Member Since: August 17, 2007

Country: United States

  • "Isn't immune to small vibrations" is an understatement; its so sensitive it barely works at all. More disturbingly it occasionally fails to close the switch at all in the upright position, seems that the mechanism can fall into a position such that the ball is resting on the pins but doesn't close the contact. I suspect the ball isn't sufficiently weighted and/or smooth to reliably create an electrical circuit with just the force of gravity.
    Might work better as a vibration sensor, or use two in parallel to reduce the probability that they will both fail at the same time?

  • Note that the rebel Blue and Royal Blue LEDs are IEC825 class 2 for eye safety.
    The Royal Blue in particular will appear to be very dim but can still cause eye damage because most of its output power is in the UV spectrum. The ordinary blue (LXML-PB01-xxxx) is a better choice for lighting applications.

  • there are 5 different part numbers with the "cool white" label, which one is this?

  • Nice product, it works fine on my Macbook Pro running inside VMWare with a Windoze-XP image.
    The super-small size is really nice for portability. Much better than a standalone scope. I never used a logic analyzer before and now I think its an essential tool.

  • What you want is a shift register:

  • Windows users: even though its a class-compliant USB serial device I believe you still need to install a "driver" (actually just a .INF file) from Microchip (comes with the USB Framework). On other OSes it should work as-is.

  • Silkscreen has a typo, it should say "LMD1820x" not "LDM1820x"... :)

  • A trademark search would have uncovered this potential conflict, and they are actually quite inexpensive. You should probably fire your lawyer for not suggesting it a long time ago. You do have one right? This is America after all. Good luck...

  • Note that the machine pin size is too small to make reliable connections with breadboards and jumperwire. These need the larger square-pin header.

  • Nice idea, terrible implementation.
    The traces are on both sides of the board and you have to cut the same trace on both sides.
    As pointed out by another poster here, having the traces on both sides gives no advantage since its still effectively a one layer board. Paths cannot cross without shorting, so its mandatory to cut the traces on both sides.
    And, when you flip the board over to cut the opposite trace you are looking at the mirror image of the circuit so it is difficult and error prone to locate the correct trace.
    Also the traces require significant effort to cut with a utility knife. Once you have components on the board it will be even harder since it won't lie flat on a surface. Furthermore if you forget to cut a trace before mounting a component, you will have to desolder it because the traces are covered by the component (such as an IC).
    Finally, implicit connections are less obvious than wired ones, so expect to spend a lot of time double checking your work.
    Expect to spend lots of time with a multimeter testing connectivity at every step.
    Many of these problems would be fixed by putting the traces only on one side. Also I would prefer if there were fewer builtin traces, rather than fully connecting every via, it would be better if most of the connections were only in one direction.

No public wish lists :(