Loading ...
Sorry, an error occurred while loading the content.

6561Re: Membership Card - Ultra Low Power Variant

Expand Messages
  • aa3nm
    Mar 1 7:58 AM
    • 0 Attachment
      Lee,

      Thanks for clarifying the debounce. Didn't realize I was using my bad eye when I squinted at that part of the schematic. Totally not sure how I didn't see the feedback resistor as debounce.

      All good points on; the DIP Switches, layout, and on the memory power. I was thinking about all those points too. In the end, it's all a compromise anyway.

      My goal is to have a truly minimum power ELF that is as simple as possible with binary input and output. That being said, the tedium of programming anything more than ~50 bytes gets old very fast, even with toggle switches (been there too).

      Filling out the extra board space with a higher power and function expansion capability didn't seem to degrade my goal (as long as it could be disconnected). It would allow more flexibility of use. This is especially true if it enables easier loading of the 1/4 K low power RAM.

      The key point is to have the expansion board be unplugged from the micro-power ELF without loosing the 1/4 K RAM contents.

      Thanks for the inputs.

      BTW, I know you did a lot of work on the power consumption of the membership card - Do you have any sense on the consumption of this layout? I read on Herb's page that you were "running air" on the membershiip card overnight from the charge in the .047 uF cap, but adding the LCD display bumps up the power consumption (a lot less than LEDs but its still there).

      Steve


      --- In cosmacelf@yahoogroups.com, Lee Hart <leeahart@...> wrote:
      >
      > aa3nm wrote:
      > > I've taken Lee's schematic (available in the files section under
      > > Membershipcard LCD.pdf) and re-done it in the ExpressPCB schematic
      > > tool... to add the "8 bit binary LCD display" I described earlier.
      >
      > OK! Glad it was of use. I never got around to building this one, as so
      > few people seemed to be interested in it.
      >
      > > While at it, I decided to use a pair of DIP 4 wide SPDT switched for input.
      >
      > They are certainly small, but I think you'd get sick of operating them
      > pretty quickly.
      >
      > > In the process I discovered an issue with Lee's schematic – the input
      > > switch buffer for bit 0 should be U9D not 9A a second time. This
      > > puts the input on pin 10 and output as pin 9.
      >
      > Good catch. I fixed it on my copy.
      >
      > > Also, I notice there is no debounce on the Input switch SW1. I
      > > wonder about this and I'll hope Lee comments on this question.
      >
      > The switches *are* debounced. See the feedback resistors (R6, R14, R15)
      > around the buffers (U9E, U8E, U8F)? The switches (S1, S2, and S4) are
      > SPDT. When you move them from one position to the other, say, from the
      > NO (normally closed) to the NO (normally open) contact, the sequence of
      > events is:
      >
      > NC contact closed
      > NC contact bouncing open/closed
      > no connection from common to either NC or NO
      > NO contact bouncing open/closed
      > NO contact closed
      >
      > The 4503 buffer starts off with its input held low by the NC contact.
      > It's non-inverting, so the output is also low. The input and output
      > voltages are the same, so there is zero current in the resistor.
      >
      > Now you move the switch. When the NC contact starts bouncing
      > open/closed, the resistor holds the input low regardless of the switch
      > bounce. Power supply current is still zero.
      >
      > When neither NC nor NO contacts are closed, the resistor still holds the
      > 4503 output in its last state, still at zero power.
      >
      > When the NO contact first touches, it forces the 4503 input high in
      > spite of the resistor. There is a momentary pulse of current in the
      > resistor, for the time it takes the 4503 output to go high. Then the
      > current is zero again.
      >
      > While the NO contact is bouncing, the resistor holds the 4503 input and
      > output both high, at zero supply current.
      >
      > Thus, this circuit debounces the switches, and takes (essentially) zero
      > power to do it. This particular design was optimized for minimum power
      > consumption.
      >
      > > This space issue presents a bit of a problem for me… If I end up
      > > with a layout that is more than 11 sq-in.(meaning I cannot double the
      > > layout to get a pair of Micro Power ELFs on each board) what do I do
      > > with the extra space? It'd be a shame to have 15 sq in. of ELF on a
      > > 21 sq in. board.
      >
      > Depending on your layout skills, you could use both sides of the PCB for
      > parts, or mount parts underneath the 1802 or LCD display. But consider
      > that it will be harder to build (and to use!) if you make it too small.
      >
      > > So, my thinking is (so as not to waste the extra real-estate) to add
      > > a layout for a memory expansion card. I'm thinking 32k CMOS RAM and
      > > a couple of COMS 2K EEPROMS
      >
      > Unless you are very careful in the choice of memory chips, these newer
      > larger chips will use 100's of times more power than the rest of the
      > circuit. Almost all memory chips are optimized for maximum speed, not
      > minimum power!
      >
      > --
      > Lee A. Hart | Ring the bells that still can ring
      > 814 8th Ave N | Forget the perfect offering
      > Sartell MN 56377 | There is a crack in everything
      > leeahart earthlink.net | That's how the light gets in -- Leonard Cohen
      >
    • Show all 16 messages in this topic