Wire-Wrap Photos
I finished wiring the audio circuitry for BMOW yesterday, which marks the end of wiring for BMOW… for version 1 at least! I took the opportunity to reflect on my wire-wrapping odyssey, and took a bazillion photos of the namesake Big Mess o’ Wires before it disappears forever into an opaque case. It’s really quite a sight. There are 1253 wires with 2506 individually-wrapped connections, and I’ve got every one logged in a spreadsheet.
At first I found the slow pace of wire-wrapping to be a chore, but eventually I came to enjoy it. Once I get going, I can wrap about 25 wires in an hour. It’s almost like a form of meditation. Despite how long it takes to wrap, the wire-wrapping hasn’t really impacted my overall rate of progress. Design, debugging, and general procrastination consume the most time, and wrapping is just a small part of the total.
All this craziness is built on a 12×7 inch Augat wire-wrap board with 2832 gold wire-wrap posts, which I purchased from eBay for $50. I started wondering about all that gold… are those posts solid gold, or gold-plated? I broke one once, and its cross-section appears to be solid gold. I know the boards sold for astronomical prices when new, and the name Augat suggests the atomic symbol for gold. The board was even listed in the precious metals section of eBay, rather than anywhere electronics-related. With a postage scale and a little figuring, I estimated the total weight of all the pins at about 10 ounces. Assuming it’s solid 10 karat gold, that’s 4.1 ounces of pure gold, with a value of about $3700 at today’s gold price. Too good to be true?
In the photos, I tried to capture the feeling of order amid chaos. Click any image to see the full-size version. What I found remarkable is just how dense the wiring is in many parts. The wide shots show off the surface wiring, but the close-ups really get into some interesting rats’ nests. In the densest channels, those wires are stacked at least 10 deep! It’s hard to believe it actually implements a computer.
Behold the insanity that is Big Mess o’ Wires!
34 Comments so far
Leave a reply. For customer support issues, please use the Customer Support link instead of writing comments.
I can pretty much guarantee that the pins are copper, plated with a very thin layer of gold to prevent oxidation. Wire-wrap is very reliable. I’ve worked with 25 year old boards that still work fine, so BMOW should be good until at least 2034!
You’ve done a very neat job with the wrapping. I’m impressed!
Merlin
Very tidy, I take my hat off to you. 🙂 It looks like about fifty hours just to do the wrapping? I must admit to cheating a little for my project and using a company to produce the PCB from my layout. The IC placement and soldering however did take a couple of days because it was about 2000 pins with a similar number of traces totalling about 5000cm in length.
Wow, enough wire to reach the moon I think 🙂
Steve, If you were to start from scratch (not that I’m suggesting this, your layout looks great) would you change the method you used to route the wires?. Is it better to route directly from point A to B, or to keep everything in lanes?
I know about the cross talk that can happen with wires running in parallel, but I guess at a few Mhz this isn’t an issue.
Cheers
Dave
Haha, I’d argue the wiring is anything but tidy. I actually do route directly from point A to B, to minimize crosstalk, and also just because it’s easier. Where you see a pattern of many parallel wires in the photo, that’s because that *is* the shortest route from A to B. That happens more often than you might expect– any time two chips of the same type are in the same row, and are connected via a bus. I focused more on these areas for the photos because they’re more interesting to look at, but the top-right photo is the typical mess. Occasionally you’ll see a wire take an odd route, because I cut it a little too long and wanted to eat up extra slack, but otherwise it’s all A to B.
Amazing!
I would say off hand that you’ve got definite talent there.
Those boards look as if a pro did them.
[…] The BMOW runs at 2MHz and has 512K RAM and 16K ROM. It is constructed with primarily 7400 series logic and over 1048 wirewrap connections. […]
Some people say they “Built” a computer.
You can say “No, really, I BUILT a computer”
There are but a handful of people left in this world that could replicate this kind of work. Its hard to find someone that knows the hardware, the firmware, and the software.
I’m re-reading “Foundation” by Asimov, and you know, he may have been right. If he is, you sir qualify for a high priest! 😉
That is simply amazing. Wow.
[…] those who’ve asked for high-res photos, see this entry from February, and click any of the thumbnails to get the high-res versions of wire-wrapping […]
As an electronics engineer, you have my utmost respect. Have you thought about building a lower part count serial bit computer? A single section of the MPP Goodyear computer would be an interesting project if you could also write useful software for it. It makes you wonder how simple you can go and still be universal.
Thank you for taking me back! Much of my teen years were spent wirewrapping TTL and Z80 projects. I can still remember the pinouts to many of them. Sad to think how many of my neurons are wasted on that.
I used to cover the bottom of my boards with blank white mailing labels and poke the pins thru them. That way I could label the chips and denote pin 1. Working on the chips upside-down would confuse me otherwise. I saved up my allowance so that I could buy a 16Kx1bit DRAM each month.
A part of me died when surface mount chips came out.
pure gold pins would be very bendy.. it’s soft stuff.
Usually pins are made from gold plated brass.
Looks like a CDC Cyber mainframe back-plane ca 1975!
They were just as messy, but boy oh boy were they fast
for their day!
Kudos
Peter
This is really quite amazing – at first I thought.. why? But then I realized just how cool it is to have built your own computer, from scratch (well, using IC’s – but still).
Very nice work, and very nice wiring.. just.. very nice =)
Not only is this an incredible project, the photos make great desktop backgrounds for EE students. Like me.
Awesome! I envy your patience. 😉 Very, very nice.
That takes me back, nice work.
Have you implemented the S100 buss on the edge connector ?
Just wanted you to know that the backplane of computers that were built by Sperry Univac Defense systems built in the 50s, 60s and 70s were all wire-wrapped. The reasoning for this was that for low volume machines, it wasn’t worth the cost of building a printed circuit backplane. Also, since logic density was low, most design fixes could be implemented on the backplane.
During those days logic switching speeds were slow, so crosstalk and noise were not an issue. As speeds increased, wirewrap just would not work, unless twisted pairs with a ground return were used. Also, the interconnect density with I.C.s with of several hundred pins just was not practical to wirewrap.
High accuracy circuit simulation has made it practical to do Printed Circuit (P.C.) layout first time with minimal rework
( soldering on external wires ).
I would say, that the mid 1980s to early 90s was when it was more practical to do the first desiging on p.c. boards versus wirewrap.
Gold plated pins were used because gold does not oxidize and results in very reliable, long term connections. The pin corners provide for very high pressure, gas tight interconnect.
Art Collins,
Former Engineer with GE Healthcare Systems and Sperry Univac Defense Systems ( now Lockheed Martin )
Have you ever seen the plug board for an IBM 407, circa 1947? A fully populated board looks a lot like the BMOW. I take my hat, if I wore one, off to you. And thanks for sharing.
I used to work on Sperry Univac mainframe systems in the late 70’s which had wirewrap backplanes. There was a problem with the automated machine that did the wirewrapping, it would pull the wires too tight where it went around a pin. Eventually, the sharp corner of the pin could cut through the insulation causing intermittent faults – very difficult to trace!
Great piece of nostalgia!
Just to echo Pat Rooney’s comment: years ago, I worked on a UNIVAC 1108 mainframe; it was a 2 CPU system with 262K (36-bit) WORDS, or about a megabyte. It had the aforementioned backplane that was wire-wrapped, and we did experience the intermittent failure caused by the insulation failing on the occasional wire when it bent around another pin. After like a decade of being in service, just sitting there in the computer room with all the vibration from the AC and fans, etc. the insulation broke down. The actual wire-wrap connections are very reliable.
BTW the wire-wrap prototype board looks to be for an Intel Multibus, not the S-100 bus.
This is very nice work!
Very nice!
Watch out for bugs though.. 😉
Wow, looks like a fun week.
So what will you do with it?
If the pins really were solid gold my calculations show that the best plan would be to torch the board, cash-in the gold and immediately buy 50 or 60 more of the same boards, and an Apple III for your computing; repeating these steps until there are no more boards available – too bad they’re brass. The gold plating likely wouldn’t even buy an Apple III so you were better off building the computer.
Nice wrapping technique!
Ivan
Ow, what a sweet sight! My dad used to maintain systems with wirewrapped boards, and I actually learned how to wrap and re-wrap from him. I never used it in practice, as the soldering iron and PCB’s became the standard.
I saved your photo’s to use as a desktop wallpaper. For the moments I need some wrapping-Zen 🙂
Nicely done.
Meh-
As one of the few Main Distribution Frameworkers left at any Verizon office, I do that much in a day. Props
to your mad stacking & neatness skillz, however. Picture a rack of that 18ft high and…135ft long.
Partially wired up by outside guys who were in a hurry on a Sat…get the idea?
Are you using a hand-wrapper or an OK wrapgun?
Copper will never die; one day ya’ll will realize this. Underneath your soft-switches and Juniper fiber cabinets it
lurks, behind all your backbones and T-1’s; keeping all those new shiny cell-phones yakking along…
It’ll all still be there in a century. Slowly oxidizing, but still working.
A very nice job indeed. It takes a lot of commitment to build something that complex and debug it until it works. I am impressed, and give you my “Extreme Cleverness” award for this year. 🙂
Ah yes the joy of wrapping.. Nice work!
Thought I’d mention the magic-1 ( http://www.homebrewcpu.com/construction.htm ) also on the homebuilt webring..
Depending, the posts are gold plating (possibly relatively “thick”, but that’s still likely something like only 50 microinches) over either a brass alloy or mild steel. Pass a magnet near the pins and see whether you feel a tug. As you probably know, the 30ga wire wrap wire is silver plated. Carry on!
[…] of choice would leave you quite free to change your mind later. But, I could imagine reworking that Big Mess ‘o Wires would be a huge pain. So, I probably wouldn’t use a BMOW as my daily […]
What’s up, its fastidious post regarding media print,
we all be aware of media is a enormous source of facts.
Here is my site Buy Cash Mojo
Hey there…any chance I could use some of the photo’s you took?
These pins are not pure copper, too soft. They do have copper in them but in an alloy. I use to wirewrap serial number, model & speed on the big box tape drives made by Storage Technology. What was interesting is like handwriting, I could tell who wrapped the motherboard just by looking at the wraps. As to Kevin’s request for pictures, I have quite a collection of wirewrap boards just for the prototyping I like to do. I will be recovering from surgery, so it might be 3-4 months before I can get to them & photograph them. There’s nothing proprietary on them so copyright is not an issue.
Damn! This is just amazing.