger solid angle from
the LED, even though some of that light is wasted because the focal
length is too long.

Similar modification of optical spacing increased
the illumination from the Kinkajou optics to increase to 522 lux. 
Substituting the F-P lenses, this value was only 361 lux.  These
results imply that the F-P lens can work almost as well as the Kinkajou
is actually working, but the optical cell must be longer by 15mm.

LED life test.

The life test begun last week accumulated 100 hours of constant illumination. 
At that time, an intermittent operation switch was added to the setup,
giving one interruption every 2 seconds for a 90% duty cycle. 
As a result, the base temperature dropped to 31C (at 700mA, 6.1V) 
An additional 70h have been accumulated.

New Package Design

Ive
begun constructing a solid model of some of the packaging concepts I
want to explore.  These are shown in the illustration:

Unitized hea</span><span class="Normal--Char" style=" font-family: 'Georgia', 'Arial';
">tsink and housing.  The two ends of the housing are identical
slices of an aluminum extrusion, with fins.  They are joined in the center by a bolted bridge
strip, and across the front by another bolted extrusion.  The large
extrusions are a solid extruded shape (which means they are neither hollow nor semi-hollow)  fitting into a 7 diameter circle.  The fins will be designed for 2C/W.
Asymmetrical optics.  Use of identical extrusions results in
a joint on the centerline.  The optics can be on one side, the
circuit board on the other.
Sliding door to protect the projector lens.  The front panel
extrusion has channels to capture a sliding door.
Cylindrical condenser lens housing can be machined in a bar-fed lathe.
Focus mechanism.  Not shown yet on the model, I</span><span class="Normal--Char" style=" font-family: 'Georgia', 'Arial';
"></span><span class="Normal--Char" style=" font-family: 'Georgia', 'Arial';
">m considering moving the film gate rather than the projector lens
for focus adjustment.  This will simplify the lens mounting and seal
dust out.  The projection lens will be fixed in a hole in the front
plate, by a spring ring or adhesive.  The film gate will be two
polished metal plates, one of which attaches to the front plate by a
long flexural part.  A thumbscrew through the plat</span><span class="Normal--Char" style=" font-family: 'Georgia', 'Arial';
">e adjusts the gate</span><span class="Normal--Char" style=" font-family: 'Georgia', 'Arial';
"></span><span class="Normal--Char" style=" font-family: 'Georgia', 'Arial';
">s position.

Allen Armstrong