Cleaning PCBs, particularly ionic residue

[Howardlong]

Are there suggestions for this? Twice now in the past six months I've been caught out by unexpected behaviour in boards carrying high impedance circuitry which has turned out to be due to ionic contamination, I assume predominently from the flux in the solder I've been using.

The way I've resolved it in the short term is to firmly brush the boards with flux remover and let them dry off, and it often needs a couple of goes. i can see the effects in real time by the way the current leakage dissipates as the solvent evaborates and does its thing.

If I were to look at other methids, such as ultrasonic baths, what are the recommendations here in terms of the bath itself and the comnsumables, associated tools and process? Is a <$50 jeweller's unit off Amazon going to perform for example?

[tautech]

As a result of our score here:
https://www.eevblog.com/forum/chat/auckland-beer-and-flux-xmas-get-together/

I've been doing a little homework on cleaning flux residue and stumbled on this link that might give some little answers Howard:
http://www.smtnet.com/Forums/index.cfm?fuseaction=view_thread&Thread_ID=4488

Second post.

[helius]

Cheap ultrasonic cleaners do not have swept frequency: in practice this means that the position and spacing of the cavitation whorls is fixed and so there are spots which do not get cleaned. If you manually move the board within the machine you can get better coverage, but touching an object that is vibrating with ultrasonic energy may not be safe. The cleaners use sequestering agents like ethanolamines and morpholinium compounds, which can dissolve ionic residues without being corrosive.

[KL27x]

What frequency? If It's under 500 MHz, just change solder. Rosin flux. Done. Leave the no clean for when esthetics are more important than function. I would not use low solids no clean on any high impedance circuitry that doesn't qualify as either a horseshoe or a handgrenade.

[Howardlong]

What frequency? If It's under 500 MHz, just change solder. Rosin flux. Done. Leave the no clean for when esthetics are more important than function. I would not use low solids no clean on any high impedance circuitry that doesn't qualify as either a horseshoe or a handgrenade.

In this latest case it's DC sub microamp energy harvesting where the leakage currents from a badly spec'd cap is very significant. What solder do you recommend? The reason I ask is that the solder you use is a bit of a habit, you tend to grow an affinity to it, so I'm reluctant to just change solder without knowing why or precisely what I'm changing to. Current solder is Multicore 309. http://www.epage.se/oem_el/uploaded_files/Lodtrad%20309%20EN.pdf

[exmadscientist]

I still stand by everything I said over here about critical cleaning. I am continually amazed how much one can accomplish with just Alconox and Citranox (poor man's Citranox: simply mix Alconox plus citric acid).

[KL27x]

Ah, you're already using rosin flux. Abeit, it has some tiny amount of halides, per datasheet, which probably not what you want.

Anyhow, I noticed this part :

The way I've resolved it in the short term is to firmly brush the boards with flux remover and let them dry off, and it often needs a couple of goes. i can see the effects in real time by the way the current leakage dissipates as the solvent evaborates and does its thing.

Letting board dry off allows the ionic contamination to replant onto the pcb. If you flush board with solvent after scrubbing and then blot, rather than allowing it to evaporate, you might increase your efficiency.

Personally, I have never been able to record any conductivity in RA flux residue, even at elevated temperatures. Not even megaohm range. Supposedly, there is some conductivity at very high frequency, though.... like maybe tiny capacitance or skin effect or something. But this is as long as I don't clean it. If you start to clean it but don't do a thorough job, you can make it worse by removing the rosin but leaving bare ionic contaminants. I have done projects where I had to swap resistors many times just to tune, precisely, due to resistor tolerances... and rosin flux residue has not made any detectable difference, cleaning it off or not. This why I wrongly assumed you were using non-rosin low solids flux core.

[SeanB]

If leakage currents are an issue how about giving the capacitor some stilts to enable cleaning under it if it is leaded, or if it is SMD have a slot milled in the board under it to provide creepage distance. Even for the leaded parts a slot is a good thing, to get air instead of board to provide insulation.

I used to use some little tiny PTFE washers under certain transistors in avionics, and there were little spacers for TO can transistors, IC's and for certain of the precision Vishay glass film resistors as well, along with some spacers for the relays. Those enabled just enough room under to clean after soldering, and also wicked up the conformal coating applied after so the component was almost part of the board itself.

[cdev]

A clean PCB makes all the difference for the performance of a sensitive VHF/UHF LNA.

[Howardlong]

If leakage currents are an issue how about giving the capacitor some stilts to enable cleaning under it if it is leaded, or if it is SMD have a slot milled in the board under it to provide creepage distance. Even for the leaded parts a slot is a good thing, to get air instead of board to provide insulation.

I used to use some little tiny PTFE washers under certain transistors in avionics, and there were little spacers for TO can transistors, IC's and for certain of the precision Vishay glass film resistors as well, along with some spacers for the relays. Those enabled just enough room under to clean after soldering, and also wicked up the conformal coating applied after so the component was almost part of the board itself.

The example I gave regarding the importance of choosing the right cap was to highlight just how low current the application is rather than it being a current problem (see what I did there?). Even a 10x scope probe creates significant load when measuring this thing, 10M ohm at 2.0V is 200nA which is in the same ballpark as the device's quiescent current.

On some of the very first prototypes I used solderless breadboard and was able to reaaonably accurately achieve the datasheet specs for typical current consumption. As soon as I got the hot pointy thing out on QFP packages, I got very different results, over an order of magnitude out. for example when a device should be drawing 250nA if was drawing 6uA. Even lightly blowing on the PCB makes a marked increase in current consumption, I assume it makes the residual ionic mix more volatile/mobile (disclaimer, I am not a chemist!)

[Howardlong]

A clean PCB makes all the difference for the performance of a sensitive UHF LNA.

I assume this is more to do wih the dielectric effect than ionic leakage?

Some years ago I worked on some microwave downconverters that used interdigital filters fabricated into the board, I had to be careful to scrub the residue off those or else there was a risk of them being off frequency, or so I was told.

More recently I've developed a number of devices for commercial use that work up to S band, but have never been able to quantify the benefit of scrubbing the boards other than for aesthetics. These are inherently broadband devices though, without anything more than controlled impedance microstrip and CPW.

[Conrad Hoffman]

I do a lot of critical cleaning at work and can confirm everything exmadscientist says. We use nearly identical materials and processes.

[cdev]

@Howardlong   I don't have any means of testing it, it is subjective. Its one of those things Ive done several times and re-learned. Thats the only application Ive seen where its mattered.

[cdev]

It may be possible to use a very cheap sealed ultrasonic transducer used to make Halloween "smoke" special effects with water as a ghetto style ultrasonic cleaner. 

[f5r5e5d]

ultrasonic bath is possibly useful but don't dismiss just using a toothbrush and detergent with tap water to start, then component safe solvents

isopropanol is often easier to source not having the tax issues of pure ethanol

Acetone, MEK attack some components' plastic, should only be used cautiously for spot flux removal where there's no polystyrene, polycarbonate, acrylic....

[cdev]

I have used some off the shelf citrus degreaser I have a bunch of sparingly with a toothbrush and then after that, 91% isopropyl alcohol and drying. That works better than 91% isopropyl alcohol by itself.

[richardlawson1489]

You can use cheap sealed ultrasonic transducer.

[Howardlong]

My final solution (oops, pun) was:

o Ultrasonic bath at 40 deg C for five minutes in Elctrolube SWAS cleaning fluid
o Rinse under cold tap water
o Rinse in cold de-ionised water
o Dry with a fan heater

There is a visible difference to the solder joints, and under the microscope the boards are noticeably spotless apart from some white solid residue around the area of the pins on some boards. Perhaps some of them need another go.

The electrical results are much better, I can get down to a static 500nA quiescent current whereas I was lucky to see 3uA. The current draw does not change if you blow on the devices either: I assume this was residual ionic compounds in a slightly volatile suspension or fluid that were evaporating, changing conductivity.

[BrianHG]

The solution of my friend Frank, who used to run a PCB assembly house, when I did a few battery powered devices which leaked current:

(And I know you guys will frown on this one)
ULTRASONIC TOOTHBRUSH with firm bristle + FORMULA 409 all purpose cleaner, then a normal flux off cycle after that.
yes, this stuff > https://www.formula409.com/products/cleaners/multi-surface-cleaner/

Fringe and tell me off all you want, block this post, have it deleted, but, no matter which normal PCB solvents and chemicals personally I used to no avail, the above toothbrush and formula 409 he brought over to my lab erased the problem...

[SeanB]

Basically a nice industrial detergent, with a little colourant and some fragrance added. Nothing wrong using that, it will dissolve both polar and non polar residues, and a short application, like with all cleaners, will not harm plastics, copper or rubber. Very important step that you also used was the rinse, to remove the dissolved substances off the board.

Not as if this is a proprietary PCB cleaner, just that it is disinfectant as well, but not too corrosive when used with a proper rinse cycle. I am sure there are expensive PCB cleaning solutions sold that use the same basic chemical parts, and the only differentiation between them and this is the added fragrances and possibly a colour difference, and perhaps a thickening agent to make it a gel before you mix it to the recommended dilutions.

[Howardlong]

The Electrolube SWAS smells and behaves suspiciously like diluted washing up liquid, FWIW!

[Someone]

The Electrolube SWAS smells and behaves suspiciously like diluted washing up liquid, FWIW!

Its not dilute when delivered, try washing your hands in it and watch the skin come floating off (seriously bad for you). But I'm assured by the chemists that it is basically the same action as a strong dishwashing agent.

[BrianHG]

Basically a nice industrial detergent, with a little colourant and some fragrance added. Nothing wrong using that, it will dissolve both polar and non polar residues, and a short application, like with all cleaners, will not harm plastics, copper or rubber. Very important step that you also used was the rinse, to remove the dissolved substances off the board.

Not as if this is a proprietary PCB cleaner, just that it is disinfectant as well, but not too corrosive when used with a proper rinse cycle. I am sure there are expensive PCB cleaning solutions sold that use the same basic chemical parts, and the only differentiation between them and this is the added fragrances and possibly a colour difference, and perhaps a thickening agent to make it a gel before you mix it to the recommended dilutions.

That electric ultrasonic tooth brush is crucial as well.  I tried 409 with a regular brush and it does not much better than FluxOff on it's own.  I'll try to get the electric toothbrush model type & # tomorrow.

[Howardlong]

The Electrolube SWAS smells and behaves suspiciously like diluted washing up liquid, FWIW!

Its not dilute when delivered, try washing your hands in it and watch the skin come floating off (seriously bad for you). But I'm assured by the chemists that it is basically the same action as a strong dishwashing agent.

The documentation states it should be used as it is supplied without dilution. Probably not for the dishes though.

[Someone]

The Electrolube SWAS smells and behaves suspiciously like diluted washing up liquid, FWIW!

Its not dilute when delivered, try washing your hands in it and watch the skin come floating off (seriously bad for you). But I'm assured by the chemists that it is basically the same action as a strong dishwashing agent.

The documentation states it should be used as it is supplied without dilution. Probably not for the dishes though.

Yes, but the MSDS has some strong warnings such as:
R41 Risk of serious damage to eyes.
S39 Wear eye/face protection.

HAND PROTECTION
Use suitable protective gloves if risk of skin contact.

Its not benign like dishwashing liquid and is much more concentrated chemically. Ignore the viscosity of it, thats added to dishwashing liquids for the consumers benefit, the viscosity is no indicator of concentration.

[gbyleveldt]

I use SWAS prior to repairing water damaged automotive electronics. Basically, I clean off the worst of it with a short bristle brush dipped in SWAS and then remove under water. Once the debris is removed from the board, I dip the boards into an ultrasonic bath of SWAS (I don't heat the solution up) running it for 2 minutes both sides. This cleans up all contaminants in hard to reach places and cleans out the many vias as well. Hot air works well to dry the board and once the repair is complete it goes for another round of cleaning in the ultrasonic bath and then water.

SWAS does appear to be conductive, so it must be cleaned with water afterwards and dried properly. At least, that's my experience. It smells very similar to window cleaner.

[Bud]

ULTRASONIC TOOTHBRUSH with firm bristle + FORMULA 409 all purpose cleaner, then a normal flux off cycle after that.

Where do you buy this stuff, it seems not available in Canada.

[cdev]

He probably means "Sonicare" (available everywhere in US so probably Canada too)  You can use the handsets and brushes for PCB cleaning when they wear out for regular tooth-brushing.  (MAKE SURE TO MARK CLEARLY)

[Bryan]

Think Bud meant that 409 is not available in Canada.

[BrianHG]

Think Bud meant that 409 is not available in Canada.

My mistake, the cleaner was 'Spray Nine', sold here:
http://www.canadiantire.ca/en/pdp/spray-9-marine-multi-purpose-cleaner-946-ml-0790120p.html#

[Bud]

Thanks, will give it a try.

[Bryan]

Think Bud meant that 409 is not available in Canada.

My mistake, the cleaner was 'Spray Nine', sold here:
http://www.canadiantire.ca/en/pdp/spray-9-marine-multi-purpose-cleaner-946-ml-0790120p.html#

Thanks for the tip, picked a bottle up today and it is excellent. Although haven't tried it on a PCB it worked like magic cleaning off baked on grease grease and hardened sticker residue on some equipment.

[Bud]

I found it in Wallmart in automotive section. Tried on a piece of PCB against clean-free flux and it seemed to work. I used a regular toothbrush though, not an ultrasonic one. My concern was the white residue which you get when cleaning with IPA for example. But the board came very clean, no any white residue  under the microscope inspection. Seems to be good stuff. Thanks for the tip. It may be a bit expensive though per volume, compare to Alconox which I have not tried yet but am planning to.