Transcript RuggedPOD Proto 1 bugs

RuggedPOD Proto 1 bugs
SDS team
Wolfe Engineering
•
Hypothesis
–
–
•
–
•
Liquid Leak
Liquid leak is coming from poor welding between panels
We did not experienced liquid leak on our early prototypes welded in France
Panel material: AS7G03
– European aluminum alloy for die casting. Is it available in the US ?
Welded successfully in France with the following procedure
– Pre-heating to 200-300 C with a welding torch to remove oxidation, improve welding grip and
reduce gas emission during welding process
– Use TIG process with Argon and AG3 filler Rod
– Is this ROD available in the US ?
– Keep the material pre-heated during the process
Liquid leak due to premature failure
•
The original design was using locking mechanism between panel (see the
following top view)
•
This locking system was on top part of the panel and has been removed by
the machining process reducing potential material to material contact and
pushing the pressure support to the weld.
•
•
Extreme forces on cube edge
without locking mechanism
which could probably led to
weld failure if they are not
strongly attached
French proto style have in die
casting rail support increasing
thickness by 5 mm on some
area making the panel much
stronger
Gas Leak
• The top plate was misaligned on the prototype,
not sharing the same plane than its support
containing the Oring.
• We do not really know from where this
misalignment is coming from, but it could be
perhaps from the welding process of the support
to the tank, which has involved some twisting of
the support. Can we check ?
• One work around could be to use a thicker oring.
3 Pails doesn’t fit
• We need 3 Pails of 5 Gallons oil
• Oil is available as Pail or drum. Drum are much
less expensive but contain 45 Gallons. Can we
have a single drum of 15 Gallons compliant
with oil ?
Blade bending
• Move to stainless steel material
Screw doesn’t fit SSD mounting hole
Wire length (bug 6/7/8/9/10)
• Before
• After
• Change wire routing, and add some wire strapping.
(position will be transferred on sept 22th)
Wire length (bug 6/7/8/9/10)
• Before
• After
• Change wire routing, and add some wire strapping.
(position will be transferred on sept 22th)
Wire length
•
•
•
•
•
•
5v/12V to DCDC cable lenght : 65mm
DCDC power cable lenght : 75mm
AC/DC to backplane cable lenght : 85mm
Backplane to ethernet switch : 4x60mm
Raspberry to ethernet switch : 100mm
Switch power cable : 35mm
• All dimensions doen't inculde connectors head
length
Soldered connection on Fisher are
exposed
• We need to find a kind of threaded cork to
hide the wires.
• Fisher do not seem to provide it and the
threaded is a custom one.
• We didn't find something that fit at this time
but can continue the dev program without it
through thermal shrink protection
• Maybe the solution will be to make a custom
one
Top cover cables are not guided
Raspberry PI mounting holes
Standoff are now M2.5
Backplane PCB mounting screws
• Will be changed for next pass
USB power cable for RPI not pluggable
after backplane positionned
We can use this type of USB
cable
http://goo.gl/9n5E2h
Cables on Blade are loose
Need a slot for an optional Vacuum
Pump
•
•
Robinair Vacuum pump 1.5CFM/min
Warning: Missing cable between vacuum inlet to POD. We need the vacuum cable
•
•
•
Intake fitting: ¼ inch flare and ½ inch Acme intake port
Voltage: 110V
More info : http://www.robinair.com/products/vacumaster%C2%AE-single-stage-pump
•
Can Jabil supply us this pump ?
Need a slot for an optional oil transfer
pump
• http://www.ebay.com/itm/110v-Electric-Oil-FuelBio-Diesel-Gas-Transfer-Pump-W-Meter-HoseNozzle-/191563985417
• Can Jabil supply us this pump ?
Impossible to fully empty a POD
When serial is connected to a motherboard we
observed some random POST failure on MSI core i7
board
• MSI board wiring
COM 1
• Connectic board wiring
J52
RX
TX
TX
RX
Connect RX (com1) on RX(J52) and TX (com1) on TX (J52)
This was a bad wiring approach
POD is hard to move (handle
positioning)
Label on electronics part need to be
withdrawn
• Will be done for next batch
Top cover not aligned with support
• We believe that this is a duplicate of Gas leak
bug and that we can close this one. The main
target is to avoid any gas leak
Mechanical conflict between raspberry
pi, wires and power supply
• We will use right angle connectors to avoid
collision,
• the raspberry positioning has been changed,
Define a sealing test procedure
Explain in documentation how to
position O-ring
Lots of screws, deliver pods with
electrical screwdriver
Mechanical conflict between ethernet
wires and cover
We can use this type of RJ45
cable
http://www.touslescables.com/b.php?a=Z1EL&c=Voi&h=162
1
Mechanical conflict between power
supply and rails
• The rail must be shortned
by 15 mm
Add a cable guide for every internal
cable even on blades
Add a GFCI to every ruggedpod
• We need to find one for north america market
and one for european market
Blade collision with PSU
Add a network switch support