> On Tue, Dec 02, 2003 at 02:10:36AM -0800, vrs wrote:
> > At 11*2=22 boards/device, you are even
> > talking quantities that make sense to send out to a board fabricator.
>
> Indeed. If you were really being cheap, single-sided boards would cost
> less, but I don't have any numbers to estimate how much cheaper they
> would be. I have both coax and ribbon-cable negibus cables, and some
> of the paddle cards are about the size of a G727 Unibus grant card.
> Those would be somewhat cheap to make, especially at q25.
It looks about a buck cheaper for small quantities of single sided.
Checking online, it looks like $13.21 vs $14.46 for 1 week delivery. 5
weeks is $9.40 each for the two-sided boards, all qty 30.
30 bare-bones one-sided cards can be had in 5 weeks for $7.81 each, no
solder mask or silk-screen.
30 bare-bones two-sided cards can be had in 5 weeks for $8.62 each, no
solder mask or silk-screen.
That's just one vendor, of course, but it is the one with the steepest
quantity discount I know. Those boards are around $4 each qty 100, and
$2.50 each, qty 500 :-). As the quantity goes up, the difference in price
also goes down. If you are using 20-odd per device, it might be possible to
use 100 :-).
Of course, single sided cards are not as generic, so you might as well make
them W011 clones and use narrower headers and ribbon, for additional cost
savings :-).
> > BTW: Any market for DF32 emulators out there? I like mine, but I
wondered
> > if there was demand to get more made?
>
> I'd love one, but I might have to convert my negibus -8/i to posibus to
> use it. I suppose I could throw one on my 12K 8/L. I've always wanted
> to have disk on it. How much harder would it be to make an RF08
> emulator? :-)
I don't think it would be too hard, but then I didn't do the design work on
the DF32 emulation :-). Basically, it is like some of the talk on another
thread -- an NVRAM chip, some counters and gates for data break, and a bus
interface card that collects up all the Posibus signals. It emulates all 8
possible drives (and exports write-protect switches for each one :-)). Last
I heard, there was a diagnostic function unimplemented that keeps it from
being 100% software compatible. I should recheck the RevB schematic to see
if that has been addressed yet. Replaces a whole rack with a 1U device :-).
I just have to get my 8L working well enough to use it :-).
It also occurred to me that one could turn the generic boards into M904/W011
clones by using the pads for the second header to install the ground jumpers
needed. That would give a very clean look, but would also use more paddle
cards than ribbon cable daisy chaining.
Vince
Received on Tue Dec 02 2003 - 09:27:12 GMT
This archive was generated by hypermail 2.3.0
: Fri Oct 10 2014 - 23:35:49 BST