John Lark [snip]
>> Circular boards are cute, but are almost always a serious PITA.
>I agree that circular boards are a pain, mostly because you lose a lot
>of usable area along the curve. However they're sort of a fact of life
>for this client, because of needing to mount the optomechanics to the
>cards, and they've got good at it.
>Long skinny boards don't have a lot of surface area either. I stack
>them, they have to get skinny really fast. Besides, I can easily put
>the pringles in a steel can, sitting inside the aluminum housing. The
>can will be slightly D shaped, with the wiring coming out on the flat side.
>They connect with one flex jumper each, and there are no connections
>between them. That makes working on them pretty simple, actually--the
>whole stack just falls apart into separate pieces that you can probe any
>way you like. (It's the stacked headers that make daughterboards such
>a pain to work on.) Because they're quiet, it doesn't matter if the
>input and output leads are close together.
>Also the pringle design is more extensible--I only have so much
>diameter, but in principle I can have a mile's worth of potato chips if
>I like. ;) Seriously, doing it that way I can add more boards so I can
>use many many heating zones if I need to, to keep the thermal forcing
>down to a very low level.
I can recall making hybrid circuits for a Systron-Donner accelerometer, which was composed of a large number of circular boards stacked via interconnecting bus bars and spacers. I thought it rather cute. It would be easy enough to insert a shield between "layers". ...Jim Thompson
| James E.Thompson, CTO | mens |
| Analog Innovations, Inc. | et |
Click to see the full signature