Flux fux, and 10 Meg ohm

Bob Pease used to swear by a dishwasher, and a standard load of Calgonite dishwashing powder (this is a formula which does not contain a chlorine bleaching agent). Let the washer run through the wash and rinse cycles. When it finishes the rinse, take the boards out, do a final rinse with distilled or deionized water, shake off the excess water, then stand the boards on edge to drain and dry.

He wrote that this approach gave better results (lower leakage on the board) than the use of expensive cleaning solvents.

...

I've successfully used a technique of not putting any solder mask under the resistors so that washing can be effective.

This was in a design where I had thousands (3072 in fact) of high-voltage (

Our initial attempts at just using conformal coating showed us that water p ermeates straight through and then dissolves any ionic debris to create a l eakage path.

I agree that No-Clean flux is terrible!

kevin

Interesting idea. Seems like it would depend a fair amount on the paste volume, though.

Yup. Even Parylene is just plastic. To get a hermetic seal you need metal, glass, or fully dense ceramic. The coating does help keep crud from moving around, though, and ions diffuse much more slowly than water.

I'm actually a fan of No-Clean with leaded solder--it's just good old RMA in disguise, and so comes off easily with solvent. The residue doesn't leak much. In the lab I clean it off with HPLC grade methanol, scrub with a toothbrush, rinse with more methanol, then chase drops with canned air. Expensive per unit, but not many units.

Cheers

Phil Hobbs

Calgonite used to be pretty well just sodium tripolyphosphate, but not any more. I get TSPP from eBay, because the replacement detergent turns silverware black, and there's no way I'm going to eutrophy the lower Hudson estuary.

Cheers

Phil Hobbs

Yup and yup. Canned air is way cheaper than freeze spray, and AFAIK the only difference is that there's no feed tube inside, so it empties from the top down. Gotta be careful with new cans, though--it bubbles inside as you use it, and sometimes you get drops of freeze stuff coming out. On optics that can be fairly permanent.

That's my preference too.

Cheers

Phil Hobbs

...

Probably - we didn't experiment much once it passed our testing. I can't remember our stencil thickness.

The detailed shape of the underside of the resistors may also affect it. I was using expensive Ohmcraft 1206 resistors - $1 each in tens of thousands!!!

kevin.

Getting the switch "right", I'm guessing, is the hard part.

Grin.. I was stopping by to see me mum on the way home ... the bit of comatose, was a good thing. Sit, talk to ladies, relax....

George H.

Ted's makes a superior hotdog, charcoal fire and a heavy grate. But I go for the thick shakes and onion rings! I've never had a better ring, very light batter and lotsa onion.

George H.

Thanks Jon, These boards are all made up with switches, pots, connectors... replacing a few R's and soldering with kester 44 did the trick. What sort of flux were you using?

Out of (my) curiosity; Did you ever look at leakage as a function of voltage? George H.

Yeah I've read that. In this case the crud was all under the smd... So (as many suggest) in the future, through hole or a slot underneath the device.

George H.

Assuming (as I will) that there's a problem with standoff distance and board contamination, one would think that resistor manufacturers could tin their endcaps over a few BGA balls to create board clearance. Or, maybe a MELF option with a real end flange?

Until then, the folk who were satisfied with leakage did long-time-duration agitation things ? many minutes in ultrasonic cleaner, ? wash/rinse-but not /dry - cycle of dishwasher ? water-pik spot cleaning

General agreement on a final clean-solvent rinse and blowing away the droplets (instead of letting them dry and leave residue). Even auto wash folk know to do the final rinse in distilled water...

The agitation and long time is because there's a deep narrow crevice to be cleaned. Diffusion of crud through solvent goes as cm**2/second, and if you want fewer seconds, you want more under-chip volume and less under-chip reach distance.

Has anyone tried MELF packages,because they have less- cramped space underneath? Is there any merit to asking for thicker component-side copper, to create some standoff?

Bob Pease's 'calgonite' wash reads like an old-wives-tale recipe: I'm dubio us that any part of the cleaner except the wetting agent was really useful. A washerload of Calgonite (modern substitute: Cascade) introduces maybe 35 grams of active ingredients and wetting agent, but the real need is just the wetting agent (Cascade product "Rinse Aid" recommendation is about 3 grams per load).

In the matter of blowing off droplets of the final rinse, has anyone set up an assembly-line with 'air knife' nozzles, and run a tests with a standard contaminant? Even just dipping a test board in molasses, then rinsing, blo wing dry, and weighing, would tell you how effective the rinse/dry is.

Get a grad student on it, there's a real publication prospect here!

On Oct 7, 2016, whit3rd wrote (in article):

Beware of ultrasonic cleaning of PWBs with components soldered on - many components cannot handle much exposure to ultrasonic energy.

A big part of the ingredients of automatic dishwasher detergent of the day was sodium tripolyphosphate,which softens the water and strips the grease off the plates (or PWBs). One can use the phosphates alone, but a little detergent greatly improves things. But detergent alone will not really work.

The modern Cascade is phosphate-free, and does not work nearly well enough for PWBs. Or for plates either.

Instead, use AlcoJet from Alconox, Inc on PWBs. Works very well on dinnerware, but will wash the glaze of pottery a bit too quickly, as dinnerware is washed many many times. You do not need much AlcoJet, and it?s easy to use too much.

Be sure to filter all the oil and water out of the air, or the subsequent application of conformal coatings may go badly.

Joe Gwinn

Surface tension won't lift the part. The main source of gap is the copper thoclness.

Yes, I worried about that. We've cleaned hundreds of this particular board, with a mix of parts, including big oxicaps and SMT inductors, never had a problem with them.

Jon

Hmmm, I don't know. This high-impedance stuff is done with some PbSn solder, and I'm guessing some "no-clean" flux. This is all solder paste, P&P machine and then reflow in a toaster oven.

Yes, it is NOT a simple resistor! Nowhere CLOSE!

So, this particular board has 32 channels, each one has a 10 nF coupling cap and a 10 M bias resistor on the detector side. Plus, there are 2 of the same caps in a bias filter RC. So, for the up to 500 V bias supply, it looks like all 34 caps are in parallel. We kept cleaning until the leakage at 300 V was less than 100 nA after allowing the current to stabilize for a few minutes. So, that's better than 3 G Ohms for the 34 in parallel, or about 100 G Ohms each.

Because the leakage current changes with time, and if you breathe on the board it shoots up, I don't have anything like a graph. But, it does seem like the leakage increases to about 100 V, and then kind of levels off.

Jon

If it's dramatically humidity sensitive, there's ionic contamination from something water soluble, most likely the flux.

This board has voltages up to 1400. They all work great, assembled with rosin flux and solvent wash.

Here's a little test board, glopped with rosin flux, not cleaned.

The needle is pinned on the 1e14 ohm range.

AND dirty oil.

paste

Sure it can. If there's enough solder the part will float.

Cheers

Phil Hobbs

Yes, for sure. As we go through several cleaning cycles, it gets less sensitive. But, exhale slowly over any high-impedance (G-Ohms) circuit, and you are bound to see the resistances go down.

Yes, but leave it exposed to the St. Louis summer humidity, and I'll bet that reading will go down after a day or two.

Jon