bwc@ganehd.UUCP (Brantley Coile) (05/02/85)
I want to build some one-of-a-kind stuff to try out ideas on parallel systems. My first question is how should I fabricate the equipment. While this may be an easy question to answer for production design, it is not so easy for a OOAK system. My choices are: 1) wire wrap 2) make a pc board I don't like (2) because of the cost and effort. Doing the artwork and having made sounds time consuming and expensive. It does give me very reliable stuff. It also make debugging impossible. I don't like (1) because of reliability. I understand that as wire wrap ages it develops the flakies. This leaves me with - 3) point to point I could do (3) if I used a solderless circuit board to prototype and do point to point when the board was debugged. This led me to another problem; I would need one heck of a solderless circuit board to debug stuff. Why? Why do boards have to be so big? If I were going to build a million of them I could see the benefit of reduced chip count that could come from a large board. I assume that there could also be a lot of savings in other things with a larger board size. But I don't care about any of them. If system design is similar for hardware as software, the smaller, functional boards would be better. I am not saying that the boards should be interfaced with some bus definition like STD, S-100, MultiBus, etc. Each board would be more of a module than a board, each with inputs and output. If I did this I could debug them one the solderless circuit board and when working make the point- to-point board. Well, what do you think. As you might have guessed I'm not a hardware engineer although I have built a small micro. Please send me you thoughts on the matter. -- Brantley Coile CCP ..!akgua!ganehd!bwc Northeast Health District, Athens, Ga
karsh@geowhiz.UUCP (Bruce Karsh) (05/08/85)
> > I want to build some one-of-a-kind stuff to try out ideas on > parallel systems. My first question is how should I > fabricate the equipment. While this may be an easy question > to answer for production design, it is not so easy for a > OOAK system. > > My choices are: > > 1) wire wrap > > 2) make a pc board > > I don't like (1) because of reliability. I understand that > as wire wrap ages it develops the flakies. I disagree. Wirewrap is more reliable than hand soldered pc boards. We have instruments which we use in our field operations that have both pc and wirewrap boards in them. These boards are subject to a lot of vibration and temperature cycling. The wirewrap boards are much better than the pc boards. The wirewrap boards are also more dense than the pc boards. The reason that wirewrap is so reliable is that: 1) The wire is held under tension against the post. That keeps it from falling off due to thermal expansion/ contraction. 2) The corners of the posts are sharp and form an airtight connection with the wire. 3) Since there are 4 corners on each wirewrap post, and about 8 wraps, there are 32 connections per post. Thats a awful lot of redundancy. It is very unlikely to fail. For one of a kind projects and small production runs, wirewrap is hard to beat. A good reference is a publication called Wire-Wrap from The Cooper Groop, PO Box 728 APex NC 27502. (919)362-7510. Catalog No. 55548. This goes into the materials science aspects of wire wrap. Of course, if you don't do the wraps properly, it won't work. But it's pretty easy to do right, and a proper wrap has about 100,000 lbs/sq in of contact force. Wire wrap is covered under mil-spec MIL-STD 1130. -- Bruce Karsh | U. Wisc. Dept. Geology and Geophysics | 1215 W Dayton, Madison, WI 53706 | This space for rent. (608) 262-1697 | {ihnp4,seismo}!uwvax!geowhiz!karsh |
luscher@nic_vax.UUCP (05/08/85)
> > I want to build some one-of-a-kind stuff to try out ideas. > How should I fabricate the equipment? > > My choices are: > 1) wire wrap > 2) make a pc board > 3) point to point (solderless breadboard) > My additional choices are: 4) point to point, directly soldering wires to either IC legs (tacky but cuts down cost), or to sockets. 5) insulation displacement I have used #4 at home and when soldering to ICs or sockets I find it a nuisance to add a wire withour having a previously connected one pop off occasionally. Otherwise this is quick, dirty, cheap and works. I have used #5 at work, and am preparing to use it at home (S100 board with insulation displacement sockets from 3M, bought as a kit some time ago). These are *EXPENSIVE* (I'm cheap?) but I found them VERY easy and *QUICK* to set up and change, a valuable advantage for OOAK equipment designed for fooling around and cut-and-paste design. They can be damaged however if wires are pulled out of the bifurcated terminals towards the side instead of straight up. If this happens the terminals spread and the next wire into that terminal will not be gripped tightly enough. Also don't pull wires too tightly around corners on any method at they tend to short through insulation (very difficult to find!). GOOD LUCK! -- Jim Luscher / Nicolet Instruments / Oscilloscope Div. 5225 Verona Rd Bldg-2 / Madison Wi 53711 USA / 608/271-3333x2274
ron@brl-tgr.ARPA (Ron Natalie <ron>) (05/09/85)
> My choices are: > > 1) wire wrap > > 2) make a pc board > > I don't like (1) because of reliability. I understand that > as wire wrap ages it develops the flakies. > SCOCTHFLEX is nice. It's expensive, but cheaper than making PC boards for prototypes.
rbc@houxu.UUCP (R.CONNAGHAN) (05/11/85)
No one has mentioned Multiwire. Can be fabricated from schematics or a net list. Generally very fast turn-around compared to Printed Circuits. I have used most prototyping techniques, Wire-Wrap, Multiwire, PCB, and point to point. The favorite has gone from wire-wrap to Multiwire. A comment on wire-wrap robutness. It's not the wrapped connection reliability that causes problems. It't the intergratiy of the insulation. The insulation can often be nicked as it goes around a post. It can also were away if the board is flexed or moved alot. A wire-wrap system carefully built (automatically prefferred) and NOT MOVED or FLEXED much is a very reliable way to build. -- Robert Connaghan Microprocessor Group AT&T Bell Labs - Holmdel, N.J. houxu!rbc
jlw@ariel.UUCP (J.WOOD) (05/12/85)
For a short prototype run I like multi-wire. It's much more rugged than wirewrap, can be run off the wirewrap wiring list and component placement, and is multilayer; you can really pack it on the card. The components can be soldered on to the card, and changes can be made fairly easily. Joseph L. Wood, III AT&T Information Systems Laboratories, Holmdel (201) 834-3759 <ariel!>titania!jlw
jss@sjuvax.UUCP (J. Shapiro) (05/12/85)
Regardgin wire wrapping... > > 2) The corners of the posts are sharp and form an airtight > connection with the wire. > Nay - better than, this, the seal is basically a cold weld - damn unlikely that it will randomly remove itself. Things to beware of, however: At high speed, wire wrap temds to crap out. (i.e. > 8 Mhz or so). Then again, so do sockets. Beware of kinked wires - these *do* go with time, but are easy to spot once gone. In particular, the repair time given a known break is quite low. Also, get yourself an electric wire wrap gun - it will save many many hours of precious time, and will prevent a lot of bad wraps. I just completed a VAX I/O interface which was wirewrapped... I will be prototyping a small computer shortly using wirewrap. Jon Shapiro
caf@omen.UUCP (Chuck Forsberg WA7KGX) (05/13/85)
I recently developed a moderately large board (a laser printer controller with 1 MB ram, etc.) with the insulation displacement system. We had considerable trouble with the sockets coming apart after repeated insertions and withdrawl of the chips. As for my next project, I might use the method if the board was small, wasn't going to be used for very long, and I didn't expect to fuss with it too much before getting it to work. A prime advantage of the insulation displacement method is less board thickness compared to wirewrap. It's faster too, although a good stripper and/or prestripped wires make wirewrap much faster. With wirewrap, I find a manual tool ($5) better than the cheap electric types. -- Chuck Forsberg WA7KGX ..!tektronix!reed!omen!caf Omen Technology Inc 17505-V NW Sauvie IS RD Portland OR 97231 Voice: 503-621-3406 Modem: 503-621-3746 (Hit CR's for speed detect)
esco@ssc-vax.UUCP (Michael Esco) (05/14/85)
> Things to beware of, however: > > At high speed, wire wrap temds to crap out. > (i.e. > 8 Mhz or so). Then again, so do sockets. Not necessarily true. At NCR we used twisted pair wire wrap in a large, multiboard, 38-nanosecond processor. > Beware of kinked wires - these *do* go with time, but are easy > to spot once gone. In particular, the repair time given a known break > is quite low. > > Also, get yourself an electric wire wrap gun - it will save many many hours > of precious time, and will prevent a lot of bad wraps. > > Jon Shapiro Wire wrap is a perfectly good technique for limited-production devices. But it requires careful assembly with the proper tools. I'd venture to say that most complaints about wire wrap come from those who didn't put it together carefully or handled it roughly after assembly (No, you can't play frisbee with it). Michael Esco (formerly of) NCR E&M San Diego (but now at) Boeing Aerospace
b-davis@utah-cs.UUCP (Brad Davis) (05/16/85)
>> Things to beware of, however: >> >> At high speed, wire wrap temds to crap out. >> (i.e. > 8 Mhz or so). Then again, so do sockets. > >Not necessarily true. At NCR we used twisted pair wire wrap in a large, >multiboard, 38-nanosecond processor. > I once helped (did software) on a project that had 15 nanosecond signals running around on the board. The only twisted pairs ran off the board to a high resolution monitor. -- Brad Davis {ihnp4, decvax, seismo}!utah-cs!b-davis b-davis@utah-cs.ARPA
ron@brl-tgr.ARPA (Ron Natalie <ron>) (05/17/85)
> > >> Things to beware of, however: > >> > >> At high speed, wire wrap temds to crap out. > >> (i.e. > 8 Mhz or so). Then again, so do sockets. > > > >Not necessarily true. At NCR we used twisted pair wire wrap in a large, > >multiboard, 38-nanosecond processor. > > > I once helped (did software) on a project that had 15 nanosecond signals > running around on the board. The only twisted pairs ran off the board to > a high resolution monitor. > -- > The backplanes of the HEP, one of the worlds fastest computers with the worlds slowest I/O system are wire wrapped.
yrdbrd@bmcg.UUCP (Larry J. Huntley) (05/18/85)
In article <612@houxu.UUCP> rbc@houxu.UUCP (R.CONNAGHAN) writes: >No one has mentioned Multiwire. >Generally very fast turn-around compared to Printed Circuits. > >The favorite has gone from wire-wrap to Multiwire. > >A comment on wire-wrap robustness. >(etc.) > >Robert Connaghan >Microprocessor Group >AT&T Bell Labs - Holmdel, N.J. >houxu!rbc Although I can agree with Robert about Multiwire, I think that it is a little out of sync for the application the original poster had in mind (hobbyist or university lab "one-off" system, I think.) Multi- wire is great for something that you need 25 -- 1000 of. Computervision used the technique to build add-on boards in Data General computers, and they were at least as reliable as the multi-layer PCBs that were in the computers. It's a bit expensive for a single board, though. I have been using wire-wrap for 13 years now, and have never had any problems with well-planned, well-wrapped boards/systems. Some further comments: (1) The problems encountered with wire-wrap usually involve (1) the insulation failing where a wire passes around the corner of a socket/pin and cuts through the insulation, causing a short, or (2) a wire under so much tension that the inner conductor breaks (usually at the point where the wrap starts around the post) but the insulation stays intact, making the problem virtually impossible to find visually. (2) There are wire-wrap bits and sleeves which cut and strip the insulation as the wrap is being made. One of these pays for itself with the first board you build with more than 25 wires. (NO, this is NOT the "slit & wrap" tool discussed earlier. This is a genuine wire-wrap tool and it makes a proper 'modified' wrap on standard 0.025" sq. pins. If you need more information, let me know. I'll look it up.) (3) Be sure to use "levels" on your pins, and plan your physical layout to minimize wire lengths. Leveling the wires looks like this: Wiring Side | | | | | | <- pins |--------| |--------| |-------| <-- level 2 | | | | | | | |-----------| |---------| | <-- level 1 | | | | | | ------------------------------------------------------ <- board ------------------------------------------------------ | | | | | | | | | | | | <- sockets ------------ ------------ ----------- Component Side (4) Beware of stray inductance in wire-wrapped boards. If possible make your connections to power and ground with 20 ga. solid wire, soldered to busses on the board. If you must wrap the power leads, keep the wires as short as possible. A loop of 1" of 30 AWG wire can exhibit enough inductance to induce voltage spikes as high as 1 volt on the power leads of devices which are switching at 4 MHz. Wire-wrapping dynamic RAM arrays can be particularly exciting. Have great fun, 'brd -- Larry J. Huntley Burroughs -(B)- Corporation Advanced Systems Group MS-703 10850 Via Frontera San Diego, CA 92128 (619) 485-4544 -*- Non Circum Copulae -*-
kensmith@sunybcs.UUCP (Ken Smith) (05/20/85)
Well, since most of the net has seemed to suggest wire-wrapping you might like some pointers on how to go about it. First, make reasonable circuit schematics, and get a blueprint made. Using the blueprint start making connections using sockets and, for discrete parts get the pins that you insert into the board and solder the discretes onto later. (The only place these pins aren't necessary is when you have lots of resistors of the same value in generally the same area, in which case it is easier to just use another socket.) Keep going for as long as you can at one shot, this helps reduce the chances for errors. Keep a diagram that shows which parts are what (i.e. a layout diagram) so you can find them later. Labels also help (especially those labels for the back-side of sockets with the pin numbers, though these get expensive). Whether you start checking while constructing the circuit or after it is all connected is up to you, but do continuity checks on ALL connections, and (this is where the blue- print comes in) use one of those high-lighting markers to mark all the connections you have checked. The continuity checks are important, it's very possible to have breaks in the wires but the insulation is still intact so it LOOKS connected. A multi-meter that beeps for continuity checks is nice, as are clip probes and those headers you can stick in the socket that have pins on both sides. When done solder in the discretes and start to debug it... I'm still relatively new to wire-wrapping so there are probably many other things that make life easier that I haven't tripped over yet, but this should get anyone interested started on the right track. -- ----- Ken Smith - Don't rattle my bars!
arcorp@utcsri.UUCP (Alias Research Corporation) (05/20/85)
What is Scotchflex? Steve
jer@peora.UUCP (J. Eric Roskos) (05/20/85)
ron@brl-tgr.ARPA (Ron Natalie <ron> at Ballistic Research Lab) writes: > The backplanes of the HEP, one of the worlds fastest computers with the > worlds slowest I/O system are wire wrapped. For that matter, wire-wrapped backplanes were extremely common until fairly recently. Look at most of the old DEC (KL10, PDP8-L, etc) and NCR (100 family) machines, for example. These contain backplanes wire-wrapped many layers deep. However, the wire-wrap was with stripped wire. My experience with the "just wraps" style tools, particularly when used by people who were relatively inexperienced with them, has not been good. But... to raise a different question... why does the HEP have "the world's slowest I/O system"? -- Full-Name: J. Eric Roskos UUCP: ..!{decvax,ucbvax,ihnp4}!vax135!petsd!peora!jer US Mail: MS 795; Perkin-Elmer SDC; 2486 Sand Lake Road, Orlando, FL 32809-7642 "Vg'f whfg guvf yvggyr puebzvhz fjvgpu urer... lbh thlf ner FB fhcrefgvgvbhf!"
brad@gcc-bill.ARPA (Brad Parker) (05/21/85)
In article <1143@sjuvax.UUCP> jss@sjuvax.UUCP (J. Shapiro) writes: >Regardgin wire wrapping... >> 2) The corners of the posts are sharp and form an airtight >> connection with the wire. This is known as "gas tight". >Things to beware of, however: > > At high speed, wire wrap temds to crap out. > (i.e. > 8 Mhz or so). Then again, so do sockets. This is not true. Get a good Augat board - worth every penny in noise reduction. Hell, people do ECL with these boards... > Beware of kinked wires - these *do* go with time, but are easy > to spot once gone. In particular, the repair time given a known break > is quite low. > >Also, get yourself an electric wire wrap gun - it will save many many hours >of precious time, and will prevent a lot of bad wraps. Of dubious use for small fixes (1-2 wire). Very handy for major jobs. -- J Bradford Parker uucp: seismo!harvard!gcc-bill!brad "She said you know how to spell AUDACIOUSLY? I could tell I was in love... You want to go to heaven? or would you rather not be saved?" - Lloyd Coal
roy@phri.UUCP (Roy Smith) (05/22/85)
> [...] but do continuity checks on ALL connections, and (this is where > the blueprint comes in) use one of those high-lighting markers to mark > all the connections you have checked. As someone who has fab'ed a few small-scale wire-wrap boards, I heartily agree -- If you don't ohm out EVERY connection, you are asking for trouble. Now, a question. I've gotten sick of those "Just Wrap" tools you can buy at Radio Schlock or through the catalogs. Either my technique is the pits (possible) or the tools just plain don't work right. I find that my percentage of good wraps is only about 2 out of 3. Every time I need to do another board, I debate shelling out the bucks for one of those OK electric jobs. Is it worth it? Any recommendations of pre-cut, pre-stripped vs. continuous feed wire? -- allegra!phri!roy (Roy Smith) System Administrator, Public Health Research Institute
esco@ssc-vax.UUCP (Michael Esco) (05/23/85)
> > Well, since most of the net has seemed to suggest wire-wrapping > you might like some pointers on how to go about it. First, make reasonable > circuit schematics, and get a blueprint made. Using the blueprint start > making connections using sockets... > Ken Smith For all but the simplist systems, I would strongly suggest you prepare a wire list. This could save quite a bit of time trying to debug a flakey circuit. Michael Esco Boeing Aerospace
neal@weitek.UUCP (Neal Bedard) (05/24/85)
I assume we're talking TTL digital here. I have had experience with both Wire-Wrap (tm) and Scotchflex (tm) systems, and I will summarize my opinions of both here. Wire-wrap: Extremely tedious unless you have the right tools, most important of which are: 1. a *good* wire-wrap gun - that works with the gauge wire you're using. Though expensive (~ $160) my favorite is the NiCd-powered Gardener-Denver unit. 2. wire strippers that don't nick the wire. 3. an unwrapping tool. Nets should be split into two levels, alternating between them serially so that one does not have to undo more than two wraps to make a fix. Wiring should be done serially, as to avoid stubs. Wires should be routed so that no tension are placed on them to prevent the pins from cutting the insulation. Scotchflex: This is a system marketed by 3M that uses insulation displacement to make a gas-tight connection. You insert strips on the wiring side, and mating "sockets" on the other. Connecting wires are then punched down into the strip's "tangs" on the wiring side. Daisy-chaining bus wires is trivially easy, compared to Wire-Wrap. The tangs protrude about 1/8" on the wiring side, as opposed to 3/4" for Wire-Wrap pins. This allows Scotchflexed boards to be used in normal cardcages without sacrificing slots to allow for Wire-Wrap pin spacing. While quick and easy, this system is unreliable unless one takes care to route the wires at a right angle to the tangs. If this is not done, the wires break quite easily. I usually tie bundles of wires together to give them additional support. The plug sockets aren't the greatest, as they are tin-plated and tend to oxidize - no match for the machined contact sockets used in Wire-Wrap (3M, are you listening?) Probably more important is the choice of board. One should choose a board with the lowest gnd and Vcc impedance possible, as current dumping probably causes more proto boards to fail than any other cause. Augat panels (with pins already on them) are good, since they have internal power and ground planes, but are *very* expensive. I use bare boards from a local company called Twin Industries that feature gnd and Vcc interspersed between all holes, as well as decoupling at every IC location. Other vendors also make good bare boards. Wire-Wrap connections (and, I assume, Scotchflex) ring more than multilayer boards. The ringing problem shouldn't be an issue if one is careful with stubs and current dumping. Someone mentioned Multi-Wire (tm). Multi-Wire isn't quite appropriate for one-off systems due to it's cost and not so great turnaround (about 4 weeks is my experience.) It is tough to rework, also. Multi-Wire's key advantage is density over multilayer (but not over Wire-Wrap) since no feedthoughs are needed in Multi-Wire. -Neal B. Wire-Wrap is a trademark of Gardener-Denver Scotchflex is a trademark of 3M Co. Multi-Wire is a trademark of Kollmorgen Corp.
pauldan@hou2e.UUCP (P.SAUNDERS) (05/24/85)
<<>> > Every time I need to do another board, I debate shelling out the > bucks for one of those OK electric jobs. Is it worth it? Any > recommendations of pre-cut, pre-stripped vs. continuous feed wire? I have obtained terrible results from using the "Just Wrap" tool; mine was disposed of a few years ago. Seems that the wire would never cut in through the insulation consistantly; also, the stuff was more prone to coming through the insulation when the wire was bent around a pin (not a good thing to do anyway, but it happens). I've had best results using an OK-type hand-wrapping tool. I've found that the electric gun doesn't save as much time as one might think, and the wraps are usually better with the hand wrapper. The time taken to actually wrap the wire is maybe 10% of the time taken for each wrap, so the gun's speedup isn't very effective. Pre-cut, pre-stripped wire is a blessing, and is worth every penny. When not using this, I use wire in rolls, and a "clip-and-strip" is a *must* for this. Dan Masi
kensmith@sunybcs.UUCP (Ken Smith) (05/28/85)
> Every time I need to do another board, I debate shelling out the > bucks for one of those OK electric jobs. Is it worth it? Any > recommendations of pre-cut, pre-stripped vs. continuous feed wire? > -- > allegra!phri!roy (Roy Smith) > System Administrator, Public Health Research Institute I've never used the "Just-Wrap" stuff, but the people I work for say it's definitely not for novices; experienced people may even have trouble. They loan me one of the battery-powered OK guns and that really saves my sanity! They are really nice compared to hand-wrappers. Considering the time saved they are definitely worth the $45 - $50 (Digikey or Jameco prices). I don't use pre-cut because the boards I've been doing don't follow very regular patterns. I'd consider pre-cut for memory boards or the like. I use those plastic spool holders with the stripper and cutter built in. Just be careful how you grab the wire when you go to strip it; it's easy to bend and (cringe) break the wire. I grip the wire with needle-nose pliers to strip it which seems to work well. Typically I don't have trouble with broken wire, just connections I missed... -- Ken Smith UUCP : ..![bbnca,decvax,dual,rocksanne,watmath]!sunybcs!kensmith
karsh@geowhiz.UUCP (Bruce Karsh) (05/29/85)
In article <765@ssc-vax.UUCP> esco@ssc-vax.UUCP (Michael Esco) writes: > >For all but the simplist systems, I would strongly suggest you prepare a >wire list. This could save quite a bit of time trying to debug a flakey >circuit. > Does anybody have a good wire list generator that they could post to net/mod.sources? If nobody has a good one, I have a minimal one that I could post if there is any interest. Bruce Karsh -- Bruce Karsh | U. Wisc. Dept. Geology and Geophysics | 1215 W Dayton, Madison, WI 53706 | This space for rent. (608) 262-1697 | {ihnp4,seismo}!uwvax!geowhiz!karsh |
jack@boring.UUCP (05/29/85)
About tools: I find that it doesn't make much of a difference
whether you use an electric gun or a hand-operated one (as long
as it's good quality), only the electrical guns produce a wrap of
a more constant quality.
Since I seem to spend most of the time (50%??) cutting the wire
to the right length and stripping it, I tried a lot of different
tools for that. The one I like best is a very simple thing (costs
about $10, I guess), shaped like this:
|| \
==========================
cut strip
It has a very nice grip, strips to the right length, hardly ever fails
to strip the wire, and the same tool is usable with a wide variety
of wire gauges. I haven't got it near at hand, but I can lookup the
mfg/type, if anyone is interested.
Isn't it great that a $10 tool can outperform various $50-$200(really!)
competitors........
--
Jack Jansen, jack@mcvax.UUCP
The shell is my oyster.