Reverse polarity, Scotty.

Remember my deliberately-massively-flux-crudded leakage experiment?

Rosin flux, unwashed, high humidity, deliberate fingerprints: pinned the needle for weeks on the 1e14 ohm range.

The only leakage problems we've had were on boards that we sent out, processed with water wash. In-house, we use rosin flux and solvent wash, no problems.

:envy.

Cheers

Phil Hobbs

On Thu, 12 Sep 2013 11:32:05 -0400, Phil Hobbs wrote: [snip]

All I use now is Amtech LF-4300. Water Washable. Stencil and tack time are great. Good joints, good repeatability.

That's the problem with no-clean flux: it's so hard to clean.

No idea, but when they say something like that they usually have done pretty good sims and validations on earth with the same hardware.

How is it with stuff in the high gigohms?

Cheers

Phil Hobbs

I cannot really help you with that, above my range. But on some high value boards I usually visually inspect the joints under a microscope and the boards are clean. If you have a way to do a couple of test boards with this stuff I would give it a try. It's not cheap but cheaper than your other options.

Thanks, I'll pass that along.

Cheers

Phil Hobbs

If a flux is water soluble, it will leave ions behind, potential humidity sensors. Stuff likes to get trapped under parts. A dirty board can usually be fixed with a serious nook+cranny blast treatment with a water-pic dental thing and really clean (distilled or DI) water, then baked. Expect to get wet.

Yup, 63/37 Kester 44 is the ticket.

Had some noticable problems in a precision AC-bridge front end design where 5-10G leakage made a difference.

Vaguely.

What was the relative humidity?

I see that you are in San Francisco, so it may be reasonable.

I recall having problems with static electricity sparks disrupting a printer, in the 1970s in Baltimore, MD, during the winter. The computer vendor, located in Ft Lauderdale, Florida, had never had any such problems.

Anyway, those ions are the reason for the trip through the dishwasher. Alconox and competitors have specialized detergents for exactly such problems.

Joe Gwinn

Stopping the leakage current also reduces the noise.

No-clean is definitely a compromise.

I agree with the other posters that Kester 44 followed by a solvent wash works far better, especially of the wash is followed or replaced by a vapor degreaser using a solvent that can dissolve rosin.

I recall reading an article from a scientist that was doing femtoamp experiments saying that the rosin flux also left stuff in the boards, so he did a solvent wash followed by hot water followed by distilled water followed by alcohol.

If the intent of the hydrogen or forming gas is to prevent oxides from forming during soft soldering, carbon dioxide or nitrogen may be sufficient, and are not explosive.

As for the nooks under components, one must design such that one can get at and clean those nooks. The dishwasher will get into such nooks, and requires no touch labor.

What is the highest humidity at which these boards must work?

Anyway, it sounds like some experiments are in your future.

Joe Gwinn

It will reduce the 1/f noise, but not the Johnson noise due to the shunt resistance, and not the multiplication of the voltage noise of the amp due to the increased noise gain.

Plus Kester 44 is an RA flux, which cleans oxide and crud off the metal better. I just bought another 4 pounds of it from eBay for $60 total--I'm up to nearly a career's worth, but I have my son to think of. ;)

Usually a polar solvent followed by a nonpolar solvent (or the other way round) works best. Iteration may be required if the contamination is heavy.

The feedback resistors have unplated holes under them, specifically to allow good cleaning.

In the fab, about 45%. In test, probably much higher. They're made in Atlanta.

The client sent me a board yesterday, but I'm hip deep in legal hair-splitting at the moment.

Expensive lawyers really are good at what they do, for the most part--they're very persuasive. Fortunately I have enough technical _nous_ to see what they're up to most of the time (I think).

Expert witness work uses a different part of the brain, for sure.

Fortunately I get to go back to building stuff in a week or so.

Cheers

Phil Hobbs

What's the lowest frequency of interest?

I think it's non-polar first, because much of what non-polar dissolves is waterproof. But there is definitely a chicken-and-egg problem here, and an iteration can solve that.

Ahh, good. But also need to ensure that there is a minimum spacing between the bottom of the component and the board surface, so the cleaning solutions can flow freely. And there are no creepage paths in the air gaps.

How about in use?

Just make sure not to lose virtus in that environment.

It sounds like escape planning to me.

Joe Gwinn

A few hundred hertz.

I tried to get them to use 4-oz copper on the top, but they didn't, at least not for the first batch. There's not much copper there, so there shouldn't be a warpage issue.

The fab is where they're used. (It's a semiconductor manufacturing tool.)

It's the other side's lawyers I'm talking about--I'm fisking some really excellent rhetoric from the other side, in which they're defending what on examination turn out to be weak technical arguments.

One of the good things about being an expert witness is that you're paid to be an advocate for the facts. The lawyers are advocates for their clients, so you do have to keep your balance. Of course if you fall apart in deposition or cross-examination, it's worse than if you were never there at all. So if the lawyer is a bit too insistent on some point, all you have to say is that you wouldn't be comfortable saying that in a depo.

This is my seventh or eighth case, and I must say that I've never felt pressure to say anything I didn't believe, regardless of how much money was at stake. (I've been retained by some big outfits, too, such as Intersil, Samsung, and Lockheed Martin.) Despite pervasive public cynicism, IME the patent litigation system works pretty well.

Cheers

Phil Hobbs

I would submit that you need to stop all leakage then.

OK, but warpage isn't the issue for cleaning under most parts, at least the small ones.

Sounds a bit like sausage making.

Joe Gwinn

PS: I'll be offline for a few weeks.

Right. But the usual objection to heavy copper on an outside layer is possible warpage.

Bismarck thought so, but then he was putting together an empire, and I'm just making a living by telling the truth, which is a better gig than lots. ;)

Hopefully for a pleasant reason, such as sailing your yacht around the world?

Cheers

Phil Hobbs