another hack

I scrub the copper-clad with steel wool, then Comet & rinse to remove oil, then spray with Krylon clear acrylic spray.

You can solder right through it, and it keeps the copper clean and shiny. Did my first in ~1977. It has seen salt air, heat, and humidity over the years, and it still looks brand-new.

Cheers, James Arthur

40ps on FR4? You won't even notice the gold and nickel.

The green ScotchBrite pads work wonders. No Comet needed but it doesn't hurt. Steel wool disintegrates and gets into everything (including skin).

Silver creeps across insulators and forms whiskers under humid conditions - that property is the main reason it fell out of favor. Conductivity is affected by process conditions. Cranking up the voltage to speed the process can embed hydrogen into the lattice. Most plating companies know this now and stick with current densities that give a high quality plate.

Silver plated coils are routine in HF electronics. The plating does conduct better than copper - 2-5% better Q - not much but every little bit helps.

--
Grizzly H.

Ni is a no no in the RF world due to IMD issues. It is never used at cell sites.

ScotchBrite's a good idea. The steel wool cleans oxidation well, but it's oiled. I haven't had any trouble with steel wool crumbs. I've heard they get embedded in copper, assume it's true, but if so, it's never made any difference.

I use broken 1/8" shank carbide drill bits for Dremeling FR4--works great.

Oh, I should've mentioned--etch or Dremel the board, clean, spray, THEN assemble. You can solder right through the Krylon, perfectly, and it doesn't even reek.

Cheers, James Arthur

I dip the board in 10% HCl solution and it really cleans up all the oxides. Then rinse it off. This also makes resist stay on the board.

NMR hardware gets a white bronze (CuSnZn) plating if it has any chance of getting into a magnet. Huber & Suhner for sure, and most other companies if you ask, sell connectors plated with it. The bath is proprietary so not every plating house will have it. There are various trade names - Albaloy, Miralloy, Optalloy. The phone companies switched over decades ago. A lot of costume jewelry is a bright variant. Doesn't flake like nickel either.

--
Grizzly H.

I would totally use this as an excuse to by a small drill/mill.

--
Les Cargill

By bad reputation, do you mean PIM (passive intermodulation distortion)? I didn't know that nickel creates tunnel diode junctions. I've always assumed that the PIM problem was caused by the non-linear V/I relationship (hysteresis) of shoving power through a ferromagnetic material, such as nickel, iron, and steel. That's a bulk effect which is bidirectional, while the tunnel diode junctions are possibly unidirectional. So, is the problem with nickel plating (or the nickel under the gold plating) PIM, tunnel diode rectification, or are they both the same thing?

I also hadn't heard of that problem. Can I get a link to whatever I'm missing?

I don't think any of the aformentioned is going to be a problem with JL's ENIC PCB's. The power and current levels are not high enough to create PIM problems with non-linear conduction through ferromagnetic material. However, scattering tunnel diodes around the PCB seems like something worth worrying about.

Incidentally, I threw together a PIM tester from a pair of commercial

60 watt UHF radios and a butchered cavity duplexer. It was sensitive enough to identify that the steel screws used to hold down an SO-239 flange mounted connector were a source of PIM. I also confirmed that a thicker layer of gold over nickel is a good but expensive way to reduce PIM. Replacing everything with silver plated connectors is probably the best.

--
Jeff Liebermann     jeffl@cruzio.com 
150 Felker St #D    http://www.LearnByDestroying.com 
Santa Cruz CA 95060 http://802.11junk.com 
Skype: JeffLiebermann     AE6KS    831-336-2558

I spent about seven years of my life building Ni-NiO-Ni junctions for optical communications, and managed to get about a 50x improvement on the state of the art for efficiency, so I'm sort of a primary reference on that point.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
ElectroOptical Innovations LLC 
Optics, Electro-optics, Photonics, Analog Electronics 

160 North State Road #203 
Briarcliff Manor NY 10510 

hobbs at electrooptical dot net 
http://electrooptical.net

I know, but it seems John really wants to continue doing it by hand so I was pointing out that he could get the best of both worlds by using a cnc router in manual mode. Consistent cut depth, straight lines, bare minimum set up time, and no programming. Just stick a piece of board on the table and start cutting lines. If he did get one, I'm sure he would eventually feel compelled to connect a pc and play with programming, but he doesn't need to to get started.

----- Regards, Carl Ijames

Den onsdag den 9. december 2015 kl. 19.29.19 UTC+1 skrev Carl Ijames:

gcode is just text and extremely simple, you can easily write it by hand. or just use one of the numerous programs to draw and make gcode

-Lasse

Hopefully, I didn't say anything wrong or offensive. I'm certainly not questioning your expertise on the subject. I just want to know if we're talking about two different things. Is the sensitivity to "a few tens of mV" problem the same as PIM, which requires much higher power levels? I would think they're different, but I would like to be certain.

--
Jeff Liebermann     jeffl@cruzio.com 
150 Felker St #D    http://www.LearnByDestroying.com 
Santa Cruz CA 95060 http://802.11junk.com 
Skype: JeffLiebermann     AE6KS    831-336-2558

Works great. There's nothing wrong with using FR4 for fast stuff. Just do it right.

I use SoftScrub on both copper and gold-plated copper. It shines it up beautifully without scratching.

there's no special process about them, it is just that normally most of a pcb will be covered with soldermask before you put on gold so you use very little gold

-Lasse

[snip]

That was my impression initially, but the newer machines are a big improvement. We have two, a huge SLS (Selective Laser Sintering) machine from Z Corp and a newer smaller machine from Ultimaker that extrudes ABS or PLA.

They both get used a lot, mostly for in-house jigs and fixtures. The SLS machine is a bit cumbersome, the parts are dimensionally pretty stable but very brittle. There is a post-processing step which impregnates the part with epoxy resin, this improves the strength a lot.

The Ultimaker was about EUR 1800 IIRC, it's paid for itself many times over. One problem is that as you say, the factory settings produce very flimsy parts and a lot of people give up at this point. However a bit of tweaking of feed rates etc can produce a massive improvement. One of our guys printed some toys for his 5yo and they are still in one piece, I'd consider that a fairly harsh test.

FR4 does make for dispersive transmission lines. 40psec is 8mm of propagation, so John probably hasn't built anything large enough for it to matter.

Microstrip is intrinsically dispersive, so the FR4 probably doesn't make it much worse. If he ever went to trouble of making buried stripline transmission lines, he might need to worry.

--
Bill Sloman, Sydney

No worries, I was just in a grouchy mood. Sorry I was a bit curt.

The relevant voltage is the one across the tunnel junction, i.e. in series with the shield, not the one across the dielectric.

I haven't done any high power RF measurements on connectors, but anecdotall y I understand that the IMD effect depends on what you had for breakfast--r eplacing the connector with one of the same type. mating and unmating etc. Can change the IM a lot, the same way it does with ground loops. (Ni makes ground loop problems worse.)

That sure sounds like a surface problem to me, and oxidized nickel forms th ese super low impedance junctions that are also super fast (about 1 fs for the optical rectification process).

Cheers

Phil Hobbs