My initial proof of concept test with copper wire, shakey start but I think it will work. I think in the end it will merely be an equivalent option, not necessarily a slam dunk killer most practical option, unless you can source the wire cheaper.
I tried to film the process but all I got was 12 minutes of the back of my hand. ;) The wire I'm testing with is thicker than the wire I intend to use for real, and the pcb I'm testing with was not designed specifically for this, it was designed for TE Connectivity leadframes or millmax/keystone micro pins. So a couple of the holes were tight and the wire was hard to push through. Obviously that's just a matter of getting the sizes right for the holes and the wire, not a problem with the concept. Next problem was 4 out of 28 pins soldered themselves into the socket. That IS a fundamental problem with the concept, but maybe addressable by just throwing a piece of painters tape across the bottom of the board, and punch the wires through the tape. Or use a lot less flux. Or be more consistent with heat time. Or all of the above. This was bare copper wire too, and maybe gold won't wick quite as aggressively as bare copper. Maybe no extra flux, just rely on what's inside the solder? Practicality? All in all, it took several minutes to insert & cut the pins and then solder them, but this was the first attempt. I'm sure it will get quite a bit quicker even the 2nd time. I had to stop and figure things out several times, like the tight holes, and I didn't cut a long enough length of wire and ran out 3/4 of tye way through, and I had to work out a good way to grip the wire to push down into the socket. Even with all that it wasn't too bad and I can see how a lot of that was just first-time-bumbling. The basic process of poke the wire in with one hand and flush-cut with the other actually went about as easily as I imagined. I did the corner pins first and that was enough to hold the pcb in place so I didn't even have to rig up some way to hold the pcb down. I had a round pins socket in a large breadboard to make a stable base to hold the socket. However if the cost of the wire is going to work out to just over 6 cents per pin (buying in short 5ft lengths from ebay or amazon in the US) then the idea isn't much of an advantage over just buying the fancy mill-max pins at 10 cents per pin. And the TE leadframes are only 3 cents per pin. The time to assemble & solder the mill-max pins is about the same or less than the wire even assuming some practice working out a good procedure for doing the wire. I have some wire from different sources coming in that I'll just have to see if they work when they arrive. Some are much cheaper than others yet still claim to be real gold on half-hard brass or half-hard copper, in 26 and 24 gauge, so, as long as that description turns out to be true, then you can get much lower than 3 cents per pin and the idea is worth doing. Maybe it's even worth it at 6 cents just because all else being equal, it's better not to depend on some special manufactured product (just look at the option rom socket for a lesson on why). Gold plated brass wire is a little uncommon, but it's still just wire. -- bkw On Sun, Feb 28, 2021, 7:01 AM Brian K. White <[email protected]> wrote: > On 2/28/21 2:43 AM, Brian K. White wrote: > > > >> https://www.amazon.com/dp/B07S9JFK2V gold plated brass, 300ft, $14 > >> or > >> https://www.amazon.com/dp/B07H9RBTPF gold plated copper, 60ft, $5 > > > > It's actually kind of hard to tell for sure what many of the Amazon > > links are really made of. But this Aliexpress item at least explicitly > > says copper wire and real gold plating. > > https://www.aliexpress.com/item/1005002047296739.html > > > > I think brass would be a little better but maybe copper is strong enough. > > Found some gold on brass half-hard. 26 and 24 awg. Much more gold than > necessary, and costs more, but it's available in a few days. > $16 for 5ft = 254 legs = $0.063/leg. Not that much better than the fancy > machined Mill-Max or Keystone ones. > https://www.ebay.com/itm/273191001282 > > There's others like that. Looks like the ones with varnish usually say > so somewhere, or they say "tarnish resistant". > > -- > bkw > >
