We sent out a batch of 35 VME module kits to be built by a contract assembly house. There are lots of high-value resistors on this board
formatting link
in the filters and such. Boards started failing in test and it seems to be caused by ionic contamination trapped under parts. The stuffers used water-soluble flux (which is contrary to our rules) and obviously didn't clean the boards enough. They claim to use a super high-pressure conveyerized spray cleaner with super-clean water. I'm skeptical about the cleanliness of their system, and they just told us that the cleaning line "just broke" so now they can't rerun the boards.
So, what's your experience? Can a water-soluble flux be reliably cleaned off to decent leakage levels? Can they really clean under surface-mount parts?
Last time this happened, some years ago with another assembler, we nabbed a sample of their wash water, and it was 20x as conductive as tap water.
I'm thinking in terms of slowly hand-scanning each board with a water-pic sort of high-pressure wand, with single-use distilled water, or something like that. It looks tricky to clean under surfmount parts, especially with water. Our normal process is RMA flux followed by solvent wash in a vapor degreaser.
Is there anything that can be added to the wash water to reduce its surface tension, to get under parts better, but that isn't itself a source of leakage? Maybe an alcohol/water mix?
Are there any lead-free implications as regards leakage? This board uses regular pb/sn solder, but it may become a concerm some day.
Did this "just broke" happen right after you complained about the PCBs adn asked that they be rewashed?
If so, I'd start to suspect everything they've told you. Chances are it really failed before your PCBs. It wouldn't suprise me to dicover that they cleaned them with a crub brush and a bucket of water.
As far as I know the board house that did the last PCBs for me used what is really a dishwasher with a different trim option. The trick seem to be just to throw enough water at it for long enough.
The only flux I've really had trouble with was the "no clean" type. I don't think it should ever be used.
Using dirty solvent is trouble, whether or not it is water.
... or ... You could attach the PCBs to the wall and get 20 guys with fire hoses to blast away at them for a day. If one squirt with water only lowers the flux level by 10%, 1000 squirts will get the board clean.
I believe that there is a detergent of some sort added to the first wash. After that it is washed with pure water.
A lead free soldered PCB will short out due to tin whiskers, so the ionic stuff won't matter at all. :)
The unobtainium in the lead free flux may be harder to move.
Due to allowing ESD devices to be used as clamps ?:-)
...Jim Thompson
-- | James E.Thompson, P.E. | mens | | Analog Innovations, Inc. | et | | Analog/Mixed-Signal ASIC's and Discrete Systems | manus | | Phoenix, Arizona Voice:(480)460-2350 | | | E-mail Address at Website Fax:(480)460-2142 | Brass Rat | |
formatting link
| 1962 | I love to cook with wine. Sometimes I even put it in the food.
In one case, I successfully improved the leakage resistance of a board with surface mount parts, but it didn't include a battery or any electrolytic capacitors. I first soaked the board in warm 99% pure isopropyl alcohol for an hour or so, to remove any covering of rosin flux (some parts had been hand soldered), and then cooked the boards for an hour or so in near boiling (perhaps 85 C) distilled water, followed by a couple rinses in cold, clean distilled water.
I then baked the boards in a vacuum oven at 50 C, overnight, to make sure all the surface bound water escaped.
My real concern is leaving hygroscopic crud under the parts. If we bake them dry, they'll pass test, but may well later suck moisture out of the air and fail in the field.
We've never had any leakage problems with rosin flux. It's not conductive whether you clean it or not.
Had this sort of thing happen in the mid 90's on a board with a hi-Z (10Meg) node. We got on the assembly house's case and made sure they gave the boards a final wash in a fresh batch of cleaner. I believe they were using water soluble cleaner. Problem was contamination under a SOT23 Schottky diode. In retrospect, I wonder if it would have helped if the soldermask was removed under the part to give more clearance between the underbelly of the part and the board?
Don't knock dishwashers. They work great. The problem is getting flux out from under those high-value parts--it should happen eventually, but you'll have to calibrate it. Try bringing a board home and running it through your DW three or four times, first with detergent and then with ordinary tap water.
Yes, you would want to soak them in the highest specified humidity before testing, so the thorough dry out is probably pointless.
But ionic crud could have contaminated the surfaces before the hand soldering added a layer of rosin. I had to remove the rosin before the crud could be dissolved off with water.
I'll second that. It's nasty, greasy, dirty stuff, more like "can't clean" flux. But the no-clean crud I've used doesn't seem to conduct.
Organic solvent loaded with a bit of rosin flux seems to be safe, if not very cosmetic. Rosin residues don't seem to conduct. In our in-house vapor degreaser, we submerge the board for a while in a boiling solvent that has a deflux agent in it, then spray it down with a wand that pumps clean, freshly-distilled solvent. The solvent wets the pcb and parts, unlike water which beads.
We'd just need a distilled-water fire hydrant. I considered using my big ole 1500 psi pressure washer, but it would use tap water, too. Hence the dental water-pic idea, fillable with distilled or deionized water, whichever ohms highest.
I just filled my red plastic Official Presidential Drinking Cup with local tap water, and poked in a pair of standard Fluke probes 1" apart, and got about 1.2 Mohms, not too bad. Same test with water from the office cooler, Alhambra Mountain Spring Water, is about half that.
Dude, man up! If the boards don't meet your spec, they're rejects. Send the boards back and tell the stuffer they don't get paid until the boards are right, or make them pay for the rework.
We own the boards and the parts, and customers are waiting for delivery. The only thing we can do to the vendor is not pay them for assembly and never use them again. Meanwhile, the problem has to be fixed.
Last (and only) time I had this problem, cleaning the boards using solvent scrub in the critical area worked. The crud was quite difficult to remove, a simply solvent wash was not enough. But in our case, I think it was on the surface-- a nasty hydrophilic (?) layer of something, rather than underneath the parts. Try axing the chemical manufacturers, they are generally pretty good (at least I've had good luck with Kester).
Best regards, Spehro Pefhany
--
"it\'s the network..." "The Journey is the reward"
speff@interlog.com Info for manufacturers: http://www.trexon.com
Embedded software/hardware/analog Info for designers: http://www.speff.com
You need distilled water only for the final spray rinse to displace the film of tap water. Tap water is fine to remove the vast majority of the contamination.
I've had reasonable luck with an ultrasonic tank filled with the appropriate solvent. I'd probably wash in a mild soap and water solution followed by a rinse in distilled.
Here are some posts I collected from Bob Wilson on cleaning flux from pcb's. His comment on activator salts leaving a white residue being dynamite is absolutely correct. I ruined a perfectly good dvm by trying to clean it after it started acting up in high humididy.
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Subject: Re: Flux clean > Dave Rosenbloom wrote:
Never mind fooling around with stuff whose solvency you don't understand. Partial or incorrect removal of resin based flux is FAR worse than doing nothing at all.
Resin fluxes require both polar and nonpolar solvents to remove all components completely. Using a bit of solvent (even the correct blend) and a brush, is REALLY bad news. What is mainly does is just spread the flux around all over the place.
There are 2 main ingredients in resin fluxes. First, there is the resin itself. The active ingredient in resin is Abetic Acid. Resin requires a nonpolar sovent to remove it (such as trichloroethane, or even toluene).
The other ingredient is the activator salts (typically chlorides and fluorides). These are soluable in a polar solvent ONLY (such as isopropanol, or even water). They are ABSOLUTELY UNAFFECTED by the nonpolar solvents that will dissolve the resin component.
Attempting to use Disc Brake cleaner, or any other aggressive NON-polar solvent will simply remove the protective resin (that previously and harmlessly encapsulated the hygroscopic and conductive activator salts), and expose these salts to the air. This results in high impedance conductive paths all over the PCB, and possible corrosion.
You cannot see the activator salts, in most cases, but in higher concentrations they do appear as a slight white residue. This is dynamite!
So before anyone uses whatever snake oil thay happen to find laying around, it is wise to understand the chemistry that underlies this operation.
For years an excellent flux remover was a 30% mixture of isopropyl alcohol, with the remaining 70% being trichloroethane or freon. Since both trichloroethane and freon are no longer easily available, a good substitute is plain ordinary "lacquer thinner" (mainly toluene). This mixture will remove ALL parts of the flux. Pure 99% Isopropanol (aka "rubbing alcohol") WILL work as well, since it DOES dissolve both polar and non polar residues, BUT it is extremely slow to dissolve the non-polar stuff (resin). The above mixture acts much faster,and still does not harm most components (polystyrene caps are the exception, but they dissolve in nearly anything anyway).
The other alternative is to remove the resin based flux with a proprietry Saponifier in a water solution (Kester makes one). This requires very strong agitation, and hot water/saponifier solution.
Finally, one can just use water-soluable-flux solder. Some brands so not have very good "fluxing" action, although I have had good luck with the stuff made by Alpha Metals. Warm, HIGH PRESSURE water spray (to get under ICs and so on) is a must to remove this. One good way is to stick the PCBs in a dishwasher. Personally, I use water soluable flux wherever possible. It works well and DOES remove completely. One final point: you must ALWAYS clean water soluable flux off the PCB. NEVER leave it on there. It is extremely hygrosopic and will result is a malfunctioning PCB after it has absorbed atmospheric water (1 to 4 weeks later).
Bpb.
You need a solvent blend that contains both ionic and non-ionic solvents. A very good blend is 70% trichloroethane and 30% isopropanol (isopropyl, or "rubbing" alcohol). The trichlorethane is an aggrressive non-ionic solvent and the isopropanol attacks the ionic salts.
Although isopropanol is a drugstore item, trichloroethane is not. If you have a problem with this, xylene or toluene (or even ordinary lacquer) can be used instead, although these are more aggressive solvents and some parts may be affected (test first). Increasing the proportions to 50-50 will reduce the aggressiveness of the blend.
Even straight isopropanol can be used, since it can dissolve bothe ionic and non-ionic residues. However it is a very feeble non-ionic solvent, so it will take a long time to attach the resin and other non-ionic residues.
Be sure to immerse the PCB in a reasonable volume of solvent, and rinse in a fresh bath. **Use lots!!** DO NOT JUST SWAB IT AROUND WITH A COTTON SWAB!!! All this will do is to spread the flux all over the place, and partially dissolve the rosin, exposing the activator salts that were previously trapped harmlessly. EITHER CLEAN THE PCB PROPERLY OR NOT AT ALL!
BOb.
Naptha would work well if the crud were purely non-ionic, however I suggested the Full Monte cleaning because the true makup of the crud was not known, and also because there will CERTAINLY be some small amount of flux residue if the motherboard were made in the orient (which most are).
You are correct that there is no rule that a solvent mix must be used. Sequential cleaning (as long as the rosin is removed first to expose the activators) is perfectly acceptable.
I agree with Bob. Most of the flux is no clean today. Why use an extra processing step (washing) if you don't need to. Especially when you have some non-washable parts on the board.
Now, I use a turpine based solvent that:
1) Smells great (it is orange peeling oil really) 2) Is predominantly non-toxic and non polluting. 3) Most importantly, does a great job in cleaning the boards.
The only drawback is that it takes a long time (when compared to freon) to evaporate. If you use compressed air to get the majority of it off, then it will dry in a few minutes. It does a great job and if used in liberal quantities and some scrubbing, leaves no residue. That cannot usually be said about acohols and freon based solvents.
Chuck
I agree, that clean water wash is good if all the components are sealed. Another method that does not require careful drying is to use a turpine (orange/lemon peeling oil) based solvent. It smells good, cleans good, takes a while to evaporate (I suggest using compressed air) and leaves very little or no residue. It will eat styrene but that is the only plastic that I have found that it doesn't like. BTW never use any type of alcohol on poly carbonate components. Like the plugs on the ends of phone cords.
I would guess not, We had a simlar problem with a crystal and residue under it from the water cleaning process.
They all seem to be aware of the problem
We tried a alcohol wash after the fact, and the crap was still under the crystal. It seem that once it dries its very tuff to wash out. John P's wash and rinse might work. I would recomend a Milled slot under the part in the next design, similar to HV stuff.
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here.
All logos and trade names are the property of their respective owners.