wire wrap advice

Sanmina-SCI in Salem NH?

--
  Keith

hear hear!

Its amazing what a good production department can do with rework :)

a contract mfg we used in NH showed us a rework they did for one customer, who forgot a wire under a 512-pin BGA. So these guys (alas, cant remember name) whipped the BGA off, re-balled it, soldered a dangly wire onto the BGA then replaced it on the PCB. 100% success rate, almost

1000 boards. wow.

an OEM ups I got lumbered with had 90 rework mods. It took longer to rework the pcb than to assemble it in the first place. So we built

17,000 like that :) [long story why we didnt re-spin the pcb]

careful design + careful spice + careful layout = its close enough to working that we can make it go.

Cheers Terry

It still isn't worth it. Just use regular sockets, and solder wires to them. It's not quite as quick as wire-wrapping, but it's nowhere near as kloogey-looking. Can you even still get wire-wrap stuff?

Somebody was asking about soldering to a vectorboard. This isn't chips, but hopefully shows that it's doable:

Ew. In the close-up, it looks like pretty crappy work. But WTF, I got paid $150.00 for six of the little bastards, and the guy who ordered them was buying the drinks.

Why the big fat (#18) wire around the perimeter of the board? EMI protection. It's a motor speed controller for a spool gun for a battery-powered MIG welder; it mounts right alongside the ~60A weld current cable. I first did it with a 555, and it died instantly, so the primary consideration was robustness. The main current switch isn't shown

- it mounts to the heatsink that the little board gets bolted to. Its leads poke up through the three enlarged holes by the turret terminal at the bottom. It was also fun cutting notches in the heatsink fins with an endmill bit in a drill press, to accommodate the mounting screws. He also showed me how to use a tap in a hand drill.

Yes, it's a Radio Shack perfboard. :-)

Cheers! Rich

I once saw a "wire-wrap gun" that was just a hand wire-wrapper clamped in an electric eraser.

Cheers! Rich

The other end is an unwrapper, and actually, is a very convenient handle for spinning.

Cheers! Rich

Tom Woodrow

snipped-for-privacy@yahoo.com wrote:

Hello Rich,

Agree. Also, 10 years isn't a whole lot of life. Even 20 isn't for some stuff. When wire wrap has turned 30 and the posts look black, oh boy...

It can be as quick. When in a rush I use wire where the lacquer insulation melts while soldering. For very speedy work you have to learn how to simultaneously hold the spool and wire cutters in one hand, solder iron in the other. But that's no harder than learning to eat with chop sticks. Sometimes I clamp a contraption to the lamp above me and roll off the spool from there.

Regards, Joerg

I took one of these tools, a WSU-30M:

and ground off the rivet holding the little stripper blade, and clamped it in an X-acto handle. It was handier than a rat, especially for doing buses. I'd strip a couple of inches off the end of a spool of wrap wire, and solder just the tip end to the one point, then go to the insulation, and cut it at the length I needed to go to the next point; slide the insulation up to the joint, route it, and solder the next point; then again for the next one, and so on. So, essentially, it was one wire with gaps in the insulation. You can do it the other direction, but if it's a long run, sometimes the insulation doesn't slide along the whole length of wire very well.

Cheers! Rich

Hello Rich,

That's pretty clever.

In the days when I had to debug wrapped boards (not mine, I'll never WW a board or anything else for that matter) I used mostly hand tools like these because the techs had the electric wrap guns. Except late at night, then we could have the wrap guns. Got a lot of blisters from the repetitive motion.

For daisy chains I also use the melting-insulation wire. It's pretty easy, no tools.

BTW, the modified wrap they show can cause some grief. Many people don't realize that they create a small inductor of unknown properties that way. Usually ok for old TTL but for ECL that's another story.

Regards, Joerg

Out techs hated it when we used their tools. We had our own and had to take care of 'em. Touching their 30ga strippers was a crime. Using them as pliers was a capital offense!

Kynar? We were required to use teflon wire. Kynar resulted int too many "cold" (really cut insulation) connections.

Ok, but the "modified" wrap isn't in any way "modified". It is the standard wire-wrap connection. Using their "regular" wrap results in broken wires down the road. Gardner-Denver always made the "modified" wrap. ...and yes WW, even with the "modified wrap" was used for *years* with ECL. Mainframes boards used tons of WW, and were all ECL. I used WW with ECL in the 70s, and with 74Fxx and 74ASxx in the 80s. It works, but it takes some care. I had several boards with 6K wires (my techs loved the OT ;-).

--
  Keith

Hello Keith,

Oh yes, I remember how closely the techs guarded their tools.

I should have said ECL in an ultrasound machine. There half a nanosecond deviation can mess up the whole beamformer performance. So one day I got pretty tired of debugging other engineer's WW boards and asked the techs to solder one. Should have seen their looks.... but then again they were paid overtime for it so their grudge only lasted until payday. That fixed all the delay inconsistencies.

BTW, back at the university I soldered the backplanes of some dead mainframes that had been written off as too expensive to repair. After that they raised like Phoenix from the ashes.

Regards, Joerg

Where about in NH? I've worked with Sanmina in Salem (formerly Hadco) and a company in Dover NH, though I can't place their name. Both companies were great to work with.

--
  Keith

a *rose* by any other name?

Cheers Terry

The name doesnt ring any bells. But shit, their techs could solder.

Cheers Terry

Hello Terry,

Um, yes, should have said rose from the ashes. I guess I should allow the brain to adjust for a minute after switching languages...

Regards, Joerg

Ok, a half nanosecond translates to 2GHz. No, I wasn't doing 2GHz in the

70s. ...and no, I don't wire-wrap such circuits today. ;-)

It seems your engineers didn't understand the requirements (or physics). BTW, I don't know how you're maintinging 500pS with SSI anyway, ECL or not.

I had one case where the technician decided that he knew how to wire power and ground better than I. He daisy-chained *all* the connections. He then went on to exacerbate his stupidity by demonstrating that they all checked out on a continutiy tester. I went rip-shit, but a couple of weeks of work were already down the drain. He didn't work for me anymore. Instead, I got a new kid and trained him right (he ended up as an engineer in the research division).

OTOH, we also had an engineer that argued grounds. He couldn't undestand why "his" ( he took over responsibility) analog stuff looked so ratty on the scope, but it still worked, sorta. His scope ground was connected to the bat handle of a toggle switch on the test fixture. It took an hour to convince him that his "ground" wasn't *ground*.

As long as it's junk, why not try?

--
  Keith

Hello Keith,

What mattered was phase shift between dozens of transmit pulses. 500psec can start messing things up. The frequency was much lower.

Oh boy. I have seen that too, but not quite this bad. Except for the guy that wired all his stuff up on white board.

They became remarkably reliable.

Regards, Joerg

I understand, but even ECL has a Tpd variation more than 500pS. I designed an ECL clock generator in the mid '70s that had 64 programmable clocks with 1nS edge resolution over a 25nS cycle. I had to use tri-lead (twin-lead with an extra ground) to tune the delays in each channel. It was a real PITA. The control logic was all wire wrapped, but the clock drivers were on PCBs.

"white board"?

Properly done WW is as reliable as solder. I certainly wasn't qualified to WW customer deliverables, but someone wrapped many thousands of wires in each. Of course, GD machines did most of it. ;-)

--
  Keith

Ah, I've always called these things "EL sockets". IIRC, EL was the company that first made these things in the early '70s. I hated them then. After graduating I never had to look at one again.

I don't see their purpose. They may work for some really simple circuits, but it's not difficult to solder up a bread-board of similar complexity. I *hate* chasing intermittent connections. Time is important.

Whatever I tried it turned into a rat's nest. I remember losing compensation on 709s with these things. Ouch! Them things get *hot*.

Not IMO. You can have 'em.

Those are more interesting. Much pricey though.

My last few PCB designs I went straight to PCB. There were a few circuits that I was concerned about. *Those* worked perfectly. I F'd up the simple stuff. Though where I did screw up, the wires were all on the outer planes. Dumb, but easily recoverable.

--
  Keith