[chbot] Wanting real data - How best to mount sensors outside in the elements?

[Mark Atherton]

Wow Stephen, thank you. This is fascinating.

I have seen some commercial products that had a vent at the bottom, 
allowing water to escape and conformally coat the assembly inside, to 
keep the electrics safe.

Military product milled out of block aluminium, with O ring seals and 
hermetically sealed connectors suddenly doesn't seem quite as much 
overkill as it did once.

-Mark


On 29/05/2020 9:13 PM, Stephen Irons wrote:
> Waterproofing is hard, really, really hard, for housings fully exposed 
> to the weather.
>
>   * IP-codes (IP-65, etc) don't mean what you think they mean [1].
>   * Common IP-65 rated housings suck in water that collects in the
>     groove between base and lid as the internal pressure changes with
>     temperature.
>   * PTFE pressure-relief vents prevent this, as long as water does not
>     pool on the vent mebrane. PTFE is the magic stuff in outdoor
>     clothing that claims to keep the water out, yet lets your skin
>     breathe.
>   * PTFE pressure-relief vents are permeable to gas, including
>     water-vapour. The humidity inside the housing will approach the
>     ambient humidity in about a week [2].
>   * All polymers (including ABS and PC) are permeable to gases,
>     including water-vapour [3]. The humidity inside the housing will
>     approach the ambient humidity in about a month [2].
>   * If at all possible, use a potting compound to exclude gas. This
>     solves the pressure-change problem, and many epoxies and
>     polyurethanes are waterproof (polyester resins were not quite
>     waterproof enough, as older boats with osmosis blisters show).
>       o ...but you need to find a compound that will stick to the
>         insulation of all of the cables that penetrate the potting.
>         Some cable jackets are made of polymers that resins don't likt
>         to stick to.
>   * Otherwise, install the housing inside a cabinet sheltered from sun
>     and rain.
>   * Those end-user replacable batteries are also in a waterproof
>     housing, and have exactly the same problems.
>   * It is really tricky to seal the battery-compartment lid properly.
>     Too loose, and water gets in. Too tight, and the lid bends and
>     water gets in.
>   * Waterproof connectors have the same issues.
>   * Multi-core cables and with braided screens have voids between the
>     strands. Capilliary action will draw water into these voids,
>     corroding the conductors. If this it not enough, they are subject
>     to the same pressure variations as the main housing which helps to
>     draw water into the cable.
>
> [1] IP-codes apply to the product as a whole, not to the housing. 
> Water might be acceptable inside an IP-67 or IP-68 rated product, as 
> long as the product still works. Obviously, it will be easier to make 
> a water-resistant product using a 'waterproof' housing.
>
> [2] IP-65 rated ABS and PC housings from Jaycar, with Bluetooth 
> environment sensors sealed in them, installed in a 100% humidity 
> chamber (a plastic crate with a bowl of water) with its own Bluetooth 
> environment sensor to monitor the humidity chamber itself. 
> Temperature, pressure and humidity were measured hourly for 2 months. 
> The humidity chamber was indoors, but not temperature controlled so it 
> was subject to daily temperature variation.
>
> After a few weeks, it is advisable to get the intern to open the 
> humdity chamber, preferably somewhere far away from me.
>
> We tested 6 configurations, each with a Bluetooth environment sensor 
> sealed inside:
>
>   * ABS and PC housing screwed shut with the supplied seal and screws.
>   * ABS and PC housing screwed shut with the supplied seal and screws,
>     with PTFE vent installed. PTFE vent on a vertical edge to avoid
>     water pooling.
>   * ABS and PC housing epoxied shut.
>
> [3] PDF file with some gas permeability of various polymers at 
> https://www.sciencedirect.com/science/article/pii/B9781884207976500024/pdf
>
> Good luck!
>
> Stephen Irons
>
>
> On Thu, May 28, 2020 at 21:14, Andrew Sands <andrew at theatrix.org.nz> 
> wrote:
>> Hey all list dwellers, I'm looking for advise, suggestions and 
>> general pointers to what other have done in order to get various 
>> types of sensors working out in the real world - that is beyond the 
>> bench. What sensors perform better / marginal, what methods provide 
>> the best mechanical / environmental protection. Where should I 
>> position say an outdoors temperature sensor? Mostly I'm after less of 
>> the theory and more of the I tried this but it mutated a spider which 
>> ate my cat kind of thing. So don't do that. Thanks for reading, stay 
>> safe. Wash your hands. Regards, Andrew

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.ourshack.com/pipermail/chchrobotics/attachments/20200530/a4aaf13a/attachment.html>