|
Ray et.al. Thanks, glad I read more messages, excellent
summary, let’s vote. Let me play out 1 further review: Board
size. Project board size was originally mandated
at 100 x 100 mm but was based on our discussions of a ‘standard’
enclosure which had 100 (+) mm card slide groves and a manufactured depth to
the backplane from the front of the enclosure of 100 mm when card was seated. Pretty
much a VME standard with a fixed depth so all cards would project out of a
‘front panel’ with connectors on it. Efficient but customized. The ATX power supply and computer
enclosure packaging IHMO was a GREAT idea. We were getting into ‘custom
everything’, and since we ARE hobbyists, the more we customize the more
burden for the group to produce ‘low quantity higher priced customized
circuits’. I’t would be nice NOT to have to be discussing an
Atlas board in lieu of an inexpensive PC backplane, for that reason, but the
decision has been made to produce our backplane. If we stick with the “computer
enclosure” scenario then the cards can be the height of a standard
L-Bracket (? mm) and ANY depth. The distance between the connector on the Atlas
board and the L-bracket in a computer case is the only strict dimension. I’ll make this statement one more
time, as someone who is willing to produce and distribute ‘kits?’
for these projects, and having done 3 SDR -1000 “Tony – Kits” We are a SMALL but obviously very creative
group. We are self funding. Generic design with ‘off the shelf’
parts, without sacrificing our goals is ESSENTIAL. Universal appeal of any
project we produce ESPECIALLY the FPGA board and the Atlas BUSS makes the difference
between success or failure of the project. Losing to a single board custom
design for all projects, each with it’s own FPGA. We have the choice
here. Comments embedded From:
xylo-sdr-bounces@lists.ae5k.us [mailto:xylo-sdr-bounces@lists.ae5k.us] On Behalf Of Ray Anderson Based on the interchanges over the past couple days it seems that a
several basic ATLAS candidates have arisen: 2 layer No LVDS on PCB Arguments for: less expensive to fab
LVDS can be distributed with twisted pair cables
Fewer or no termination issues 4-layer 4 LVDS lanes on board Arguments for: LVDS lanes make it more general purpose
for future applications [Eric:] Difference
between ‘spurt’ selling 40 boards at $100 and 1000 boards at $10.
(I just got a request for 1 (One) Delta – 44 interface card! (smile)) We
should provide a couple and maybe 4 LVDS compatible channels. 4 slot Arguments for: Fits in 4 slot Aria case [Eric]
Discussion – It Does not have to be Aria, but the Aria IS a pretty nice
enclosure for projects as Phil_C and several of us have discussed. The choice
really is 4 or 5 slots if we use existing computer enclosures as a house for
anyone’s projects, and hobbyists. You guys don’t know how “Totally
I am INTO” Phil_C’s Mercury, depending on Phil_H’s bandpass
board, and Bill and Phil’s Janus board and Leon’s Lionheart , BUT that
is a Prodcuct in itself worthy of an extension to the future backplane, and can
be accommodated now for experimentation (including impedance bumps depending
where you plug the cards in) with 2 Atlas boards. Stick with 4, if the Atlas
works well then stretch the board in a future production to six slots and
customize the case. Give the whole shooting match to Gerald to produce. (Put a
big tuning knob on it for good measure!) This is really a prototyping project,
just a little more refined. I vote for 4 slots, daisy chain as needed. 6 slot Arguments for: We’ll probably need more than 4
slots eventually
More general purpose appeal for other applications
Don’t need buss extender boards There are probably other arguments for each case which I haven’t
captured. So it seems we need to come to concensus on which fork to take. From my perspective I’m thinking the 2 layer 6 slot variant makes
the most sense to me. Let the comments/suggestions/arguments flow….. -Ray WB6TPU |