Wire wrap
Dave Boland
a discussion on equipment, technique, performance, and reliability. The
reason for this is that pre-prototype printed cards seem to be a waist
of money because of all of the changes to the cards as soon as they
arrive (most due to the "hey, while your at it" syndrome).
The only downside is the use of surface mount devices, but we can likely
use adapters at that development stage.
Failing any documentation that would impress management, can anyone
answer these questions?
1. If a card is going to be WW, are pads and PTH's really needed? The
reason for the question is that there is no ground plane anyway, and
soldering to the socket lead seems like it is as good as soldering to a
pad. Components soldered are bypass capacitors as close to the device
as possible.
2. What is the number of connections that make power WW equipment
justifiable? I like to see development done (early stages) modularly,
so each card may have between 125 to 250 connections. A typical project
is only a few cards (modules). Once the design is stable we go for PC
cards anyway.
Any useful help would be appreciated.
Thanks,
Dave
Paul Burke
> Can anyone point me to a good document(s) on wire wrap? I'm looking for
> a discussion on equipment, technique, performance, and reliability. The
> reason for this is that pre-prototype printed cards seem to be a waist
> of money because of all of the changes to the cards as soon as they
> arrive (most due to the "hey, while your at it" syndrome).
DON'T wire wrap except for trivial circuitry. It takes an age to do,
it's difficult to check, it's bulky and uncomfortable when someone
leaves one on your seat, you need adaptors for everything except DIL and
leaded passives, modifications tend to snowball in complexity. I used to
use it a lot, until low- cost PCB prototypes like PCB Pool came along.
Changes are an advantage, you are actually creating debugged production
tools as you carry out modifications.
No, if you value your time and sanity, use PCBs.
If you want rapid and flexible prototyping, create some of your PCB
prototypes so that they can be used as sub- modules in future projects,
on wire-wrap or matrix, but that's as far as I'd go now.
>
> 1. If a card is going to be WW, are pads and PTH's really needed?
They stop the pins from flopping about. It's already a cat's cradle.
Paul Burke
Tam/WB2TT
We gave up on wirewrap about 15 years ago. You could sometimes spend a week
looking for broken wires, or wires that had shorted to a pin that it went
around. Not useful at high frequencies, especially analog. SM requires a
cluge, and at the time, SM only used .050 spacing.
For a while we used Multiwire for digital boards, but that is harder to make
channges on than a PC board.
Tam
"Dave Boland" <NOSPAMd...@stny.rr.com> wrote in message
news:OMyCc.246630$hY.3...@twister.nyroc.rr.com...
John Larkin
<NOSPAMd...@stny.rr.com> wrote:
>Can anyone point me to a good document(s) on wire wrap? I'm looking for
>a discussion on equipment, technique, performance, and reliability. The
>reason for this is that pre-prototype printed cards seem to be a waist
>of money because of all of the changes to the cards as soon as they
>arrive (most due to the "hey, while your at it" syndrome).
>
If you make them pay for multilayer boards, and pay for revisions,
they'll be a little more willing to do preliminary design reviews, and
stick with the results.
>The only downside is the use of surface mount devices, but we can likely
>use adapters at that development stage.
The big downside is the huge amount of labor involved, and the rotten
high-speed performance. 5-day multilayer boards are dirt cheap these
days. Think about it, design it carefully, make a board, and sell it.
Design-by-breadboard is a progressive, destructive addiction.
John
Ruediger Kluge
if you have a lot of DIL I Cs you can wire wrap. For wire wrap you need
sockets with pins with the right shape (rectangular) and length(number
of turns).
http://www.tecratools.com/pages/tecalert/wirewrap_guide.html
And you need a grid or 0.1".
So my proposal is nearly the same as from Paul.
I use electro mechanical interface modules (S-sub, smb, modular jacks,
BNC, power connectors...) which are in 0.1" grid, which can be wire
wrapped, soldered or plugged into standard sockets. Or they can even
plugged in breadboards. From here I make the connections to more complex
and special PCB modules.
regards
Rüdiger
--
http://www.conelek.com/
Norm Dresner
news:OMyCc.246630$hY.3...@twister.nyroc.rr.com...
I have a substantial investment in W/W equipment but my own rules are:
1. I'll wire-wrap (almost) all one-off [and sometimes two's as well]
circuits that aren't "too large"
2. I won't wire-wrap any circuit that I can't test easily. This means
a) it's a small circuit
b) it's a medium [or even somewhat large] circuit that can be built and
tested *modularly* and I build and test as I go.
c) it's not either RF or too high in frequency for logic. I'm very shy
going above 10-15 MHz this way
d) Unless it's a very simple circuit in which I can route wires very
carefully, it won't be an analog circuit either (but 555's are exempt from
this rule).
e) If it's a large circuit and I can't follow rule (b) then I use the
likelihood of changes as a determining factor.
3. There is no [convenient] PCB layout available. Since I use EagleCAD,
this generally forces me forego W/W unless there's substantial uncertainty
about the circuit design.
Norm
Jan Panteltje
On a sunny day (Thu, 24 Jun 2004 07:59:37 -0700) it happened John Larkin
<jjla...@highlandSNIPtechTHISnologyPLEASE.com> wrote in
<4qqld09l8dskj8r38...@4ax.com>:
Yea, but how good are you?
I have seen these guys soldering little wires, cutting tracks...
Not one but 50 boards, hehe
At least calculate for 2?
JP
-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
*** Usenet.com - The #1 Usenet Newsgroup Service on The Planet! ***
http://www.usenet.com
Unlimited Download - 19 Seperate Servers - 90,000 groups - Uncensored
-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
Dave Boland
Clearly, everyone seems against use of Wire wrap. I'm not that fond of
it myself, which explains why I don't use it much and have some
questions. Especially when you consider the cost of some of the card
shops today, it seems like wire wrap is not needed.
The problem is the reality of early development. A company pays an
engineer to layout a card with either a PCB program or a CAD program.
Card data is sent and in a few work days the card arrives. Within the
next few days, lands are cut, more holes drilled, a lot of wire wrap
wire used for "engineering changes", until the card is worthless. Let's
not forget that most of the quick turn cards are only two sided, so they
don't have a reference plane to stabilize the trace impedance anyway.
So, after seeing this go on for a while, it seems like there should be a
faster, more flexible way to do early development. One that the design
engineer can do in hours, and change as needed, etc. I thought wire
wrap may be that answer, but perhaps not. Is there a better way to do this?
Keep in mind that these are concept cards, well ahead of any product
prototype, so looks and long term reliability are not an issue. In
fact, I would consider the white board if the contacts were a little
more reliable (I have seen components fall out of the white boards when
moved).
Dave,
Jan Panteltje
>
>So, after seeing this go on for a while, it seems like there should be a
>faster, more flexible way to do early development. One that the design
>engineer can do in hours, and change as needed, etc. I thought wire
>wrap may be that answer, but perhaps not. Is there a better way to do this?
>
>Keep in mind that these are concept cards, well ahead of any product
>prototype, so looks and long term reliability are not an issue. In
>fact, I would consider the white board if the contacts were a little
>more reliable (I have seen components fall out of the white boards when
>moved).
>
>Dave,
The way I do that is use veroboard with round iles (.2 inch round holes, no
stripes, solder in sockets, use some flatcable, split the wires, and
solder the connections one by one.
Have some Eurocards with uP and say 30 CMOS chips that way, still work after
20 years... You can throw them around.
You need good soldering, good eyesight, check every soldering connection.
Of cause the ones I did that way were all 100% working, so I could have made
the PCB directly hehe.
But nice to add and test things.
These days with FPGA and SMD .. not so useful anymore.
John Larkin
<NOSPAMd...@stny.rr.com> wrote:
>All,
>
>Clearly, everyone seems against use of Wire wrap. I'm not that fond of
>it myself, which explains why I don't use it much and have some
>questions. Especially when you consider the cost of some of the card
>shops today, it seems like wire wrap is not needed.
>
>The problem is the reality of early development. A company pays an
>engineer to layout a card with either a PCB program or a CAD program.
>Card data is sent and in a few work days the card arrives. Within the
>next few days, lands are cut, more holes drilled, a lot of wire wrap
>wire used for "engineering changes", until the card is worthless.
Sounds like you should spend more time thinking and less hacking. Most
of our 6 or 8-layer VME boards work the first time. You can do that if
you want to.
> Let's
>not forget that most of the quick turn cards are only two sided, so they
>don't have a reference plane to stabilize the trace impedance anyway.
Multilayer protos are cheap now. See the ads in the backs of EE Times
or EDN.
>
>So, after seeing this go on for a while, it seems like there should be a
>faster, more flexible way to do early development. One that the design
>engineer can do in hours, and change as needed, etc. I thought wire
>wrap may be that answer, but perhaps not. Is there a better way to do this?
>
>Keep in mind that these are concept cards, well ahead of any product
>prototype, so looks and long term reliability are not an issue. In
>fact, I would consider the white board if the contacts were a little
>more reliable (I have seen components fall out of the white boards when
>moved).
Why does a concept need a card?
John
Joerg
While you compare against prototype PCBs keep in mind the incredible
amount of time some poor technician needs for wire wrapping. And the
cost this adds.
Also, even vanilla logic chips are several times faster today than 25
years ago. This adds to the signal integrity problems on a wrapped board.
I have never, ever, allowed any of my designs to be wire wrapped. As a
result my boards were done a lot faster than the wrapped ones. It is no
fun waiting for the others who are still chasing crosstalk, loose wires
and so on. I have seen engineers on the verge of bursting into tears
because their wrapped designs didn't work, the deadline was just hours
away and the boss nervously standing behind them.
My advice is the same that others provided: Don't wire wrap.
Regards, Joerg
Peter Bennett
<NOSPAMd...@stny.rr.com> wrote:
>Can anyone point me to a good document(s) on wire wrap? I'm looking for
>a discussion on equipment, technique, performance, and reliability. The
>reason for this is that pre-prototype printed cards seem to be a waist
>of money because of all of the changes to the cards as soon as they
>arrive (most due to the "hey, while your at it" syndrome).
>
3M makes (or used to make) a prototyping system that used wire-wrap
wire, and IDC connection points rather than wire wrap posts. I found
it much easier to use (and re-work) than wire-wrap.
However, I haven't used that system for some time - most of my
projects of any complexity have all the logic inside Altera FPGAs,
which are even easier to modify than the IDC system.
--
Peter Bennett VE7CEI
email: peterbb4 (at) interchange.ubc.ca
GPS and NMEA info and programs: http://vancouver-webpages.com/peter/index.html
Newsgroup new user info: http://vancouver-webpages.com/nnq
Nico Coesel
>All,
>
>Clearly, everyone seems against use of Wire wrap. I'm not that fond of
>it myself, which explains why I don't use it much and have some
>questions. Especially when you consider the cost of some of the card
>shops today, it seems like wire wrap is not needed.
>
>Keep in mind that these are concept cards, well ahead of any product
>prototype, so looks and long term reliability are not an issue. In
>fact, I would consider the white board if the contacts were a little
>more reliable (I have seen components fall out of the white boards when
>moved).
I used to use thin insulated wire. Nowadays I use thin wire which is
intended to wind transformers and so on. The main advantage is you
don't need to strip it. Just heat it with the soldering iron, the
insulation will melt and presto: you have a tinned wire. There is also
special wire available for this purpose in different colors.
--
Reply to nico@nctdevpuntnl (punt=.)
Bedrijven en winkels vindt U op www.adresboekje.nl
Nico Coesel
>Hi Dave,
>
>While you compare against prototype PCBs keep in mind the incredible
>amount of time some poor technician needs for wire wrapping. And the
>cost this adds.
>
>I have never, ever, allowed any of my designs to be wire wrapped. As a
>result my boards were done a lot faster than the wrapped ones. It is no
>fun waiting for the others who are still chasing crosstalk, loose wires
>and so on. I have seen engineers on the verge of bursting into tears
>because their wrapped designs didn't work, the deadline was just hours
>away and the boss nervously standing behind them.
I've seen large wire-wrapped -production- boards (in a computer
tape-drive). I guess wire wrapping is in a way a true art. If done
well, wire wrapping is more reliable than soldering (say >100 years
versus 30 years).
Rene Tschaggelar
> Can anyone point me to a good document(s) on wire wrap?
>
> [ don't wire wrap ]
>
> Any useful help would be appreciated.
To start with, digital stuff is done with CPLDs and FPGAs.
And for some analog stuff, there are systems on a chip available.
both for in circuit programming.
Rene
--
Ing.Buero R.Tschaggelar - http://www.ibrtses.com
& commercial newsgroups - http://www.talkto.net
Joerg
>I used to use thin insulated wire. Nowadays I use thin wire which is
>intended to wind transformers and so on. The main advantage is you
>don't need to strip it. Just heat it with the soldering iron, the
>insulation will melt and presto: you have a tinned wire. There is also
>special wire available for this purpose in different colors.
>
>
Just make sure you are ventilating the fumes away. I have some of that
stuff as well and the stench that comes off when the lacquer melts is
awful. The usual transformer wire may not be such a good idea as the
temps to melt are higher and the fumes might be toxic. So be careful.
Regards, Joerg
JeffM
The concensus of documents on the web (as in this thread)
is *avoid wirewrap*.
A nice overview page for PCB/PWB is
http://en.wikipedia.org/wiki/Printed_circuit_board
Joerg
> To start with, digital stuff is done with CPLDs and FPGAs.
Yes, and that takes away all this prototyping effort to a large extent
because you can "re-wire" things on the PC.
> And for some analog stuff, there are systems on a chip available.
> both for in circuit programming.
I know this is off topic here but are there any newer ones out that can
do more than a few MHz? What I saw so far doesn't contain much in terms
of speed and quantity but was really expensive.
For analog I'd seriously advise against wire wrap. I have seen people
try it but I have never seen that work in the end. Sometimes the radio
in the lab went quiet when they turned on the power ;-)
Regards, Joerg
Joerg
>I've seen large wire-wrapped -production- boards (in a computer
>tape-drive). I guess wire wrapping is in a way a true art. If done
>well, wire wrapping is more reliable than soldering (say >100 years
>versus 30 years).
>
>
To be honest I wouldn't buy a product if I knew it had wrap boards in
there. Art? Yes, in one company there were only two techs that could do
it. And I have seen wrapped boards work, at least for a while, but
usually only after they soldered all the posts.
Regards, Joerg
Greg Neff
<NOSPAMd...@stny.rr.com> wrote:
>Can anyone point me to a good document(s) on wire wrap? I'm looking for
>a discussion on equipment, technique, performance, and reliability. The
>reason for this is that pre-prototype printed cards seem to be a waist
>of money because of all of the changes to the cards as soon as they
>arrive (most due to the "hey, while your at it" syndrome).
>
>The only downside is the use of surface mount devices, but we can likely
>use adapters at that development stage.
>
>Failing any documentation that would impress management, can anyone
>answer these questions?
>
>1. If a card is going to be WW, are pads and PTH's really needed? The
>reason for the question is that there is no ground plane anyway, and
>soldering to the socket lead seems like it is as good as soldering to a
>pad. Components soldered are bypass capacitors as close to the device
>as possible.
>
1) Forget copper clad boards. Use bare 0.62" thick FR4 boards with
0.042" holes on a 0.1" grid.
2) If you have any parts that won't fit through the holes then now is
the time to mark and drill. Don't forget mounting holes.
3) Power distribution is the most important issue. Stake Vector
T46-5-9 terminals (using a proper staking tool) in two rows at a space
of 0.3" Place 0.1uF axial ceramic caps between each pair of posts,
wrap the leads around the posts once (360 degrees) and don't trim the
ends. Lay 14 AWG or even 12 AWG solid copper wire along the outside
of the posts on top of the capacitor leads. Wrap each capacitor lead
up over the copper wire and back on to the post. This ties the copper
wire to the posts. Trim the leads. Oh yeah, add some tantalum caps
along the way, maybe every 10 pairs of posts. Use a heavy soldering
iron or gun to solder this all together. Careful with the heat on the
tantalums. Enjoy a beverage while the contraption cools. Now you
have reasonably low impedance power for low-speed applications, and a
rigid PCB.
4) Place and solder any miscellaneous stuff that isn't going to be
wrapped.
5) When you place the IC sockets beside the bus bars, keep the ground
end of the IC closest to the bus. I liked to glue down my sockets
using epoxy so they wouldn't squirm while wrapping. Mark the
reference numbers on the bottom of the board. Don't forget that the
chips are upside down when you are counting pins. If you want to get
fancy use OK Industries Socket-Wrap I.D. tags.
6) Make your ground connections to the ICs first, then the VCC
connections. Have fun with the rest.
7) For troublesome high speed or sensitive signals you can try running
twisted pairs. For single-ended stuff use a ground wire as the second
wire in the pair.
8) Say goodbye to your family for a couple of weeks, because that's
how long it's going to take you to get the mess to work.
>2. What is the number of connections that make power WW equipment
>justifiable? I like to see development done (early stages) modularly,
>so each card may have between 125 to 250 connections. A typical project
>is only a few cards (modules). Once the design is stable we go for PC
>cards anyway.
>
>Any useful help would be appreciated.
>
>Thanks,
>Dave
The last time I bothered with wire-wrap I was driving an AMC Gremlin.
Never mind why I was driving an AMC Gremlin. It could have been worse,
it could have been a Pacer or a Pinto. Anyway, if done properly
wire-wrap can be reliable. The wire-wrap connections are gas tight if
done correctly. IMHO, the main reliability problem is the quality of
the socket contacts that grab the IC leads. Also, signal integrity is
a nightmare. If you are using plain old slow TTL or CMOS then it can
work. With todays fast logic families I would not hesitate to laugh
out loud at an engineer that would suggest it. These days wire-wrap
is strictly starving-student stuff, and the enclosed pointers are for
their benefit.
================================
Greg Neff
VP Engineering
*Microsym* Computers Inc.
gr...@guesswhichwordgoeshere.com
j.b. miller
retired last year( 50), I still enjoy the 'art of wire wrapping'. I've made
my money from wrapping 8051 series MCU to do remote energy control systems
using 3 tall wirewrap sockets on Radio Shack perfboard. Used standard
11M0592 xtals and some LM34 analog chips,never had a noise problem,even in
the,ah, crappiest enviroments known( shop full of 3 phase welders).
It's too bad all the neat chips are SMT otherwise I'd still be 'playing' in
electronics. Today, it's quick and easy to have a PC draw up the schematic,
design the PCB, Email the design to a fab shop and have it delivered in a
week or too. Takes all the fun out of 'wrappin'. I still have a whack of
wire ,sockets and a lot of pogo pins here,relics of a glorious former time.
Jay
Joerg
Seriously, you should post your story somewhere, especially how you
managed to retire at 50. Assuming you have accumulated enough dough to
last through the next 40 years or so :-)
For SMT you can get tiny stand-up solder boards, even for these
miniscule MSOP packages. But it's hard to properly place the MSOP parts
on there. Theoretically these could then be soldered onto one side of
your wrap DIP sockets and you could happily wire wrap again. However,
since this stuff became a lot faster respectively higher in bandwidth it
may be hard to tame in terms of circuit stability.
Regards, Joerg
John G
"j.b. miller" <invalid...@cogeco.ca> wrote in message
news:PgJCc.17668$XY6.2...@read2.cgocable.net...
are easier to design and make but...
All those 1960 1970 IBM mainframes and there were a lot, probably many
still in use, had miles and miles of wire wraps on the back panels.
One of the things taught to all customer engineers working on such
machines was how to do engineering changes amongst all those wraps.
--
John G
Wot's Your Real Problem?
John Larkin
wrote:
Solder joints die in 30 years? I still use GR and Tek instruments that
are 40 years old, and, if they break, it's not the soldering.
John
K Williams
> Dave,
> We gave up on wirewrap about 15 years ago. You could sometimes
> spend a week looking for broken wires, or wires that had shorted
> to a pin that it went around. Not useful at high frequencies,
> especially analog. SM requires a cluge, and at the time, SM only
> used .050 spacing.
I used wire-wrap for a couple of decades. It's fine for what it was
intended. Wire-wrap was used extensively in mainframes in the
'70s. Indeed, I've built entire test systems with wire-wrap. One
simply has to be a tad careful. One (close to thirty years ago) had
an 80MHz oscillator (quickly divided by two) and was done in
MECL-10K. A well done wire-wrap board is a rather good 100ohm
transmission environment. Decoupling and power-distribution was
the area that got most people. Augat had some very nice WW
prototype boards, with these miniscule clips to wire power/ground
directly to the pins of interest (technicians hated them).
One system I built had a half dozen 12"x12" boards with a few
thousand wires each. A skilled technician could do a board in a
week with minimal errors and *no* "cold-flows". Of course the key
is "skilled". This particular system ran at ~50MHz using 74ASxxx,
FS, and ALS DIP stuff, along with some DRAM.
> For a while we used Multiwire for digital boards, but that is
> harder to make channges on than a PC board.
I did some multi-wire. It was far harder than wire-wrap. We never
had enough to do to justify keeping the technicians' skills up.
OTOH, I'd not use wire-wrap for any major project these days. As
has been pointed out here, PC boards, even multi-layer are simply
too cheap. ...and modern packaging doesn't fit well with the
technology. Programmable logic has pretty much replaced 74xx
random logic, and it doesn't fit on WW boards very well. WireWrap
wasn't all that bad, and had its place.
James T. White
> 3M makes (or used to make) a prototyping system that used wire-wrap
> wire, and IDC connection points rather than wire wrap posts. I found
> it much easier to use (and re-work) than wire-wrap.
>
> However, I haven't used that system for some time - most of my
> projects of any complexity have all the logic inside Altera FPGAs,
> which are even easier to modify than the IDC system.
I used to use the 3M IDC stuff for prototypes as an alternative to wire wrapping
back in the early '80's and it worked pretty well for the relatively low speed
micros we were working with at the time. I think 3M discontinued the product as
I haven't seen any references to it in many, many years. The only sockets they
had available were for .3" and .6" DIPS plus you had the same problems with this
system as with wire wrap when it came to discrete components. You could mix
wire wrap and the 3M socket strips so you could handle most board connectors and
even PLCCs if you could find a wire wrap socket for them.
Fine pitch, surface mount parts and higher speed devices have pretty much put an
end to this sort of prototyping unless you are plan to build modules and
interconnect them as some of the other folks have suggested.
In the vein of "never saying never" though, I would be interested if anyone
knows of a source for the 3M IDC socket strips.
--
James T. White
SPAMjtwh...@SPAMhal-pcGUARD.org
Note: Remove SPAM-GUARD to reply.
Thaas
<NOSPAMd...@stny.rr.com> wrote:
>Can anyone point me to a good document(s) on wire wrap? I'm looking for
>a discussion on equipment, technique, performance, and reliability. The
>reason for this is that pre-prototype printed cards seem to be a waist
>of money because of all of the changes to the cards as soon as they
>arrive (most due to the "hey, while your at it" syndrome).
>
>The only downside is the use of surface mount devices, but we can likely
>use adapters at that development stage.
>
>Failing any documentation that would impress management, can anyone
>answer these questions?
>
>1. If a card is going to be WW, are pads and PTH's really needed? The
>reason for the question is that there is no ground plane anyway, and
>soldering to the socket lead seems like it is as good as soldering to a
>pad. Components soldered are bypass capacitors as close to the device
>as possible.
>
>2. What is the number of connections that make power WW equipment
>justifiable? I like to see development done (early stages) modularly,
>so each card may have between 125 to 250 connections. A typical project
>is only a few cards (modules). Once the design is stable we go for PC
>cards anyway.
>
>Any useful help would be appreciated.
>
>Thanks,
>Dave
In the '70s and early '80s minicomputers and their peripheral
controllers were usually wire wrapped. I can't point you to any
documentation or formal design guides on the web. I may have some
stuff lying about that addresses design issues, but nothing useful for
impressing PHBs or guiding neophytes.
To address question 1, we had power and ground planes on opposite
sides of the boards. There were pads for the signal pins, and the
boards were flow-soldered. We're talking ~500 16-pin ICs per board.
0.01uf decoupling caps were placed for every two ICs, except where
more were needed near line drivers and such. 47uF caps were
distributed across the boards.
In 1997 I consulted on the production of a copy of a wire-wrap
controller created by Raytheon for the Ballistic Missile Early Warning
program. The original manufacturer had discarded their wire-wrap
machines and production of the copies was contracted to a company in
Dallas. These people were miserable at translating a schematic to
wire-wrap and clueless about netlists.
The art/science of mass production wire-wrap technology involves
getting the tension right (so the wires don't cut themselves on the
sharp edges of the stake pins); limiting the number of wires in a
channel to avoid bulk; sequencing connections of clocks to avoid
transmission line induced race conditions; routing buses and clocks in
channels to avoid crosstalk at the wrong time.
A proper netlist lists the level of each wrap on the stake pins, the
wire's length and route. Some of this consisted of arcane parameters
like Table Rotation Position of the board in the wire-wrap machine
and routing pattern. All of this information is crucial to writing a
Design Change Order. A program that can generate a viable wirelist
for a Gardner Denver machine is non-trivial. Re-working boards
produced by hand without benefit of an organized wiring plan
contributes to alcoholism.
We had a few re-work girls who could shoot a 500 wire DCO blind drunk
without a mistake. But that performance can't be duplicated on a
board with wires going from one level on the pin to the next in
whatever order the original assembler felt like producing at the time.
Also note that wire color is important. An all yellow-wire board will
blind you. When installing a change use a contrasting color to aid
verification.
Problems with wire-wrapped boards ranged from nicked wires that
resulted from tension caused by overlaying wires tightening the wire
against a stake pin it turned past, wire scraps, bent and touching
pins, and noisy channels.
Digital logic only works when everything jumps at the same time. I
regularly had to re-address design problems in a CPU that used AMD's
2901 bit slice, whenever AMD produced a faster version. Buses would
switch at different times inducing noise into clock and other signals
that were previously happy where they were routed.
The speeds involved were only 10MHz. The lessons learned about a
given wire-wrap design's behavior are not translatable to PCB. These
days we have to narrow traces when they pass by vias! You are better
off addressing your design problems with better simulation, and you
might consider lower speed emulation to address the wish list
phenomena.
The above paragraph indirectly addresses question 2. You might do a
search for companies providing wire-wrap services. The number you
find and the volume of business they do should answer your question.
Personally I feel that for any board big enough to raise the question,
you are, with or without previous wire-wrap expertise, asking for a
headache.
Of course as an experiment a wire-wrap prototype might be useful in
convincing the PHBs to spring for more PCB prototypes in the design
cycle.
--
Thaas
Rene Tschaggelar
> Hi Rene,
>
>> To start with, digital stuff is done with CPLDs and FPGAs.
>
>
> Yes, and that takes away all this prototyping effort to a large extent
> because you can "re-wire" things on the PC.
>
>> And for some analog stuff, there are systems on a chip available.
>> both for in circuit programming.
>
>
> I know this is off topic here but are there any newer ones out that can
> do more than a few MHz? What I saw so far doesn't contain much in terms
> of speed and quantity but was really expensive.
It is definitely not off-topic.
As far as i know the analog-systems-on-a-chip are plain vanilla only.
Little bandwidth, high offset, high bias, not rail-rail, high noise
and such.
Keep yourself updated at http://www.cypress.com and look out for PSoC.
Rene
Nico Coesel
Joerg
>Solder joints die in 30 years? I still use GR and Tek instruments that
>are 40 years old, and, if they break, it's not the soldering.
>
>
And we have a Western Electric phone with crank from around 1925, in
working order. No solder joint problems there either. The only bad
solder joints I find are either from shoddy work where someone just
wasn't able to solder, or when people mount heat sinks with TO220
devices onto a PCB without stress reliefs. There the solder joints
vibrate into a pudding. But that's simply sub-optimal design.
In contrast to that I cannot even count the times I found bad wire wrap
connections on motherboards. Same on some of these "pressed-in"
motherboards which is why I usually insist on soldered construction.
Regards, Joerg
Joerg
>> I know this is off topic here but are there any newer ones out that
>> can do more than a few MHz? What I saw so far doesn't contain much in
>> terms of speed and quantity but was really expensive.
>
>
> It is definitely not off-topic.
> As far as i know the analog-systems-on-a-chip are plain vanilla only.
> Little bandwidth, high offset, high bias, not rail-rail, high noise
> and such.
>
> Keep yourself updated at http://www.cypress.com and look out for PSoC.
Thanks, Rene. I knew these and was hoping some more had come up. Maybe
there just isn't such big of a market that TI and others don't want to
jump onto the bandwagon. The idea is great but the performance of the
analog section isn't sufficient for many tasks and usually the whole
thing is out of cost range anyways for many of my designs. You can buy a
nice quad amp such as the LM324 for under 20 cents and it does the
negative rail. That price is hard to beat with SoC unless space is
really an issue. But even then you can go to bare die.
Regards, Joerg
Joerg
In the 80's I have seen IDC prototyping, too. And I have seen the grief
that came with it. Wires popping off but not far enough that you could
see it. So, a bus didn't work. Then you tugged a little and suddenly a
whole bunch of wires popped. Now three of the buses didn't work. I
believe the technique relied on multiple wires stacked into each IDC
position. Is that right? In telecom that's an absolute no-no.
But at least you could solder them shut once the board worked. If it
ever did, that is. With wrapping that was harder because that wire
didn't always accept solder easily.
Regards, Joerg
John Fields
---
You obviously either don't know what you're talking about, you don't
know how to wirewrap properly, or you've had the misfortune of being
surrounded by incompetents. I've been wirewrapping analog, digital,
and mixed signal breadboards and prototypes for over 30 years and have
had little trouble working up to about 30-40MHz reliably with AS edge
rates.
Of course forethought in part placement and wiring technique are
important, just as they are in PCB layout, but for some reason many
people seem to believe that just because they have a wirewrap gun in
their hand, that's an invitation to sloppiness. Then, when they
mistakenly take the invitation and turn out a piece of crap they blame
wirewrapping instead of themselves for the disaster. Your overall
blanket condemnation of wirewrapping seems to put you in that
category.
--
John Fields
John Fields
---
Wire-wrapping is usually done with silver-plated OFHC copper wire, so
soldering the wire to the posts should have been duck soup. That is,
I've never seen a problem with it unless the board was in service for
some years and then, for some reason, someone decided to try to solder
over the tarnish. Or perhaps they used magnet wire for the wrap,
thinking that the square posts would break through the enamel...
--
John Fields
Joerg
>You obviously either don't know what you're talking about, you don't
>know how to wirewrap properly, or you've had the misfortune of being
>surrounded by incompetents. I've been wirewrapping analog, digital,
>and mixed signal breadboards and prototypes for over 30 years and have
>had little trouble working up to about 30-40MHz reliably with AS edge
>rates.
>
>
What I said wasn't meant as condemnation. My apologies if it came over
differently. I just said I won't do it and do not consider it reliable
enough. That is my personal opinion and not a condemnation. The same
goes for some other mentioned techniques such as IDC or pressed in DIN
connectors. 20 years of experience taught me what works for me and what
doesn't.
Yes, I did wirewrapping as well when helping others debug their boards.
We were (usually) surrounded by well trained folks who knew this
technology. Prototype unit copies were even done at top notch wrapping
service providers. Still, the reliability did not reach the level of
soldered boards and soldered motherboards, especially for stuff that had
to be shipped air cargo.
30-40 MHz is doable. Even some more since the wire is quite consistent
if channeled properly and when avoiding insulation binding on the posts.
Problem was that even in the 80's our clocks and buses ran faster than
that. But most failures related to rough flights or truck rides,
something that soldered solutions weathered flawlessly.
>Of course forethought in part placement and wiring technique are
>important, just as they are in PCB layout, but for some reason many
>people seem to believe that just because they have a wirewrap gun in
>their hand, that's an invitation to sloppiness. Then, when they
>mistakenly take the invitation and turn out a piece of crap they blame
>wirewrapping instead of themselves for the disaster.
>
You are right but the techs we worked with were well trained in
wirewrap. Actually we didn't let anyone else near the wrap guns because
they were so expensive. They didn't worry about me as they knew I
wouldn't touch them anyways :-)
> Your overall
>blanket condemnation of wirewrapping seems to put you in that
>category.
>
>
>
I don't think so.
Regards, Joerg
Joerg
>Wire-wrapping is usually done with silver-plated OFHC copper wire, so
>soldering the wire to the posts should have been duck soup. That is,
>I've never seen a problem with it unless the board was in service for
>some years and then, for some reason, someone decided to try to solder
>over the tarnish. Or perhaps they used magnet wire for the wrap,
>thinking that the square posts would break through the enamel...
>
>
You may be right about contamination. Usually it took a lot of effort to
make these wrapped boards work and by that time the posts had been
touched countless times with fingers. Add in the popcorn that engineers
tend to munch on when burning the late night oil, and maybe a slight
nicotine/tar layer if they smoked (used to be legal once upon a
time...). That can make a mess. Then the square posts of these sockets
respectively DIN connectors were so long that the usual Weller tips had
a hard time to heat all that up.
The wire was always the good stuff and locked away at night, probably so
nobody used it for mundane soldering jobs.
Regards, Joerg
K Williams
> Hi John,
>
>>You obviously either don't know what you're talking about, you
>>don't know how to wirewrap properly, or you've had the misfortune
>>of being
>>surrounded by incompetents. I've been wirewrapping analog,
>>digital, and mixed signal breadboards and prototypes for over 30
>>years and have had little trouble working up to about 30-40MHz
>>reliably with AS edge rates.
>>
>>
> What I said wasn't meant as condemnation. My apologies if it came
> over differently. I just said I won't do it and do not consider it
> reliable enough. That is my personal opinion and not a
> condemnation. The same goes for some other mentioned techniques
> such as IDC or pressed in DIN connectors. 20 years of experience
> taught me what works for me and what doesn't.
Is that why IBM /360 and /370 mainframes were
wire-wrapped? ...because it is unreliable? Sheesh!
> Yes, I did wirewrapping as well when helping others debug their
> boards. We were (usually) surrounded by well trained folks who
> knew this technology. Prototype unit copies were even done at top
> notch wrapping service providers. Still, the reliability did not
> reach the level of soldered boards and soldered motherboards,
> especially for stuff that had to be shipped air cargo.
You were surrounded by incompetence, obviously.
> 30-40 MHz is doable. Even some more since the wire is quite
> consistent if channeled properly and when avoiding insulation
> binding on the posts. Problem was that even in the 80's our clocks
> and buses ran faster than that. But most failures related to rough
> flights or truck rides, something that soldered solutions
> weathered flawlessly.
I've done 80MHz ECL on WW, though admittedly not much was really at
80MHz (just the front end of a multiphase timing circuit). Yes,
but the mid 80's we'd move on, but WireWrap was used extensively in
the '70s and before. It is *not* unreliable. Gardner-Denver
showed that it was at least as reliable as printed wire.
>>Of course forethought in part placement and wiring technique are
>>important, just as they are in PCB layout, but for some reason
>>many people seem to believe that just because they have a wirewrap
>>gun in
>>their hand, that's an invitation to sloppiness. Then, when they
>>mistakenly take the invitation and turn out a piece of crap they
>>blame wirewrapping instead of themselves for the disaster.
>>
> You are right but the techs we worked with were well trained in
> wirewrap. Actually we didn't let anyone else near the wrap guns
> because they were so expensive. They didn't worry about me as they
> knew I wouldn't touch them anyways :-)
You were *surrounded* by incompetence. I've done a boat-load of WW,
though I wasn't certified for production (and had techs do most of
the work - better things to do). With the right tools and a little
practice it's not all that hard.
>> Your overall
>>blanket condemnation of wirewrapping seems to put you in that
>>category.
>>
>>
>>
> I don't think so.
I *know* so. WW is not in any way unreliable, except in the hands
of incompetence. That statement alone blows your credibility on
this issue. Wire-Wrap is perfectly fine, though it has been
eclipsed by printed-wire due to component packaging
incompatibilities, clock speed, and cheap PWB processes (including
tools).
--
Keith
James T. White
When I first started using the 3M IDCs I was a bit concerned about having wires
come loose, but only remember one instance where I had a problem with wires
popping out due to rough handling, It was an instance where I had two or three
wires stacked and got a little careless handling the board. Granted, loose
wires could be a pain to find but comparable to finding pin-to-pin shorts on a
wire wrapped board. If I remember correctly the 3M products were rated for two
wires in the IDCs. In practice you rarely had more than 1 wire in a contact
because unlike wire wrap, you could continue on to the next connection point
with the same wire. One thing that tended to help keep the wires in the
contacts was that we would install a clear plastic shield on the wire side of
boards once prototyping was finished. I've also seen the plastic shields used
with wire wrap boards to minimize the possibility of bending the wrap tails and
creating shorts.
Again this was just an aid to getting the design right before committing it to a
PCB layout which was considerably more costly than it is with the tools we have
today.
--
James T. White
SPAMjtwh...@SPAMhal-pcGUARD.org
Note: Remove SPAM-GUARD to reply.
"Joerg" <notthis...@removethispacbell.net> wrote in message
news:Ge%Cc.77386$Xq3....@newssvr29.news.prodigy.com...
Mark Zenier
They stuck it into their AP Products line for a couple of years and
then discontinued them.
I used it a lot but you had to remember the cardinal rules.
1) never more than two wires per "fork", (pushing in a third broke
the bottom one), and 2) never flex the wire sideways (metal fatigue in
the pinched part of the wire).
A possible cheap substitute may be wire wrap sockets and the Molex or
Panduit connectors that mate to .1 inch single row headers and have
IDC termials on the back.
Mark Zenier mze...@eskimo.com Washington State resident
John Fields
<notthis...@removethispacbell.net> wrote:
>Hi John,
>
>>You obviously either don't know what you're talking about, you don't
>>know how to wirewrap properly, or you've had the misfortune of being
>>surrounded by incompetents. I've been wirewrapping analog, digital,
>>and mixed signal breadboards and prototypes for over 30 years and have
>>had little trouble working up to about 30-40MHz reliably with AS edge
>>rates.
>>
>>
>What I said wasn't meant as condemnation. My apologies if it came over
>differently. I just said I won't do it and do not consider it reliable
>enough. That is my personal opinion and not a condemnation. The same
>goes for some other mentioned techniques such as IDC or pressed in DIN
>connectors. 20 years of experience taught me what works for me and what
>doesn't.
---
What works for you and what doesn't obviously has little to with the
reliability of the process and more to do with your capabilities and
what you consider to be beneath your "standards". Toolmakers like
Gardner-Denver, Standard Pneumatics, OK Industries,
http://www.cooperhandtools.com/brands/wire_wrap/
http://www.standardpneumatic.com/product/wirewrap/index.html
http://www.okindustries.com/products/4.1.1.htm
and their customers obviously don't share your views, in that
wire-wrap is still very much alive and well after all these years,
with myriad tools, accessories, and wire available.
Vector
http://www.vectorelect.com/Catpdf/Pages%2069-70.pdf
even makes single stamped, bifucated wire-wrap pins for discretes, and
sockets for all maner of chips are available from lots of folks; check
Digi-Key or any distributor for a clue. For a little further
enlightenment, Google "wire-wrap" for a surprise.
---
>Yes, I did wirewrapping as well when helping others debug their boards.
>We were (usually) surrounded by well trained folks who knew this
>technology. Prototype unit copies were even done at top notch wrapping
>service providers. Still, the reliability did not reach the level of
>soldered boards and soldered motherboards, especially for stuff that had
>to be shipped air cargo.
---
Poppycock. I've wire-wrapped stuff that's been used by the US Navy,
at sea, as well as stuff that's been used on-shore in a marine
environment, and never had any complaints about failures.
---
>30-40 MHz is doable. Even some more since the wire is quite consistent
>if channeled properly and when avoiding insulation binding on the posts.
>Problem was that even in the 80's our clocks and buses ran faster than
>that. But most failures related to rough flights or truck rides,
>something that soldered solutions weathered flawlessly.
---
If you had failures because of rough flights or truck rides, and those
failures were attributable to the wire-wrapped connections, then they
were improperly made, period.
---
>>Of course forethought in part placement and wiring technique are
>>important, just as they are in PCB layout, but for some reason many
>>people seem to believe that just because they have a wirewrap gun in
>>their hand, that's an invitation to sloppiness. Then, when they
>>mistakenly take the invitation and turn out a piece of crap they blame
>>wirewrapping instead of themselves for the disaster.
>>
>You are right but the techs we worked with were well trained in
>wirewrap.
---
Well trained, perhaps. Properly trained? Not if they caused the
kinds of problems you say came up.
---
>Actually we didn't let anyone else near the wrap guns because
>they were so expensive. They didn't worry about me as they knew I
>wouldn't touch them anyways :-)
---
And yet you say you did wirewrapping when you were helping to debug?
Hmmm...
---
>> Your overall
>>blanket condemnation of wirewrapping seems to put you in that
>>category.
>>
>>
>>
>I don't think so.
---
Think what you like, your admitted inability to successfully use or
supervise the use of a (even then) mature, well proven, and reliable
technology speaks for itself.
--
John Fields
Joerg
>Is that why IBM /360 and /370 mainframes were
>wire-wrapped? ...because it is unreliable? Sheesh!
>
>
They were indeed quite good for their days. But have you seen a WW
backplane after 10 years in a heavy industry location? As long as we
need steel products, electricity and so on we will have coal plants.
Therefore, we must accept the fact that air in some regions will contain
sulfur which causes silver sulfide and other problems in WW. I haven't
seen an IBM but several PDP-11 in these situations and it wasn't pretty.
Ok, you don't have to listen to me and I don't have time to dig all this
out again but here is one paper by Brent Sorensen. Check for "wire-wrap"
about half way down. Only one paragraph but it sums it up and I guess we
cannot say that he isn't an expert here.
http://www.evaluationengineering.com/archive/articles/0604/0604modern_electronics.asp
Regards, Joerg
Joerg
>... In practice you rarely had more than 1 wire in a contact because unlike wire wrap, you could continue on to the next connection point with the same wire. ...
>
Yes, in that case IDC can be quite reliable. But with prototyping this
became difficult. Oh, I forgot that I also need the strobe signal over
there...
Double placement is a concern and telecoms usually do not allow it. Even
the manufacturers of the 66 and 110 blocks strongly discourage that.
>... One thing that tended to help keep the wires in the contacts was that we would install a clear plastic shield on the wire side of boards once prototyping was finished. I've also seen the plastic shields used with wire wrap boards to minimize the possibility of bending the wrap tails and creating shorts. ...
>
>
Another device we used with IDC where combs made out of soft plastic.
They had barbs at the end of the fingers so wires would have less of a
chance to slip out. These comb rows were affixed to the board before
starting the wiring job. This also created a very organized appearance.
The IDC technology didn't last long at the companies I worked for. Wire
wrap kind of vanished towards the end of the 80's but had been used a
long time by then. IDC was often tried out as "the new thing" but I have
only seen it for a couple years or so. Then it all went to direct layout.
Regards, Joerg
John Larkin
AMP had a system that used a big NC machine that used reel-fed metal
clips that were shot over stranded wire onto rectangular posts, up to
three or maybe four stacked clips per post. It was called TermiPoint
or something. I wrote a lot of the netlist-sorting and coordinate
conversion software that generated the paper tapes the beast read...
the old salesman's route optimization and all that. We used it on a
couple of marine automation systems, but it was a real dog, and we
went to real semi-automatic wire-wrap. Later on we got smart, invented
a bus system, and went to PCB backplanes.
The GE marine/industrial systems were wrapped in the 1970 timeframe -
huge 45-mil posts and fat wire - and seemed OK.
John
John Fields
<notthis...@removethispacbell.net> wrote:
>Hi Keith,
>
>>Is that why IBM /360 and /370 mainframes were
>>wire-wrapped? ...because it is unreliable? Sheesh!
>>
>>
>They were indeed quite good for their days. But have you seen a WW
>backplane after 10 years in a heavy industry location? As long as we
>need steel products, electricity and so on we will have coal plants.
>Therefore, we must accept the fact that air in some regions will contain
>sulfur which causes silver sulfide and other problems in WW. I haven't
>seen an IBM but several PDP-11 in these situations and it wasn't pretty.
---
Read about gas-tight connections below...
---
>Ok, you don't have to listen to me and I don't have time to dig all this
>out again but here is one paper by Brent Sorensen. Check for "wire-wrap"
>about half way down. Only one paragraph but it sums it up and I guess we
>cannot say that he isn't an expert here.
---
Yes, we can, because he's certainly no expert as far as wire-wrapping
goes.
In the first place, that "paper" was no more than an advertisement,
and that single paragraph doesn't even bother to mention (if he even
knew) that the wrap, done properly, with proper tools, results in the
displacement of the copper and silverplate at the corners of the post
so that there are four gas-tight connections made per turn of wire
wrapped. For a one inch MIL-SPEC (yes, MIL-SPEC) modified wrap on a
.025" square post, that'll result in about 40 gas-tight contacts per
wrap, which is precisely one of the reasons why wire-wrap _is_
reliable. Another reason is that a bit and sleeve which does a
'modified' wrap puts about a turn of _insulated_ wire at the bottom of
the wrap, eliminating fatigue failure of the conductor which would
otherwise be caused by flexure of the bare wire at the beginning of
the wrap in the presence of extremely severe and persistent vibration.
--
John Fields
K Williams
> Hi Keith,
>
>>Is that why IBM /360 and /370 mainframes were
>>wire-wrapped? ...because it is unreliable? Sheesh!
>>
>>
> They were indeed quite good for their days. But have you seen a WW
> backplane after 10 years in a heavy industry location?
Yes! WW is a *very* good system. It is just as reliable as PWBs.
> As long as
> we need steel products, electricity and so on we will have coal
> plants. Therefore, we must accept the fact that air in some
> regions will contain sulfur which causes silver sulfide and other
> problems in WW.
You're simply *wrong*. WW done right is perfectly reliable, even in
areas I'm not.
> I haven't seen an IBM but several PDP-11 in these
> situations and it wasn't pretty.
You're colleagues are simple incompetent, apparently.
> Ok, you don't have to listen to me and I don't have time to dig
> all this out again but here is one paper by Brent Sorensen. Check
> for "wire-wrap" about half way down. Only one paragraph but it
> sums it up and I guess we cannot say that he isn't an expert here.
I'll trust my 30 years in the business, the first twenty happily
doing WireWrap. You're simply *wrong*. No, I wouldn't use it
today, simply because it's too expensive and printed-wiring is
cheaper. Reliability doesn't figure here.
http://www.evaluationengineering.com/archive/articles/0604/0604modern_electronics.asp
--
Keith
Tam/WB2TT
"K Williams" <k...@att.biz> wrote in message
news:3aadnX6d9fZ...@adelphia.com...
>
> I used wire-wrap for a couple of decades. It's fine for what it was
> intended. Wire-wrap was used extensively in mainframes in the
> '70s. Indeed, I've built entire test systems with wire-wrap. One
> simply has to be a tad careful. One (close to thirty years ago) had
> an 80MHz oscillator (quickly divided by two) and was done in
> MECL-10K. A well done wire-wrap board is a rather good 100ohm
> transmission environment. Decoupling and power-distribution was
> the area that got most people. Augat had some very nice WW
> prototype boards, with these miniscule clips to wire power/ground
> directly to the pins of interest (technicians hated them).
Sounds about the same kind of stuff we did. Our boards were custom, about
8.5 by 13 inches, with a 200 pin connector on one of the skinny ends. There
were rows of pins running the length of the board, and spaced vertically so
that you could use .3, .4, or .6 spacing. DIPs did not need sockets. Biggest
board I did had 126 DIPs. Usually we generated a wirelist off a Mentor
schematic, which was later also used to generate artwork. Eventually all
boards were machine wrapped. A pain with SM (mostly PLCC) and the required
adapters, which also had to be custom made. I think we gave up on wirewrap
when the time to get a 6 layer board came down from about 6 weeks to 6 days.
Most of the time difference was probably due to the fact that the bosses no
longer insisted on having our own factory make prototype boards.
Interesting what you said about MECL. A friend of mine had a MECL design,
and powered up a board with no bypass caps connected. It worked. He figured
that it was due to near constant current drain with no spikes. Somebody else
had a 7474 on a Vector board with 3 foot power/ground leads. It did not work
until he added a bypass cap.
Tam
Joerg
>>>Is that why IBM /360 and /370 mainframes were
>>>wire-wrapped? ...because it is unreliable? Sheesh!
>>>
>>>
>>>
>>>
>>They were indeed quite good for their days. But have you seen a WW
>>backplane after 10 years in a heavy industry location?
>>
>>
>
>Yes! WW is a *very* good system. It is just as reliable as PWBs.
>
>
Just why is it then that several agencies do not endorse it for printed
circuit boards? Maybe you can enlighten us here. All the WW boards I
have seen (not the back planes) were essentially printed circuit boards
with a ground and power plane. The socket corner pins were soldered to
the GND and power planes respectively and decoupling caps were on the
boards as well. To see an example of what agencies prescribe scroll to
"3.2.2.2 Wire Wrap" in this document:
http://www.faa.gov/and/and300/and360/coldfusion/library/specifications/FAA-E-2100g.pdf
>>I haven't seen an IBM but several PDP-11 in these
>>situations and it wasn't pretty.
>>
>>
>
>You're colleagues are simple incompetent, apparently.
>
>
The PDP-11 was made by DEC, not by my colleagues. What I was talking
about is contamination.
There is another issue that occasionally comes up with WW and this one
has nothing to do with the gas-tight connection and all. It is the fact
that you end up with lots of dense and tall posts that the wire is
wrapped around. That invites the accumulation of contaminants and I
remember a case (in the UK) where that was considered a factor in a
very dangerous situation in the air. Luckily the crew got the plane
somewhat stabilized and back to the ground. I believe the yaw damper
control is what had failed and the investigation concentrated on some of
the wire wrapped connectors.
With a circuit board a coating takes care of this but I have not seen
coating on a WW assembly.
Regards, Joerg
John Fields
<notthis...@removethispacbell.net> wrote:
>Just why is it then that several agencies do not endorse it for printed
>circuit boards? Maybe you can enlighten us here.
---
OK.
For several reasons, one of which is that ww boards are more
voluminous and weigh more than a PCB supporting the same component
population and function, another being that sockets, which are often
used on ww boards, are more or less verboten in a high-rel environment
because of galvanic corrosion problems and the likelihood of chips
falling out in a high-vibration environment.
---
>All the WW boards I
>have seen (not the back planes) were essentially printed circuit boards
>with a ground and power plane.
---
Not really. They were wire-wrap boards with ground and power planes.
Big difference.
---
>The socket corner pins were soldered to
>the GND and power planes respectively and decoupling caps were on the
>boards as well. To see an example of what agencies prescribe scroll to
>"3.2.2.2 Wire Wrap" in this document:
>
>http://www.faa.gov/and/and300/and360/coldfusion/library/specifications/FAA-E-2100g.pdf
---
They said "Wire wrap shall not be used on printed circuit boards." ,
which is a whole different thing from "Wire wrap boards shall not be
used."
---
>The PDP-11 was made by DEC, not by my colleagues. What I was talking
>about is contamination.
---
What you're talking about is something you obviously have very little
experience with but which, for the purpose of saving face, you have to
keep on prattling about.
---
>There is another issue that occasionally comes up with WW and this one
>has nothing to do with the gas-tight connection and all. It is the fact
>that you end up with lots of dense and tall posts that the wire is
>wrapped around. That invites the accumulation of contaminants and I
>remember a case (in the UK) where that was considered a factor in a
>very dangerous situation in the air. Luckily the crew got the plane
>somewhat stabilized and back to the ground. I believe the yaw damper
>control is what had failed and the investigation concentrated on some of
>the wire wrapped connectors.
---
And do you remember the outcome of the investigation, or does the
outcome of the investigation not matter as long as you can use the
investigation itself to conveniently imply guilt?
---
>With a circuit board a coating takes care of this but I have not seen
>coating on a WW assembly.
---
there are mechanical contrivances available which are designed to keep
the posts from touching each other, as is compliant foam.
--
John Fields
Guy Macon
Joerg <notthis...@removethispacbell.net> says...
>Just why is it then that several agencies do not endorse [wire wrap]
>for printed circuit boards?
I am trying to reconcile this with the fact that A Level 1 Avionics
Technician must have the "ability to use small hand tools and test
equipment such as wirewrap tools..."
K Williams
> Hi Keith,
>
>>>>Is that why IBM /360 and /370 mainframes were
>>>>wire-wrapped? ...because it is unreliable? Sheesh!
>>>>
>>>>
>>>>
>>>>
>>>They were indeed quite good for their days. But have you seen a
>>>WW backplane after 10 years in a heavy industry location?
>>>
>>>
>>
>>Yes! WW is a *very* good system. It is just as reliable as PWBs.
>>
>>
> Just why is it then that several agencies do not endorse it for
> printed circuit boards?
Perhaps they knew your competency level? The fact is that it *WAS*
used for high-rel circuits. ...yes even PCBs. IBM didn't use it
because it sucked. Of course people had the right tools and wire.
The most important tool was the stripper. Once dropped it was
junk. Techs protected their strippers (they looked like long-nosed
pliers) like they were their daughters. IIRC they wre about
$60/pr. Kynar wire was forbidden, as were auto-stripping bits,
though I let my techs use auto-strippers for my prototype stuff.
> Maybe you can enlighten us here. All the
> WW boards I have seen (not the back planes) were essentially
> printed circuit boards with a ground and power plane.
Sure. Power planes were necessary when you were still shiting
yellow! Are you trying to say you attempted WireWrap without solid
power planes? Amazing.
> The socket
> corner pins were soldered to the GND and power planes respectively
> and decoupling caps were on the boards as well. To see an example
> of what agencies prescribe scroll to "3.2.2.2 Wire Wrap" in this
> document:
Your people *were* incompetent. Look at a decent supplier of
wire-wrap components. Augat was my fav. They knew what it took to
do a decent product, though it was a PITA for the techs. ...and
expensive stuff (certainly by today's standards).
>
http://www.faa.gov/and/and300/and360/coldfusion/library/specifications/FAA-E-2100g.pdf
>
>>>I haven't seen an IBM but several PDP-11 in these
>>>situations and it wasn't pretty.
>>>
>>>
>>
>>You're colleagues are simple incompetent, apparently.
>>
>>
> The PDP-11 was made by DEC, not by my colleagues. What I was
> talking about is contamination.
I never worked for DEC, but I do know IBM never had such problems.
Like I said, it was used in all mainframes up until the 308x stuff,
when they went to twisted-pair welded transmission lines, on the
surface. You're simply wrong in your assertions.
> There is another issue that occasionally comes up with WW and this
> one has nothing to do with the gas-tight connection and all. It is
> the fact that you end up with lots of dense and tall posts that
> the wire is wrapped around. That invites the accumulation of
> contaminants and I
Hogwash! Youre assertions don't meet the smile test. Wire-wrap has
been used for *decades*, and in harsh climates, to boot. Your
assertions aside.
> remember a case (in the UK) where that was considered a factor in
> a very dangerous situation in the air. Luckily the crew got the
> plane somewhat stabilized and back to the ground. I believe the
> yaw damper control is what had failed and the investigation
> concentrated on some of the wire wrapped connectors.
Citation?
> With a circuit board a coating takes care of this but I have not
> seen coating on a WW assembly.
Good thing! Coatings are more problems than their worth, in most
circumstances.
You're showing your incompetence again.
--
Keith
K Williams
> On Sun, 27 Jun 2004 20:53:22 GMT, Joerg
> <notthis...@removethispacbell.net> wrote:
>
>
>>Just why is it then that several agencies do not endorse it for
>>printed circuit boards? Maybe you can enlighten us here.
>
> ---
> OK.
>
> For several reasons, one of which is that ww boards are more
> voluminous and weigh more than a PCB supporting the same component
> population and function, another being that sockets, which are
> often used on ww boards, are more or less verboten in a high-rel
> environment because of galvanic corrosion problems and the
> likelihood of chips falling out in a high-vibration environment.
Have "you" used the machined Augat style pins? These things hold
like hell, though are very expensive. They make wonderful
prototyping boards for DIP stuff. Of course they're far less
useful today.
> ---
>
>
>>All the WW boards I
>>have seen (not the back planes) were essentially printed circuit
>>boards with a ground and power plane.
>
> ---
> Not really. They were wire-wrap boards with ground and power
> planes. Big difference.
Sure. I never understood people trying to prototype a high-speed
widget daisy-chaining power and ground on a Wire-wrap board,
because it was "only" a prototype. One must be even more cautious,
though it is possible. Power distribution is the biggest trap I've
seen engineers fall into. Why should a prototype be any lesser of
a problem?
> ---
>
>>The socket corner pins were soldered to
>>the GND and power planes respectively and decoupling caps were on
>>the boards as well. To see an example of what agencies prescribe
>>scroll to "3.2.2.2 Wire Wrap" in this document:
>>
>>http://www.faa.gov/and/and300/and360/coldfusion/library/specifications/FAA-E-2100g.pdf
>
> ---
> They said "Wire wrap shall not be used on printed circuit boards."
> , which is a whole different thing from "Wire wrap boards shall
> not be used."
Dunno. I've seen both used on the same "board". Maybe we have a
difference of opinion of what constitutes a "board". IBM
Mainframes used both on the "mother-boards". The cards (plug into
boards) were all printed, with soldered over-flows, mainly because
there was no space for WW pins. IBM mainframes aren't generally
known to be unreliable.
> ---
>
>>The PDP-11 was made by DEC, not by my colleagues. What I was
>>talking about is contamination.
>
> ---
> What you're talking about is something you obviously have very
> little experience with but which, for the purpose of saving face,
> you have to keep on prattling about.
Indeed.
>
>>There is another issue that occasionally comes up with WW and this
>>one has nothing to do with the gas-tight connection and all. It is
>>the fact that you end up with lots of dense and tall posts that
>>the wire is wrapped around. That invites the accumulation of
>>contaminants and I
>>remember a case (in the UK) where that was considered a factor in
>>a very dangerous situation in the air. Luckily the crew got the
>>plane somewhat stabilized and back to the ground. I believe the
>>yaw damper control is what had failed and the investigation
>>concentrated on some of the wire wrapped connectors.
>
> ---
> And do you remember the outcome of the investigation, or does the
> outcome of the investigation not matter as long as you can use the
> investigation itself to conveniently imply guilt?
By association...
>>With a circuit board a coating takes care of this but I have not
>>seen coating on a WW assembly.
>
> ---
> there are mechanical contrivances available which are designed to
> keep the posts from touching each other, as is compliant foam.
Coating PCBs isn't always a win either. Often one is better off
leaving things alone. That said, I can't imagine 2-level posts
finding their way over to meet each other, unless there is severe
trauma (that would likely kill a modern PCB).
--
Keith
K Williams
> On Thu, 24 Jun 2004 19:09:37 GMT, ni...@puntnl.niks (Nico Coesel)
> wrote:
>
>>Joerg <notthis...@removethispacbell.net> wrote:
>>
>>>Hi Dave,
>>>
>>>While you compare against prototype PCBs keep in mind the
>>>incredible amount of time some poor technician needs for wire
>>>wrapping. And the cost this adds.
>>>
>>>I have never, ever, allowed any of my designs to be wire wrapped.
>>>As a result my boards were done a lot faster than the wrapped
>>>ones. It is no fun waiting for the others who are still chasing
>>>crosstalk, loose wires and so on. I have seen engineers on the
>>>verge of bursting into tears because their wrapped designs didn't
>>>work, the deadline was just hours away and the boss nervously
>>>standing behind them.
>>
>>I've seen large wire-wrapped -production- boards (in a computer
>>tape-drive). I guess wire wrapping is in a way a true art. If done
>>well, wire wrapping is more reliable than soldering (say >100
>>years versus 30 years).
>
> Solder joints die in 30 years? I still use GR and Tek instruments
> that are 40 years old, and, if they break, it's not the soldering.
The 30-40YO Teks I worked on had some rather hefty ceramic barrier
strips for mounting the components. I do hope you've never used
anythign but silver-solder on them. The lugs will pull right out
of the barrier strip with anything else. They came with a small
spool of silver solder attached to the inside of the case.
--
Keith (a one-time technician who worked on Tek and HP scopes in
college, thirty+ years ago)
K Williams
>
> "K Williams" <k...@att.biz> wrote in message
> news:3aadnX6d9fZ...@adelphia.com...
>>
>> I used wire-wrap for a couple of decades. It's fine for what it
>> was
>> intended. Wire-wrap was used extensively in mainframes in the
>> '70s. Indeed, I've built entire test systems with wire-wrap.
>> One simply has to be a tad careful. One (close to thirty years
>> ago) had an 80MHz oscillator (quickly divided by two) and was
>> done in
>> MECL-10K. A well done wire-wrap board is a rather good 100ohm
>> transmission environment. Decoupling and power-distribution was
>> the area that got most people. Augat had some very nice WW
>> prototype boards, with these miniscule clips to wire power/ground
>> directly to the pins of interest (technicians hated them).
>
> Sounds about the same kind of stuff we did. Our boards were
> custom, about 8.5 by 13 inches, with a 200 pin connector on one of
> the skinny ends. There were rows of pins running the length of the
> board, and spaced vertically so that you could use .3, .4, or .6
> spacing. DIPs did not need sockets. Biggest board I did had 126
> DIPs.
Sounds about right, though they came in many forms. One variety I
was fond of came in .300" stripes, another in stripes to
accommodate any DIP package, as you suggest. Others were set up
for 16-pinxx.300" DIPS. I had Augat make some custom boards for me
at one point (which they added to their product line).
> Usually we generated a wirelist off a Mentor schematic,
> which was later also used to generate artwork. Eventually all
> boards were machine wrapped. A pain with SM (mostly PLCC) and the
> required adapters, which also had to be custom made.
I left that area when I could still get away with DIPS ('88?). I
still used WW for prototyping for a short time after. PWBs just
got too cheap though.
> I think we
> gave up on wirewrap when the time to get a 6 layer board came down
> from about 6 weeks to 6 days. Most of the time difference was
> probably due to the fact that the bosses no longer insisted on
> having our own factory make prototype boards.
Well, many of my WW boards took more than six days to wire. IIRC my
most dense had 6K wires. ...drove my tech batty (his wife liked
the OT checks though;). He convinced my boss to sent the other
20ish out to a PC house. ;-)
> Interesting what you said about MECL. A friend of mine had a MECL
> design, and powered up a board with no bypass caps connected. It
> worked. He figured that it was due to near constant current drain
> with no spikes. Somebody else had a 7474 on a Vector board with 3
> foot power/ground leads. It did not work until he added a bypass
> cap.
Well... I coulda predicted the same. The early IBM '370s were all
ECL designs. The 308x processors were TTL (no not 74xx). The
thought was the power-savings would offset the delta-I on the power
busses. Go figure, the 309x series went back to ECL. ;-)
Of course CMOS rules the world, but not because it was faster (it
wasn't), or quieter (it isn't).
--
Keith
Joerg
>>Just why is it then that several agencies do not endorse it for printed
>>circuit boards? Maybe you can enlighten us here.
>>
>>
>
>---
>OK.
>
>For several reasons, one of which is that ww boards are more
>voluminous and weigh more than a PCB supporting the same component
>population and function, another being that sockets, which are often
>used on ww boards, are more or less verboten in a high-rel environment
>because of galvanic corrosion problems and the likelihood of chips
>falling out in a high-vibration environment.
>
>
Thanks for explaining. The weight and volume shouldn't be a reason for a
regulation as it is a feature of a certain product and the customer
should be the one to decide whether it fits or not. But the IC socket
issue thoroughly makes sense.
>>All the WW boards I have seen (not the back planes) were essentially printed circuit boards with a ground and power plane.
>>
>>
>
>---
>Not really. They were wire-wrap boards with ground and power planes.
>Big difference.
>
>
Well, they had bypass caps soldered onto them. But if I understand right
these still won't count as printed circuit boards under the regs then.
>>The PDP-11 was made by DEC, not by my colleagues. What I was talking
>>about is contamination.
>>
>>
>
>What you're talking about is something you obviously have very little
>experience with but which, for the purpose of saving face, you have to
>keep on prattling about.
>
>
No. And no need to save face or anything since WW is now history in my
field of work. Just wanted to clear up a misundertanding or two. What I
meant was that this machine had been professionally wire wrapped by DEC.
Yet contamination took its toll. Coating is, as Keith said in another
post, controversial. It may not always be a good thing and can cause
other trouble, for example thermal issues.
But I know that it is often used on printed circuit boards and
backplanes in harsh environments. Not just aerospace but also on oil
rigs. I had to repair some stuff there out at sea, so you can believe me
on this one. Sometimes you even have to meticulously log when and why
you puncture the coating for a measurement, then document the re-sealing
and so on. Actually, I wasn't allowed to do it myself because it is a
very controlled procedure and the "punctuators" often have to be
certified to do that.
>There is another issue that occasionally comes up with WW and this one
>has nothing to do with the gas-tight connection and all. It is the fact
>that you end up with lots of dense and tall posts that the wire is
>wrapped around. That invites the accumulation of contaminants and I
>remember a case (in the UK) where that was considered a factor in a
>very dangerous situation in the air. Luckily the crew got the plane
>somewhat stabilized and back to the ground. I believe the yaw damper
>control is what had failed and the investigation concentrated on some of
>the wire wrapped connectors.
>
>
>
>---
>And do you remember the outcome of the investigation, or does the
>outcome of the investigation not matter as long as you can use the
>investigation itself to conveniently imply guilt?
>
>
It's been a while now but the outcome was not a total scolding for WW in
itself. It was slow fluid ingress which started a chemical process. So
the fluid ingress was the precipitating cause or whatever expression
they use for this. The WW was only contributing because these areas were
not coated. So the fluid was able to get to the pins and start its
corrosion process until the sub-system failed in flight.
>>With a circuit board a coating takes care of this but I have not seen
>>coating on a WW assembly.
>>
>>
>
>---
>there are mechanical contrivances available which are designed to keep
>the posts from touching each other, as is compliant foam.
>
>
Haven't seen any of these but now I learned something new. Thanks. But I
have never seen a case where posts touched unless they got hit by
something, and damage like that cannot be blamed on WW. What I saw a lot
was deposits collecting way down in between the wrap posts and this was
hard to clean out.
Regards, Joerg
Joerg
>>Just why is it then that several agencies do not endorse [wire wrap]
>>for printed circuit boards?
>>
>>
>
>I am trying to reconcile this with the fact that A Level 1 Avionics
>Technician must have the "ability to use small hand tools and test
>equipment such as wirewrap tools..."
>
>
Probably for two reasons. They only ban WW from printed circuit boards
and not wiring panels, connectors and so on. Then, aircraft often fly
for decades and they must be able to service older ones as well as new
ones.
Regards, Joerg
John Fields
<notthis...@removethispacbell.net> wrote:
>Thanks for explaining. The weight and volume shouldn't be a reason for a
>regulation as it is a feature of a certain product and the customer
>should be the one to decide whether it fits or not. But the IC socket
>issue thoroughly makes sense.
---
Your question was why some agencies disapprove of wire-wrapped boards,
and the example you gave was use of wire-wrapped boards in aircraft.
Certainly, doing an apples-to-apples comparison between a wire-wrapped
board and a PCB will yield that the PCB will always win, hands down,
as far as volume and weight are concerned for the same component
population and function. Reliability is another matter, but
unimportant since the prohibition of use because of the size and
weight issues should preclude the use of wire-wrapped boards.
Besides, no sane vendor would offer open wire-wrapped boards for FAA
approval and expect them to be declared airworthy. Potted, perhaps,
if there was some compelling reason to wire-wrap the circuit and then
pot it, but I doubt it.
---
>>>All the WW boards I have seen (not the back planes) were essentially printed circuit boards with a ground and power plane.
>>>
>>>
>>
>>---
>>Not really. They were wire-wrap boards with ground and power planes.
>>Big difference.
>>
>>
>Well, they had bypass caps soldered onto them. But if I understand right
>these still won't count as printed circuit boards under the regs then.
>
>>>The PDP-11 was made by DEC, not by my colleagues. What I was talking
>>>about is contamination.
>>>
>>>
>>
>>What you're talking about is something you obviously have very little
>>experience with but which, for the purpose of saving face, you have to
>>keep on prattling about.
>>
>>
>No. And no need to save face or anything since WW is now history in my
>field of work.
---
Fine, but most of what you've said so far about wire-wrapping shows
that you were inexperienced with it when it wasn't history, that the
position you're taking is that you were, and that now you're having to
defend that stance in order to keep from losing credibility. No big
deal, that's just how it comes across to me.
--
John Fields
Joerg
>>>>The PDP-11 was made by DEC, not by my colleagues. What I was talking
>>>>about is contamination.
>>>>
>>>>
>>>>
>>>>
>>>What you're talking about is something you obviously have very little
>>>experience with but which, for the purpose of saving face, you have to
>>>keep on prattling about.
>>>
>>>
>>>
>>>
>>No. And no need to save face or anything since WW is now history in my
>>field of work.
>>
>>
>
>Fine, but most of what you've said so far about wire-wrapping shows
>that you were inexperienced with it when it wasn't history, that the
>position you're taking is that you were, and that now you're having to
>defend that stance in order to keep from losing credibility. No big
>deal, that's just how it comes across to me.
>
>
Apologies if it came across that way. So, for the record, I am not an
expert in WW. I am just someone who was exposed to gear that had WW in
it or people who used WW. And their experience level varied, of course.
There is an old saying: A technology is only as good as the customer
says it is. So in your case they are happy with WW and that is great. I
just wanted to say that I have seen cases where that wasn't so.
Please hold the tomatoes, but let me share one more: This is more than
20 years ago, an old computer, forgot which brand but it was one of the
big ones. So it can't have been built by non-experts I believe. It was
highly erratic but they didn't want to invest anymore into it, basically
junk it. PCs were not commonplace so the students there were crying foul
because they loved their old mainframe. While some were watching in
disgust I soldered the whole WW plane and, bingo, it worked again.
It's similar to Press-fit. I won't go into detail because that's really
OT here but some of us were complaining about it and basically told that
this is proven technology, not our turf and we should get on with life.
Then one day I kind of got loud because it affected my schedule. They
agreed to send a dozen "dead" press-fit motherboards over the wave
solder. All of them worked happily ever after :-)
Regards, Joerg
Joerg
>>remember a case (in the UK) where that was considered a factor in
>>a very dangerous situation in the air. Luckily the crew got the
>>plane somewhat stabilized and back to the ground. I believe the
>>yaw damper control is what had failed and the investigation
>>concentrated on some of the wire wrapped connectors.
>>
>>
>
>Citation?
>
>
Sorry, I did not keep the paper. But I'd guess there has to be a report
with the British equivalent of the FAA.
>>With a circuit board a coating takes care of this but I have not
>>seen coating on a WW assembly.
>>
>>
>
>Good thing! Coatings are more problems than their worth, in most
>circumstances.
>
>You're showing your incompetence again.
>
>
So you think conformal coating shows an engineer's incompetence? Well, I
guess you should never set foot on an oil rig then. And you may want to
build your own airliner before you take the next business trip.
Regards, Joerg
John Fields
>John Fields wrote:
>
>> On Sun, 27 Jun 2004 20:53:22 GMT, Joerg
>> <notthis...@removethispacbell.net> wrote:
>>
>>
>>>Just why is it then that several agencies do not endorse it for
>>>printed circuit boards? Maybe you can enlighten us here.
>>
>> ---
>> OK.
>>
>> For several reasons, one of which is that ww boards are more
>> voluminous and weigh more than a PCB supporting the same component
>> population and function, another being that sockets, which are
>> often used on ww boards, are more or less verboten in a high-rel
>> environment because of galvanic corrosion problems and the
>> likelihood of chips falling out in a high-vibration environment.
>
>Have "you" used the machined Augat style pins? These things hold
>like hell, though are very expensive. They make wonderful
>prototyping boards for DIP stuff. Of course they're far less
>useful today.
---
Yeah, sure!
I've got a couple of scans of a wirewrap board specifically designed
for ECL, and I'll post them if you want to see them, but be aware that
they're huge. About 11 meg each (but a lot of nice detail, including
the wire-wrapped twisted pair on the pin side and Fairchild's
ceramic-packaged 100K ECL on the component side. I'll post them to
abse if you want to see them.
---
>Sure. I never understood people trying to prototype a high-speed
>widget daisy-chaining power and ground on a Wire-wrap board,
>because it was "only" a prototype. One must be even more cautious,
>though it is possible. Power distribution is the biggest trap I've
>seen engineers fall into. Why should a prototype be any lesser of
>a problem?
---
I agree. My experience seems to be pretty much like yours, in that if
you're not _very_ careful with high speed breadboards and prototypes,
you're going to get bit and you won't even know what's biting you!
---
>Coating PCBs isn't always a win either. Often one is better off
>leaving things alone. That said, I can't imagine 2-level posts
>finding their way over to meet each other, unless there is severe
>trauma (that would likely kill a modern PCB).
---
Some "techs" have heavy hands...
--
John Fields
Steve Taylor
>>
>> Citation?
>>
> Sorry, I did not keep the paper. But I'd guess there has to be a report
> with the British equivalent of the FAA.
That would be the Civil Aviation Authority (CAA)
http://www.caa.co.uk/
Steve
Joerg
> That would be the Civil Aviation Authority (CAA)
> http://www.caa.co.uk/
Thank you for the information. I tried a search on this site but it only
came up with two airworthiness docs which contain a change from a WW to
a soldered mother board for the spoiler control module. This is one of them:
http://www.caa.co.uk/srg/airworthiness/aan/AANDocs/23398/23398000000.pdf
The documents don't say anything about the reason though. I am not sure
whether they scanned in older reports like electronics manufacturers did
with datasheet of vintage CMOS chips that are still being produced.
Regards, Joerg