Connect with us

Wire wrap

Discussion in 'Electronic Design' started by Dave Boland, Jun 24, 2004.

Scroll to continue with content
  1. K Williams

    K Williams Guest

    Is that why IBM /360 and /370 mainframes were
    wire-wrapped? ...because it is unreliable? Sheesh!
    You were surrounded by incompetence, obviously.
    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.
    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.
    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).
     
  2. Joerg

    Joerg Guest

    Hi James
    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.
    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
     
  3. John Larkin

    John Larkin Guest


    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
     
  4. John Fields

    John Fields Guest

    ---
    Read about gas-tight connections below...
    ---
    ---
    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.
     
  5. Guy Macon

    Guy Macon Guest


    I can confirm from personal experience that there are *no*
    reliability problems with wire wrap technology when the rules
    are followed. Yes, you can make bad wraps, but you can make
    bad PWBs as well - any technology can be misapplied.

    They tested wire wrap technology very thoroughly before
    approving it for MIL-STD usage.
     
  6. K Williams

    K Williams Guest

    Yes! WW is a *very* good system. It is just as reliable as PWBs.
    You're simply *wrong*. WW done right is perfectly reliable, even in
    areas I'm not.
    You're colleagues are simple incompetent, apparently.
    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
     
  7. Tam/WB2TT

    Tam/WB2TT Guest

    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
     
  8. Joerg

    Joerg Guest

    Hi Keith,
    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
    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
     
  9. John Fields

    John Fields Guest

    ---
    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.
    ---

    ---
    Not really. They were wire-wrap boards with ground and power planes.
    Big difference.
    ---
    ---
    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."
    ---
    ---
    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.
     
  10. Guy Macon

    Guy Macon Guest

    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..."
     
  11. K Williams

    K Williams Guest

    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.
    Sure. Power planes were necessary when you were still shiting
    yellow! Are you trying to say you attempted WireWrap without solid
    power planes? Amazing.
    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).
    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.
    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.
    Good thing! Coatings are more problems than their worth, in most
    circumstances.

    You're showing your incompetence again.
     
  12. K Williams

    K Williams Guest

    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.
    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?

    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.
    By association...


    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).
     
  13. K Williams

    K Williams Guest

    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.
     
  14. K Williams

    K Williams Guest

    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).
    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.
    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. ;-)
    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).
     
  15. Joerg

    Joerg Guest

    Hi John,
    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.
    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.
    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.
    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.
    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
     
  16. Joerg

    Joerg Guest

    Hi Guy,
    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
     
  17. John Fields

    John Fields Guest

    ---
    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.
    ---
    ---
    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.
     
  18. Joerg

    Joerg Guest

    Hi John,
    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
     
  19. Joerg

    Joerg Guest

    Hi Keith,
    Sorry, I did not keep the paper. But I'd guess there has to be a report
    with the British equivalent of the FAA.
    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
     
  20. John Fields

    John Fields Guest

    ---
    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.
    ---
     
Ask a Question
Want to reply to this thread or ask your own question?
You'll need to choose a username for the site, which only take a couple of moments (here). After that, you can post your question and our members will help you out.
Electronics Point Logo
Continue to site
Quote of the day

-