26 Jan
2016
26 Jan
'16
5:46 p.m.
I just received the first of the BC11A paddle boards (so that I can now create my own
BC11A cables). I made two variants of the boards. One where the cables come out the
?top? and another where the cables come out the ?side? (so you don?t have to ?fold? the
cable in a number of different applications such as RK05s).
The ones that arrived today are the S or ?side? variation (as opposed to the T or ?top?
variation). I immediately noticed (why I didn?t see this when I was reviewing the board
layout previously I?ll never know) is that one of the cable connectors is too close to the
edge. It wouldn?t be a problem except that I chose to use connectors that have cable
retention latches. The latches protrude from the board edge by ~0.02?. So once I verify
that these work, I?ll move the one connector inward ~0.5?. I have plenty of space, I just
don?t know why I put the connector that close to the edge. *sigh*. For these boards,
I?ll probably just file the latch down a bit to get the appropriate clearance.
BTW, trying to get the connectors onto the board (I haven?t soldered them in yet) was
tight. That?s the problem with trying to get 60 pins to all line up correctly. But they
fit and are *snug*. I probably won?t go and do any testing until I get the other boards
in so it may be a couple more weeks (the boards weren?t supposed to *ship* before 2/8).
For those who?d like to know, I probably will not be making a production run of these
boards unless I get a *lot* of interest (100 boards or so). If you want some, let me know
and if I get enough interest, I?ll do a production run.
TTFN - Guy
26 Jan
26 Jan
6:03 p.m.
On Tue, Jan 26, 2016 at 5:46 PM, Guy Sotomayor <ggs at shiresoft.com> wrote:
I just received the first of the BC11A paddle boards
(so that I can now create my own BC11A cables).
Cool!
For those who?d like to know, I probably will not be
making a production run of these boards unless I get a *lot* of interest (100 boards or
so). If you want some, let me know and if I get enough interest, I?ll do a production
run.
Considering DEC did something like this for the DWBUA, but with 4
30-pin cables (by necessity because of the BI backplane), is 2x60 pins
more cost-effective than 4x30? Mostly, I'm curious about how much
60-pin ends cost and the difference in cost between 2 lengths of
60-pin ribbon cable vs 4 lengths of 30-pin cable. The copper cost is
the same, but I wasn't sure what the current cost-per-foot is for
ribbon cable. I haven't bought a reel of it in 30 years, but I do
recall that 40 pin seemed to be more expensive than 4 times the length
of 10-pin, for example (since we used both where I worked in the 80s).
Either way, these days, the cost of 2x60 vs 4x30 is probably minor
compared to the fab costs of a "large" PCB with gold fingers. I'd be
curious to know how much a 10' BC11A-work-a-like will run, but I'm
vaguely interested in some pairs of boards. I do have plenty of
Unibus machines and RK05s, most with vintage cables, but I'm sure
several of them have crimps and dings and might fail close inspection.
-ethan
6:17 p.m.
On Jan 26, 2016, at 3:03 PM, Ethan Dicks
<ethan.dicks at gmail.com> wrote:
On Tue, Jan 26, 2016 at 5:46 PM, Guy Sotomayor <ggs at shiresoft.com> wrote:
My choice of 2 60 pin connectors vs a larger number of smaller connectors
was routability on a 2 layer board. I wanted to make it as simple as possible.
With 2 60 pin connectors, I can keep the signals going from the edge connector
to the ribbon cable connector all on one side. I?m mostly worried about
signal integrity. If this doesn?t work to my satisfaction, I?ll probably go to heavier
traces and put in some ground planes.
I have a PDP-11/40 with 128KW of core. 64KW of that core is in a second
BA11F chassis in another rack. I?ll be using that system (running memory
diagnostics mostly) to test out the cables.
I?ll have to work up some prices but I would expect it to be in the $100-$150
range. I purchased a 100? spool of 60 conductor ribbon cable and was agast
at the price. I was originally thinking twisted pair ribbon cable until I saw what
100? spool of that would run and decided to try just the straight ?grey? ribbon
cable for now.
TTFN - Guy
I just received the first of the BC11A paddle
boards (so that I can now create my own BC11A cables).
Cool!
For those who?d like to know, I probably will not
be making a production run of these boards unless I get a *lot* of interest (100 boards or
so). If you want some, let me know and if I get enough interest, I?ll do a production
run.
Considering DEC did something like this for the DWBUA, but with 4
30-pin cables (by necessity because of the BI backplane), is 2x60 pins
more cost-effective than 4x30? Mostly, I'm curious about how much
60-pin ends cost and the difference in cost between 2 lengths of
60-pin ribbon cable vs 4 lengths of 30-pin cable. The copper cost is
the same, but I wasn't sure what the current cost-per-foot is for
ribbon cable. I haven't bought a reel of it in 30 years, but I do
recall that 40 pin seemed to be more expensive than 4 times the length
of 10-pin, for example (since we used both where I worked in the 80s).
Either way, these days, the cost of 2x60 vs 4x30 is probably minor
compared to the fab costs of a "large" PCB with gold fingers. I'd be
curious to know how much a 10' BC11A-work-a-like will run, but I'm
vaguely interested in some pairs of boards. I do have plenty of
Unibus machines and RK05s, most with vintage cables, but I'm sure
several of them have crimps and dings and might fail close inspection.
27 Jan
27 Jan
10:06 a.m.
Considering DEC did something like this for the DWBUA, but with 4
30-pin cables (by necessity because of the BI backplane), is 2x60 pins
DEC (also?) made a board that brought the Unibus out on 3 40-way
cahles. The 11/730 I am restoring originally had a Unibus expansion
cabinet, and the arrangement was a board in the Unibus Out slot
of the CPU with 3 off 40 pin Berg connectors on it (and nothng else).
Then 3 cables (40 pin socket on each end) to a bulkhead panel with
3 40 way headers on it. That took a special screened cable with 3
40 way ribbon cables inside to another identical bulkhead, then
3 more cables to another PCB (different M-number, so perhaps
there were connection differences) in the Unibus In of the expansion
cabinet.
I have removed that from my 11/730 as I just want a single-cabinet system,
but of course have kept all of it.
Didn't we have a thread on this a few months back?
more cost-effective than 4x30? Mostly, I'm
curious about how much
60-pin ends cost and the difference in cost between 2 lengths of
Practically, I would not want to use 60 way connectors and cable. They are
not as easy to get as the 40 way ones.
[I've never seen a 30 way IDC socket. 34 way are common of course.]
-tony
10:48 a.m.
On Wed, Jan 27, 2016 at 10:06 AM, tony duell <ard at p850ug1.demon.co.uk> wrote:
I just did some pricing and 60-way cable is a touch pricey. Through
cable surplus vendors, I saw one quote of $1.33/ft and from the same
vendor, prices close to $0.30/ft for 28-34-way (you need twice as many
feet, of course, but it's still half the cost. Not a big deal for one
10' run, but trying to cable up 4 RK05s, for example, it would start
to add up). 3x 40-way wouldn't be too bad, if designing from scratch
(I totally get the design goal of a simple double-sided paddle card to
simplify the construction of that - I'm not complaining about Guy's
design, just investigating costs for different methods. For just one
pair of paddle cards and one set of cables, the differences aren't
going to be enough to matter. Wanting/needing multiples for multiple
systems or drives might start to tip the balance).
DEC (also?) made a board that brought the Unibus out
on 3 40-way
cahles.
I don't know that I've seen that exact one, but it sounds like
something they would have done.
I have removed that from my 11/730 as I just want a
single-cabinet system,
but of course have kept all of it.
Didn't we have a thread on this a few months back?
Perhaps, but I don't remember if it was related toGuy's paddle boards
or just general kicking around of "how do I get "new" BC11A cables?"
more
cost-effective than 4x30? Mostly, I'm curious about how much
60-pin ends cost and the difference in cost between 2 lengths of
Practically, I would not want to use 60 way connectors and cable. They are
not as easy to get as the 40 way ones. [I've never seen a 30 way IDC socket. 34 way are
common of course.]
I'd have to check the paddle card with my DWBUA, but I'm pretty sure
the cables are 30-pin to 30-pin, not 30-pin to 34-pin (the VAXBI
backplane is studded in 30-pin keyed connector spots (6 per backplane
slot) and are the only way to get signals to/from BI options. We made
a COMBOARD VAXBI (only sold a handful - too late in the game to
matter) and took two 30-pin cables out to our I/O bulkhead dual EIA
connector plate. The DWBUA has 4x 30-pin cables to get Unibus from
the BI board to the paddle card in the DD11DK. They certainly aren't
common. I only mentioned them because I've seen it done. 3x40-way is
probably more sensible since 40-pin parts and cable were once
incredibly common.
-ethan
1:31 p.m.
On Jan 27, 2016, at 7:48 AM, Ethan Dicks
<ethan.dicks at gmail.com> wrote:
Practically, I would not want to use 60 way connectors and cable. They are
not as easy to get as the 40 way ones.
I just did some pricing and 60-way cable is a touch pricey. Through
cable surplus vendors, I saw one quote of $1.33/ft and from the same
vendor, prices close to $0.30/ft for 28-34-way (you need twice as many
feet, of course, but it's still half the cost. Not a big deal for one
10' run, but trying to cable up 4 RK05s, for example, it would start
to add up). 3x 40-way wouldn't be too bad, if designing from scratch
(I totally get the design goal of a simple double-sided paddle card to
simplify the construction of that - I'm not complaining about Guy's
design, just investigating costs for different methods. For just one
pair of paddle cards and one set of cables, the differences aren't
going to be enough to matter. Wanting/needing multiples for multiple
systems or drives might start to tip the balance).
I just ran some numbers (Digikey*, so YMMV) and here?s what I came
up with:
60 pin connectors (PCB & cable ends): $46.60
4 PCB connectors
8 cable connectors
2 10? 60 conductor cable: $42.12
40 pin connectors (PCB & cable ends): $50.46
6 PCB connectors
12 cable connectors
3 10? 40 conductor cable: $45.30
I believe when I started these, I looked at both. Given that the 60 pin
solution is marginally cheaper (going with new parts which if I?m selling
that?s what I use) and results in fewer connectors (and was simpler to
route).
Note, that *if* I do end up selling the paddle boards, I?ll sell just the
boards and the connectors on the boards. I will *not* be selling the
cable itself (nor the cable connectors).
As you can see from the above, any solution is going to be expensive.
This does not include the boards (since the price for either solution is
the same). I just wanted to illustrate the comparison between 60 and
40 pin cables.
TTFN - Guy
* - I used the same family of connectors for both (from 3M). The only
difference was the total number of pins (60 vs 40). This wasn?t the
most expensive solution either and was pretty close to the cheapest
as I recall but I this is what I?m using for my ?prototype? run and will
probably investigate more fully if/when I do a production run (of course
if folks want just the bare boards they?re free to ?do their own thing?).
2:03 p.m.
On Wed, Jan 27, 2016 at 1:31 PM, Guy Sotomayor <ggs at shiresoft.com> wrote:
No arguments about the Digikey pricing. I'm not shocked it comes out
that high. I get needing to price new parts from reliable vendors,
but as a consumer of kit items, I frequently hit my own junk boxes (I
have a lot of cable supplies) and definitely hit surplus vendors. In
the "real world" for the last run of COMBOARDs, I got the cost down
25% by not ordering certain parts (RAM, CPU...) from
Digikey/Mouser/Allied, etc. OK for a sunset build, but not for a new
product.
On Jan 27,
2016, at 7:48 AM, Ethan Dicks <ethan.dicks at gmail.com> wrote:
I just did some pricing and 60-way cable is a touch pricey. Through
cable surplus vendors, I saw one quote of $1.33/ft...
I just ran some numbers (Digikey*, so YMMV) and here?s what I came
up with... Note, that *if* I do end up selling the paddle boards,
I?ll sell just the
boards and the connectors on the boards. I will *not* be selling the
cable itself (nor the cable connectors).
Sure.
As you can see from the above, any solution is going
to be expensive.
This does not include the boards (since the price for either solution is
the same). I just wanted to illustrate the comparison between 60 and
40 pin cables.
I see the numbers from Digikey and, yeah... quite a bit for those
endpoints, no matter which configuration.
...this is what I?m using for my ?prototype? run and
will
probably investigate more fully if/when I do a production run (of course
if folks want just the bare boards they?re free to ?do their own thing?).
If you do bare boards (no connectors), I'd probably be interested in
several sets. If you need to supply board connectors to keep your
volume up, I'll probably still be interested but perhaps not quite as
many sets.
-ethan
2:13 p.m.
On Jan 27, 2016, at 11:03 AM, Ethan Dicks
<ethan.dicks at gmail.com> wrote:
If you do bare boards (no connectors), I'd probably be interested in
several sets. If you need to supply board connectors to keep your
volume up, I'll probably still be interested but perhaps not quite as
many sets.
I offer parts kits as a convenience not as a requirement. The big costs
are just getting the number of boards up.
Of course when I start producing products that are SMD with pre-programmed
parts (ie MEM11A), those will be fully assembled and tested just because I
don?t want to handle the support issues and component choice variations as
well as I?ll have the boards assembled before they get to me.
TTFN - Guy
2:48 p.m.
On Wed, Jan 27, 2016 at 2:13 PM, Guy Sotomayor <ggs at shiresoft.com> wrote:
Sure. I just didn't assume because some kits are priced around volume
orders of PCB+parts and it screws with the inventory to sell bare
boards. I'm usually happiest to buy bare board kits or board+weird
parts.
On Jan 27, 2016, at 11:03 AM, Ethan Dicks
<ethan.dicks at gmail.com> wrote:
If you do bare boards (no connectors), I'd probably be interested in
several sets.
I offer parts kits as a convenience not as a requirement. The big costs
are just getting the number of boards up. Of course when I start producing products that are SMD
with pre-programmed
parts (ie MEM11A), those will be fully assembled and tested just because I
don?t want to handle the support issues and component choice variations as
well as I?ll have the boards assembled before they get to me.
Sure. I've been on the support end of selling through-hole kits, but
I wouldn't want to have to support SMD kits. I have bought and
assembled many kits with SMD parts and have enjoyed success, but not
everyone's builds go smoothly (and even I've had to occasionally fix
my own screwups).
And I'm still interested in at least one MEM11A when it gets to that.
I have this 11/20 that needs some stuffing.
-ethan
3:03 p.m.
On Jan 27, 2016, at 11:48 AM, Ethan Dicks
<ethan.dicks at gmail.com> wrote:
Sure. I've been on the support end of selling through-hole kits, but
I wouldn't want to have to support SMD kits. I have bought and
assembled many kits with SMD parts and have enjoyed success, but not
everyone's builds go smoothly (and even I've had to occasionally fix
my own screwups).
And I'm still interested in at least one MEM11A when it gets to that.
I have this 11/20 that needs some stuffing.
My current plan re:MEM11A is that I?m going to build a prototype using an
FPGA eval board that I already have. It has a 100pin Hirose connector on
it that brings out a bunch of I/Os. The ?prototype? board will contain all of
the ?other? components (FRAM, UARTs, Unibus I/F, etc) that don?t exist on
the eval board. It will strictly be a bench setup. This is also why I did the
paddle boards?I need a way to get from the prototype to the Unibus. ;-)
BTW, I?ve been spending my evenings recently re-writing the simulator in
C to get the functionality that I need to complete the rest of the testing and
coding of the uCode. I?ve also started writing verilog for the various parts
on the prototype board. I?m using CPLDs at present but it?s not clear that
with the partitioning I have, that it?ll all fit (right now I have stuff partitioned
into 3 CPLDs just for I/O count).
I?m planning that most of what I do for the prototype board will translate over
to the real board so I won?t have to end up re-writing a lot of code (either
the J1 uCode or verilog) but I expect that there will be some differences
mainly around how things are packaged between FPGA and CPLD(s).
TTFN - Guy
26 Jan
26 Jan
6:26 p.m.
My memories of a bloody back of hands are coming back.
We did the same with grant cards. Made a bacth of longer ones, and then
updated again by going three wide, the 3rd just for stability.
Remember the DEC FE seeing them one day, made him very happy by giving him
a few of them.
-pete
On Tue, Jan 26, 2016 at 2:46 PM, Guy Sotomayor <ggs at shiresoft.com> wrote:
I just received the first of the BC11A paddle boards
(so that I can now
create my own BC11A cables). I made two variants of the boards. One where
the cables come out the ?top? and another where the cables come out the
?side? (so you don?t have to ?fold? the cable in a number of different
applications such as RK05s).
The ones that arrived today are the S or ?side? variation (as opposed to
the T or ?top? variation). I immediately noticed (why I didn?t see this
when I was reviewing the board layout previously I?ll never know) is that
one of the cable connectors is too close to the edge. It wouldn?t be a
problem except that I chose to use connectors that have cable retention
latches. The latches protrude from the board edge by ~0.02?. So once I
verify that these work, I?ll move the one connector inward ~0.5?. I have
plenty of space, I just don?t know why I put the connector that close to
the edge. *sigh*. For these boards, I?ll probably just file the latch
down a bit to get the appropriate clearance.
BTW, trying to get the connectors onto the board (I haven?t soldered them
in yet) was tight. That?s the problem with trying to get 60 pins to all
line up correctly. But they fit and are *snug*. I probably won?t go and
do any testing until I get the other boards in so it may be a couple more
weeks (the boards weren?t supposed to *ship* before 2/8).
For those who?d like to know, I probably will not be making a production
run of these boards unless I get a *lot* of interest (100 boards or so).
If you want some, let me know and if I get enough interest, I?ll do a
production run.
TTFN - Guy
3099
days inactive
3100
days old
10 comments
4 participants
participants (4)
-
ard＠p850ug1.demon.co.uk -
ethan.dicks＠gmail.com -
ggs＠shiresoft.com -
pete＠petelancashire.com