another hack

te:

lund:

ragelund:

hro Pefhany:

ne. There's a

this, so

possibly you could get inspired by this :

se ~ $1K

spade bit is

few codes

axis and

luminum are

:g:xXYAAOSwDN1USolI

like 100 USD per letter, so I bought the machine to mill the sign myself. Worked out great, and now I can do all kinds of funny stuff (If I ever get the time to)

eds a spare

tware is based on linuxcnc

with USB or

haven't dug deep into the SW, but I think the MACH3 outperforms GRBL by a long shot. I have the probing tool to, so just place the PCB on a support p late, probe the edges and surface, and it is calibrated to provide better r esults

is a bit of simple milling and don't have a parallelport it'll do ok

d, so you can even do 3D printing

umbersome way of making flimsy ugly plastic parts, fine for a visual protot ype but not very

CNC and let it scan a PCB for radiated EMC debugging

of nails" extract the testpoint coordinates from the PCB and use a pogo pi n instead of

now you don't use standard probes, you just tell the scope which test poin t you need measured :-)

Yes, but nobody's doing it. Maybe it's in the ToDo list of Tektronix

I've made boards which I coated with an equivalent home-made electroless silverplate powder, and they still looked quite good a couple of years later. They certainly don't lose solderability as quickly as plain copper does.

It's a bit of a bother to grind up the salt and silver chloride, but the resulting powder works fine and it's significantly less expensive than Cool-Amp.

I do like the look of that goldplated protoboard, though. Can't justify having any made (yet) but I agree it's lovely stuff.

If Cool-Amp is like the homebrew stuff I cited, it creates a *very* thin layer of silver. The silver is deposited directly on the copper... I don't think it puts additional silver on top of silver. Salt Lake Metals makes a point of saying "This method produces a very thin silver plating. Heavy Polishing will remove the plating."

As a result, it probably isn't great for resisting wear, such as you would find on a silver platter.

If you need something which will stand up to wear, electroplating is almost certainly the right way to go. You can buy use-at-home silver electroplating kits (cyanide-free) if you hunt around a bit.

I have a couple of ideas for a fixtureless, motionless bare-board tester. One more thing I may never have time to pursue.

Silver tarnishes from the sulfur in car exhaust. From

"The main product of silver tarnishing is silver sulphide. The reaction mechanisms are:"

A lacquer coating can present a barrier to tarnish. The same would apply to copper. This would give the circuit some protection at greatly reduced cost. Gold is nice, but that is from HP in the 70's. Copper would impress knowlegable clients that you know how to manage without profligate waste.

OK, this one uses 2-56 screws to connect the grounds and to be the power connections. I kind of like this better.

On Wed, 9 Dec 2015 13:00:19 -0800 (PST), Lasse Langwadt Christensen Gave us:

Silver oxide actually conducts better than raw Silver.

Silver is indeed the best metallic form conductor, but the oxide of Silver has the amazing property of actually conducting better than its base form metal. A property no other in the family possesses.

All our boards are ENIG now, gold over nickel. There's microinches of gold. ENIG is a great treatment for BGA boards; solders like mad and it's nice and planar.

In the glory days of HP and Tek thru-hole gear, lots of boards were plated with gold over copper. It takes a lot of gold, something like 1 mil, to keep the copper from diffusing into the gold and making it ugly.

On Wed, 09 Dec 2015 16:45:56 -0800, John Larkin Gave us:

And when there is that much Gold, it ends up in the solder and causes embrittlement problems.

You may have heard the joke about some of these things :

So much backlash you are climb milling no mater which way you go.

I use 28 gauge brass wire and a 13.5 thou drill. The fit is just tight enough to hold the wire in place while you solder it, but the brass is stiff enough to insert.

The technique I use is to wrap solder around the board, which holds it a "little" above your work surface. Insert brass wire, solder the top, cut, repeat. After doing all the holes, flip the board over and solder all the other sides. Again, the fit is snug enough that the wires won't fall out when you do this, despite melting the solder on both sides.

But if you're doing this often enough and the vias are big, switch to via rivets. Insert, BAM, done.

If you flip the corner ones over, you can charge extra for "built-in standoffs" :-)

But really, I was asking about potential problems with long ground paths, a problem with digital signals. Does it just not matter in this application? Would extra vias near the signal paths have any effect on the signal quality?

Your production PCB's are a different animal. Gold is part of the process. Relatively cheap.

For your prototypes, gold is a special process. I think you mentioned $100 per square foot. That is a very high cost.

For high frequency where skin effect is a factor, gold has much poorer conductivity than copper. From

Table of Resistivity and Conductivity at 20C Material Resistivity Conductivity Silver 1.59e-8 6.30e7 Copper 1.68e-8 5.96e7 Gold 2.44e-8 4.10e7 Nickel 6.99e-8 1.43e7

So if you are talking about 100ps performance, you want copper, not gold or nickel.

There are many products that will protect copper against corrosion, such as

Build the circuit, debug it, clean it, and apply protective coat. Cheap, higher performance, looks good.

p-type semiconductor. 0.63 V Not amazing. More like typical.

Sadly, plated silver is less conductive than bulk silver - apparently plate d silver has an odd crystal structure, or something - and the old-fashioned habit of silver-plating RF components for better conductivity in the skin depth died off when this became well-known.

These days I'd have expected the lunatic fringe to have used a femtosecond laser to melt the plated-on silver and turn it into bulk silver, but it loo ks as if RF hasn't got an audiophool lunatic fringe.

A little palladium or AuPd can reduce gold diffusion.

Nickel is lossy at high frequencies. When loss budgets are tight, you can get rid of it and pick up a few tenths of a db here and there.

Then there's an Etch-A-Sketch style.

Mikek

That's a reasonable price to do a weird special. They plate both sides. I'm happy.

The fastest thing we're currently making has 40 ps edges. The board is FR4, gold over nickel plated. The biggest problem wasn't the plating, it was the transition from the microstrip trace into the edge-launch SMA connector.

No thanks.

Nickel forms tunnel junctions very easily. (I used to make them on purpose for rectifying light, but they also form on their own.) The nonlinearity of those junctions at a few tens of mV is what's responsible for the bad reputation of nickel-plated RF connectors.

Nickel oxide is another of those weird materials that you don't want to deal with if you don't have to.

Cheers

Phil Hobbs