Waterproofing

[FlyingSquirrel]

hello guys and gals!

ive been looking into waterproofing some drone electronics and so far only figured this might be harder than it seemed at first
so i was hoping that some one here might be able to help me, without ending up in too many rabbit holes
and maybe ill learn one or two things without being spoiled by half-true "facts" 

the requirements include:
- long lasting waterproofing (more than just 1-2 months)
- 100% water protection up to low depths (in case it ends up in a pond or along the coast)
- vibration/crash resilient (considering speeds of up to 90-100km/h)
- low weight (which is why i havent considered general potting)
- maintainable, as in being able to resolder things without having to re-coat every thing
- allow heat exchange (since the motor speed controllers can get quite warm/hot)
- protect the camera housing without having to coat the electronics and endanger the CCD chip
- protect USB, JST, XT60 and board-to-board connectors including LiPo battery

the drones operate at 5-24V with up to 140A (might be more as LiPo's become better)
and primarily consist of the following:
- power distribution board incl. LiPo
- flight controller (which can include barometers)
- 1 speed controller per motor incl. motors (which should already be waterproof)
- CCD camera (with housing)
- a radio and separate video transmitter (with 2 pigtail and 1 cloverleaf antennas)
- also possible but not always present are an exposed microphone and buzzer

so far i came across the following:
- general purpose conformal PCB coating (didnt last very long as it started to crack)
- silicone conformal coating (heard from several guys that it also didnt last very long)
- nail varnish (skeptic)
- corrosion-x (suspect this will only prevent corrosion)
- nanoprotech (which made me think i might need to learn some more) 


regarding the last product https://youtu.be/EzEreJfRp8Y (to not link to the distributor directly, in case its phony)
i couldnt fully grasp, how it could protect from short circuits but still allow a current flow where it had to be
or is this maybe an exaggerated marketing gig?

can you help me out, or have a favorite solution/product you know and works in such or similar application?

thank you!

[james_s]

I know guys who use a product called CorrosionX and report good luck with it. It's virtually impossible to satisfy all of your requirements though, by far the most effective method is to stay away from water. Foam fixed wing planes will float, multirotors don't.

[Ian.M]

I'd like to see that demo after a pound of salt has been stirred into the tank!
Ultrapure Water has a resistivity that can be over 18 Meg·cm

[JanJansen]

I would not try that at home with a drill in bath.
edit : ok, the movie also says not to try home disclaimer.

curious : how many times do you have to spray ?, 1 time or everyday ?, why dont they use that default on everything ?

[CatalinaWOW]

Check the radio control plane websites.  The guys that fly float planes have been dealing with this for years.  As I recall the solutions generally involve silicon grease to prevent water intrusion in servos and connectors and baggies to protect receivers and ESCs.  RTV where the wires enter and leave bags.  You have more elx and wires associated with FPV, but the same principals should apply.

[FlyingSquirrel]

cheers for all the inputs so far!

@james_s - im aware ill probably not find THE 1 product to cover every little thing, but was/am hoping to find a better solution than corrosionx
obviously its better not to go and bathe, but unfortunately its not always inevitable - considering we often get rain/snow here and that stuff can also take weeks to dry & if id do wait i might only get 2-4 weeks of flying per year

@JanJansen - with the RC version of the product they claim it will give full protection for 1 year and possibly more, not sure about the amount of layers required though (assuming 1)

@CatalinaWOW - ill give that area some more research, i kinda doubt the bagging solution will work for the power distribution board and speed controllers very well,  since the point of failure is usually where the cables penetrate and thats quite a few specially since just heat shrinking already requires things to be un/re-soldered etc + it adds quite a bit of weight

a little side question; lets assume id apply a layer of WD40 to some electronic components - how would i go about calculating the insulation properties in different types of water? or would i be better off experimenting with a scrap PCB and see where it catches fire?
maybe it will become easier if i fully understand what parameters im dealing with

[Codebird]

From personal experience: Not ABS-in-acetone or model-painting lacquer. They both peel off easily.

I had middling results with candle wax, but I expect a product purpose-made for the application will be far superior to any improvised substance.

[mtdoc]

Has anyone ever tried using polyester resin - as in what is used for fiber-glassing and gel-coating boats, etc in a marine environment?

I've used it for surfboard repair for years and it's pretty easy to work with.  It seems like brushing a thin layer on a pcb might work well for a permanent, light weight waterproofing. Of course later repair or rework would would be a problem.

I don't know what it's electrical conductivity is but I'd think it would be very low. Thermal conductivity is likely high enough to allow adequate heat exchange with a thin layer.

[CatalinaWOW]

When designing military electronics we never had much success waterproofing at the board level.  Even though we did coat boards with stuff you would never want to deal with at home.  Things like parylene.  The only types that we used there that you might be interested in are the two part polyurethane varieties.  They don't come from the dime store.  No matter how good the coating is there have to be places to let the electrons through, and there are almost always pinholes and voids.  So to keep water out the electronics compartment was sealed. 

You might want to think of options along those lines.  Use food service wrap and a heat gun to seal the whole thing, except for the connector you bring out for battery charging.  Silicon grease at the motors.  It will be a PITA but TAANSTAAFL.

The other thing I recall from the RC guys sealing efforts is that they don't need, or intend to have long duration sealing.  If you are going to crash your drone in a lake and then go pick it up next week you are out of luck.  The sealing is intended to keep casual water out for the few minutes or maybe tens of minutes until recovery can be effected.  When a crash in the water occurs dis-assembly and drying are the rule.

[DTJ]

Despite their absolutely shite website these guys sell a great PCB sealer, its clear and easy to use.

http://www.sensongroup.com/senson-tek/products


You brush or spray it on as thinly or as thickly as you want. It bonds really well and does not peel.
It seems to last indefinitely. It can be soldered through or you can quickly strip it from the boards using methylated spirits / IPA.

Maybe it would work for what you need.

[JanJansen]

You can always try it on a Hobbyking drone.

[David Hess]

My experience with waterproof conformal coatings is that surface preparation is very important but the only good experiences I have had were with vacuum potting and I am not sure how you could combine the two economically.  If you had a large enough vacuum chamber to dunk the individual boards into the conformable coating, then I expect that it would work very well.

I wonder if a heat shrink film could be used to bag the boards.

[FlyingSquirrel]

cheers for the inputs every one!

@CatalinaWOW - with the resources that the military has, it does give me quite a respect for the topic, regarding they even struggle with it
wont the food wrap create electrostatic charges that zapp the components? and yes, id rather not use chemicals that will kill every fish in the pond i accidentally land in but ill give the polyurethane compounds some researching
also no, the idea is to have it survive tens of minutes up to maybe an hour and a nice to have would be able to dry it relatively quickly, so i dont necessarily have to leave home directly, in case something happens at a spot that took me half a day or more to reach

@DTJ - companies that concentrate in their core business usually tend to have ugly websites (so that can also be a good sign)
is it worth spending almost 1k on one of their kits, or would you recommend a specific product?

@JanJansen - id rather not support hobbyking even if it will cost me twice as much
i could also get a cheap 50-100buck drone from the next store, but it also wouldnt be a reliable test comparing a <10W with a >2000W drone

@David - wont the vacuum just get rid of bubbles? and wouldnt vibrations achieve the same thing? regarding there woudlnt be several centimeters of coating

[David Hess]

@David - wont the vacuum just get rid of bubbles? and wouldnt vibrations achieve the same thing? regarding there woudlnt be several centimeters of coating

It gets rid of the bubbles but it also forces the encapsulant into the surfaces of the board and parts on the board with 14.7psi when the vacuum is removed.  The problem which you identified is that you need to dunk the board to apply the encapsulant.  Actually you could brush or spray it in a vacuum but that is beyond anything simple.

If you went to this much effort, then I would also thoroughly clean and bake out the board before doing the vacuum encapsulation process.

[CatalinaWOW]

Remember (particularly when dunking boards) to mask/plug/isolate all connectors, sockets, trimmers, test points, switches, relays and the like.  It is remarkable how effective conformal coating is at eliminating the function of these devices, even when it is not successful at keeping moisture out. 

If you are coating under a vacuum be aware that some large plastic IC packages actually have a vent port in them and can have strange behavior when filled with conformal coating.  Vacuum makes the masking discussed in paragraph one much more difficult.

Troubleshooting these problems is sometimes easy, but often amazingly difficult.

Red Squirrel - I don't think the ESD problems from using food service wrap would be any worse that those caused by the plastic bodies these things have.  I have only limited experience using this wrap around electronics, but have never had a problem and never drawn a spark with it.  I suggested it because a single layer would be waterproof (barring scrapes and the like) and would weigh virtually nothing.

There is also virtue in a layered defense.  For this type of application multiple pretty good layers or methods could easily be better than one really good one.

[DTJ]

cheers for the inputs every one!


@DTJ - companies that concentrate in their core business usually tend to have ugly websites (so that can also be a good sign)
is it worth spending almost 1k on one of their kits, or would you recommend a specific product?

You just need a bottle of the sealant. Probably $30 or so. Freight will probably cost you more than the sealant. Pic attached.
There must be a similar product on your country.

Regarding the food wrap - I'd be wary of that a voltage potential could be generated as the wrap is unwrapped off the roll or off the device.

[james_s]

If you are coating under a vacuum be aware that some large plastic IC packages actually have a vent port in them and can have strange behavior when filled with conformal coating.  Vacuum makes the masking discussed in paragraph one much more difficult.

Really? Where does it vent from? The bare die? Where is it and what kind of packages? I'm not challenging your statement, it's just not something I've ever been aware of. That is, aside from things like barometric sensors, speakers and microphones that obviously need an opening.

[CatalinaWOW]

I've forgotten the specifics, but I believe it was a large Cyprus gate array that we encountered.  The port is filled with a grease like substance that matches the package visually.  I can only speculate as to why the port was there, it certainly wasn't for measuring atmospheric quantities.  One of those fun things you find in a production program that has stopped working. 

[SeanB]

That would be a part in a cheaper plastic cap over BGA, which will have a void in it and a vent. 

Would be careful on that Senson website, it seems to have some very dodgy links in the pages, pointing to payday loans and the such. Might be a good idea not to order online there for a while till that is fixed, perhaps a phone call would be safer to find a local stockist.

however, as a user of Parylene and Mesoupolet K, which are really good PCB conformal coats, they really require careful preparation of the board. At a minimum a thorough wash in solvent, a scrub all over, followed by a rinse, then baking for a few hours at elevated temperature( and there the MIL 125C is good, as you do not melt things in the 120C oven), followed by a careful brushing on of multiple coats of the conformal coat, following the directions as to mixing to the letter, a clean brush each time ( brand new, washed a bit in water then solvent and then oven dried, preferably a non painted brush so the paint does not run) and removal of all bubbles, then the baking after final coat. Not fun is the coatings are very aggressive chemicals till mixed, and have a very short shelf life unopened, and an even shorter open pack life, and a really short pot life. however they do work, and in most cases are solder through with good isocyanate fume extraction, or a 120mm fan and a window in our shop.

Interesting thing about the Mesopoulet K is that you could drop a board with it as a coat into the boiling TCE ultrasonic cleaner, and leave it there for a half hour, and the coating would be mostly fine and undamaged, though in the interim the majority of the plastic parts ( capacitors, sleeves, the epoxy the board contained, some IC packages and any plastic parts of the board connectors) would either have swollen and burst, or were completely gone. Was very good at stripping the lettering off packages, and cleaned any chassis down a treat, no matter what coated it dirt or paint wise.

[FlyingSquirrel]

thanks a million for all the inputs every one!

think ill try conducting several tests with different ideas
specially applying multiple layers and also see how i can integrate gelatinous sticky tape, to protect the coating against stress without letting the tape cause the stress since the tape i tested seems to be very resilient to heat (even with 100ºC above my normal-high soldering temperature, right next to the tape)
ill also look into the food wrap as an additional coating, since ESD should also be less of an issue on already coated components

and for now i think ill keep the vacuuming in the back of my head, regarding there are several components that probably wont be compatible with this solution
better to ie loose a button to water and replace it, rather than having it inoperable from the start

ps: a little discovery i made (in case it helps some one)
when you coat something, then heat shrink it, you can apply drop by drop of additional coating along the opening/s of the heat shrink
and the coating will wick up throughout the inside of the heat shrink
depending on the size of the heat shrink etc it might not wick through every little corner (maybe vacuuming might improve this)
but it will probably reduce/prevent moisture/water from collecting under the heat shrink

[Assafl]

In the navy active sonar transducers are waterproofed (they are usually heavily pre-loaded piezo ceramic rings and cope with over 1000V excitation voltage). One of the major ingress points are cables. So when you epoxy pot or conformally coat a board water will creep along the cable under the coating.

So one way to cope with this was to make O-rings of a compatible substance to the potting compound, and put 1-2 orings on each cable near the board. The Orings seals tightly against the cable (preventing creeping) while the other side of the O-ring bonds with the potting compound.

It is somewhat tricky as you want the potting compound to bond with the O-ring but don't want it to attack or melt or cause the O-ring to swell. I do not remember what was used by the navy...

+1 on board prep before conformal coating. I have never seen conformal coating crack before.... maybe there is too much vibration and needs a shock mount?

As I see your needs you have two options:
1. Prevent water ingress at any cost - very hard to do!
2. Allow some water to ingress - but also enable it to exit quickly without leaving damage.

Salt water makes the second choice harder - but using Teflon cable and sealing all the connections (better - solder them and seal) should harden it for salt water.

One way to prevent damage is to expose contacts that sense water and disengage a "power disconnect" relay completely disconnecting the battery packs - so that the wet electronics don't fry (no power source) and you can clean the salt water away quickly with IPA and deionized water.

[DTJ]

One of the major ingress points are cables. So when you epoxy pot or conformally coat a board water will creep along the cable under the coating.

So one way to cope with this was to make O-rings of a compatible substance to the potting compound, and put 1-2 orings on each cable near the board. The Orings seals tightly against the cable (preventing creeping) while the other side of the O-ring bonds with the potting compound.

It is somewhat tricky as you want the potting compound to bond with the O-ring but don't want it to attack or melt or cause the O-ring to swell. I do not remember what was used by the navy...

That's very good idea Assafl. I'll store that one away for when I need it.

[Ian.M]

One way to prevent damage is to expose contacts that sense water and disengage a "power disconnect" relay completely disconnecting the battery packs - so that the wet electronics don't fry (no power source) and you can clean the salt water away quickly with IPA and deionized water.

Its for an aerial drone - water sensing and immediate power disconnect is a horrible idea as one raindrop in the wrong place and it will crash hard.

OTOH if there's a telemetry link, sending back an alarm condition for water detected would be smart.  Also if water is detected,  cutting power after one second of zero throttle, or if its possible to detect if the FCB has crashed (does it have a heartbeat signal, or LED?) or on impact (shock sensor?), would go a long way towards keeping the electronics and motors in salvageable condition, especially if combined with a hydrophobic film coating from a reputable manufacturer.

[FlyingSquirrel]

ill have a look at the orings, hoping ill find some for every gauge of wire
what ill try before though, is greasing the borders of the coating - since what i know from diving, is that an oring is only as effective as its integrity, which usually requires it to be treated, so that it A doesnt become brittle and B allow a small layer of grease to completely close the gaps that may occur during temporary expansions of the oring

and i have to agree with Ian there, although false positives could be minimized with some tricks, the solenoids that can cope with the required amperages are as heavy, if not heavier as the entire drone

[Assafl]

ill have a look at the orings, hoping ill find some for every gauge of wire
what ill try before though, is greasing the borders of the coating - since what i know from diving, is that an oring is only as effective as its integrity, which usually requires it to be treated, so that it A doesnt become brittle and B allow a small layer of grease to completely close the gaps that may occur during temporary expansions of the oring

and i have to agree with Ian there, although false positives could be minimized with some tricks, the solenoids that can cope with the required amperages are as heavy, if not heavier as the entire drone

As I stated Oring compatibility is crucial. But companies like Raychem makes these types of system. The wrong lube and your Oring will swell and lose sealing properties.

Also, the disconnect doesn't have to be a contactor - but the idea is to limit the current draw that would destroy sensitive electronics. So perhaps it is not possible to disconnect the motor and driver - but just the entire control electronics. Of course control signals have to be opto isolated.