Pcb conformal coatings

[Hendry Born]

Does anybody have suggestions for using formal coatings in switchable power supplies

What are the major thermal disadvantages?
Durability of the conformal?

Any thoughts on the UC40-250, up and downside


Thank you for contributing

[T3sl4co1l]

TFW I wear formal coating but the customer only wanted informal...

(The correct term is "conformal", and no, I don't have anything useful to contribute, just saw the joke and took it)

Tim

[Hendry Born]

Thank you for the correction!

[elecman14]

I have worked with conformal coat in the past. My suggestion is if at all possible to avoid it. The application is difficult, expensive and hard to maintain quality. While I do not have any experience with UC40-250, I may be able to point out a few things to consider.

• Getting the conformal coat to adhere to all the components can sometimes be difficult depending on their cleanliness. You may need to wash and dry the PCB before coating.
• Another gotcha with UV curable coating is the UV light source probably should be validated somehow. If the light source looses intensity or goes off you could end up with not a full cure. The plants I was working in used a probe run through the curing "oven" (light source). It looked kind of similar to what you would send through a reflow oven to check the thermal profile.
• Getting a coating with UV traceable florescence dye can help show board coverage problems when put them under a black light.
• Getting a coating that can be removed chemically is super helpful if you need to do failure analysis on any parts. Otherwise you end up having to remove the coating mechanically to probe. That is not fun at all and usually ends with this followed by this 

Good luck...

[Hendry Born]

Thank you Elecman14

Another issue raised my mind, high components like inductors and electrolytic capacitors... mmm   

My experience until now is that suppliers of coating and suppliers of coating machines can not provide so much insights, trail and error is the tip of the day 

[elecman14]

The inductor and capacitor issue really depends on what your customer wants. Forgot to mention in my previous post you may want to get a copy of IPC-HDBK-830 and maybe IPC-A-610E (IPC-A-610E for section ten point eight showing defects) it is quite handy and may answer some of your questions. Again good luck, just be glad you are not doing potting  ... it could be worse

[Hendry Born]

Thank you elecman14, I will look into that for more info

[uncle_bob]

Hi

When you talk about coatings / electronics / clean you immediately get into all sorts of corner cases. "No clean" flux is not no clean when you use a coating. Parts that use coatings in there construction may / may not be compatible with your coating process. Anything that as ever been near a silicone based mould release is not likely to work with a coating process. Some of this can be corrected with a cleaning process, some only by buying a different part.

Cleaning used to be real easy. You grab XXXXX (now banned) and slap the part in a vapor degreaser. All the bad stuff winds up in the sludge in the bottom of the tank. You toss the tank full of sludge in the local waterway every so often (no not really, you send it off to the reclamation company). These days cleaning is a closed loop system and not at all cheap. What you use depends a lot on who you buy your system from.

After you get the boards clean, you need to get them dry. A board that bubbles up in the hot summer heat ... not at all good. The bottom line is that you need a bake at > 125C. Again your parts will dictate how long a bake. A day may be enough for some. A week may not be enough for others. Once the parts come out of bake, you need to coat them fairly soon. If you don't they will soak up moisture from the air. How soon? Again that depends on your parts.

Doing the coating is pretty simple. Anything worth using is a two part mix. It's got a dye in it. You spray it in a spray booth with all the vents and protection. The stuff has a pot life after you mix it. You have to get absolutely everything clean before any of it dries on any of your gear. There is a cure process for the stuff. Heat is a pretty typical way to accelerate the process.

If you really care about the result, plan on running everything through twice. Yes you need to pick a material that will work with that sort of process. You need that anyway, you *will* have voids. After the two coats, you run them past the black light and check for the nice glow on all the surfaces. Any that don't glow get re-coated.

Fun stuff.

Now you have aboard that is an absolute horror to do any sort of work on. It's not humidity proof. It simply is more immune to dirt / moisture / crud than it would be without the coating. That helps protect the parts. ... errr ... except that it really was designed for leaded parts. With SMT stuff, you trap gunk under the parts. With SMT, if the coating shrinks, you can tear the parts off the board.

Yes, there are solutions to all those problems. They very quickly get into the "do you have $1M to spend?" sort of region.

Bob

[station240]

I've got a board in my junk box that has some sort of white glue around the base of the big electros, then a thin lacquer like coating to keep the elements out. The only minor problem is you forget it's there, and go to probe pins with the multimeter and get funny readings (eg open circuit). I get the feeling it's a spray gun application.

I wonder if there are places that can clean the board and apply the coating in exchange for money.

[uncle_bob]

I've got a board in my junk box that has some sort of white glue around the base of the big electros, then a thin lacquer like coating to keep the elements out. The only minor problem is you forget it's there, and go to probe pins with the multimeter and get funny readings (eg open circuit). I get the feeling it's a spray gun application.

I wonder if there are places that can clean the board and apply the coating in exchange for money.

Hi

Once you put a proper conformal coat on, there is no getting it off. Think of it as epoxy (it's close chemically). That is the thing that makes it even less attractive once you get the job done. Going through the coating is a mechanical process. Replacing a part involves scorched mask. To maintain the performance of the mask, you need to repair it each time you mess with it.

Bob

[Hendry Born]

Thank you all for your sharing your knowledge.

Does anyone have experience with the thermal effects of the coating over time I.e. dust collecting on the pcb?

[SeanB]

Dust does collect in the usual spaces, like crevices and corners. However as you do not really want to conformal coat a heatsink, mostly because it then stops working unless you have really massively overspecced it, or it is only a heat transfer plate to a chassis or other heatsink, using a lot of thermal interface material and fasteners, it will still be an effective dust trap. Thus most boards with the need for active cooling use a fan, and either use a very good dust filter on the air inlet and outlet ( surprising how much will come back in there when power is off), or use a hermetic case with cooling done by circulating internal air with a fan and a finned exterior so there is no ingress.

The only good thing about conformal coating is that cleaning dust off is very easy, as it tends not to stick, though if you get conductive dust it does really nasty things in connectors, which is the major drive for hermetic cases along with conformal coat.

As to the coatings, try to aim for a solder through coating if you ever intend the boards to be repairable, as it does make the life easier for the poor guy doing rework, he only has poisonous fumes ( in addition to the flux) to deal with, not fumes and noxious chemical removers that do a really good job of removing the epoxy coat, the labels on components, the underlying board material, the top of the bench, the floor and your skin and nose as you try to clean off the small area you want to work on. If you want a good coating get a flexible conformal coat, as it at least is probable, and it does reduce vibrational stress on components, so can help you pass MIL shock and vibration testing which can quite easily shake components ( like larger SMD capacitors and inductors, along with anything tall and skinny) off the board or snap the leads.

[Hendry Born]

Thank you SeanB.

Tomorrow I shall post in PM some images to show you what is happening 

It is real issue the manufacturers are also involved. 

[uncle_bob]

Thank you all for your sharing your knowledge.

Does anyone have experience with the thermal effects of the coating over time I.e. dust collecting on the pcb?

Hi

Conformal coat on a board (the good kind) will have no problem being dunked in a bucket of mud for a week. Spray it off and all is well. There are an empire of materials that claim to be conformal coat that will go nuts if you get them anywhere near dirt and water. The easier they are to work with and repair ... the less they protect.

The barrier any of these provide is not perfect. Moisture *can* get through. It just takes a while. Once it gets through, getting it back out again is essentially impossible. It's sort of the Hotel California for humidity . It that a problem? It depends a lot on why you are doing any of this in the first place. Nobody does conformal coat for fun. There always is a spec driver that forces you to do it. If you need to pass salt spray / humidity / leakage after coat (and all the other nine yards of testing) ... you likely need some pretty good stuff.

All epoxy like materials tend to shrink a bit over time. Your FR-4 board shrinks a bit. It's a function of the cross linking in the "stuff" that makes it all up. That can give you issues with very small SMT parts and big globs of coating. Simple answer -- you can go overboard with this stuff. Next not at all simple answer -- put some in a chamber at 125 for a couple months. Then thermo-cycle them and see if there's an issue.

Bob

[cat87]

Like someone else already said, for conformal coating, cleanliness is  critical. The board / boards should be washed in  a tank of solvent for at least an hour before the coating process is started.
While I haven't used the UC40-250 you mention, I want to mention that I have seen some water-soluble coatings (water-soluble while the coating is not dry  is what it actually means). These offer long term protection against moisture in the air seeping in, but that's kind of it. The up-side is that if you ever want to rework the board, it can easily be removed mechanically (it can easily flake off) or with some solvent like acetone or IPA.

One of the toughest coatings I've seen is a polyurethane coating, providing both moisture and  physical protection to the board. Let me tell you, once it dries, that's kind of it....it won't come off without a fight.

[uncle_bob]

Like someone else already said, for conformal coating, cleanliness is  critical. The board / boards should be washed in  a tank of solvent for at least an hour before the coating process is started.
While I haven't used the UC40-250 you mention, I want to mention that I have seen some water-soluble coatings (water-soluble while the coating is not dry  is what it actually means). These offer long term protection against moisture in the air seeping in, but that's kind of it. The up-side is that if you ever want to rework the board, it can easily be removed mechanically (it can easily flake off) or with some solvent like acetone or IPA.

One of the toughest coatings I've seen is a polyurethane coating, providing both moisture and  physical protection to the board. Let me tell you, once it dries, that's kind of it....it won't come off without a fight.

Hi

If you run humidity testing and *really* want a coating that does you some good, the Urethanes are about the only "paint on" coatings worth using. They are not perfect, but they at least do you *some* good. Everything else I have tried is really not worth the effort for humidity.

Bob

[SeanB]

I had boards coated in a coating called Mesopoulet K, which was pretty much resistant to almost anything. Dropping a board in an ultrasonic cleaner ( 400W) using boiling trichlorethylene as a solvent for 20 minutes did not remove it, though it did totally dissolve the FR4 board underneath, leaving the components on copper traces held by the encapsulant, with fibreglass chaff between the layers and the vias holding them together. Only way was to use a scalpel to abrade it off the pads after burning through the layer with a soldering iron ( with open window and a fan blowing the fumes out) then chip the IC off the board so as not to damage the traces under the chip or the copper.

Recoating was easy, clean and sand the existing coat and wipe with acetone, then mix the exact ratio of the 2 components, and use within the 2 minute pot life. Also the shelf life of the unopened packs is 2 years, and once opened 6 months. Also only came in 1l packs, and if you think printer ink is expensive........