DIY SQM - Water ingress damage

[TimBrandt]

Hi, all, I'm after some advice,please.  I recently built a version of Rob Brown's DIY SQM.  I mounted the main electronics box in my observatory, with the main sensors in a modified weather station cover outside, on the end of a 1m cable.  The sensors I used were a BME280 pressure and temp sensor, an MLX90614 to measure sky temperature and a TSL2591 to measure lux.  The cover is a TFA Dostman weather station cover which stops direct rain and sun, but allows airflow to the sensors.  The TSL2991 is protected under a plastic lens, and the IR thermometer can is sealed and itself waterproof, but the rest of the board is not.  The sensors are connected by STEMM QT connectors.

After 3 weeks, I have extensive water damage to the sensor boards, with corroded components.  The problem is humidity, and the boards are wet at least some of the time.  I had expected some problems from this, as I live in the UK, but the damage has been greater/faster than I anticipated.  My question is what I can do to fix this in a rebuild?  I can replace the QT connectors with soldered connections, but what can I do to protect the boards?  Potting seems an obvious solution, but what would be suitable?  Also, how can I protect the BME280 and the TSL2591, given that these need the active sensors to be exposed?   I realise the question is pretty basic, but my previous experience of electronics was in aviation, where the cost is several orders of magniture greater, but buys better spec'ed items.

Any advice gratefully received.

Tim

       

[Oddsocks]

Tim,

When you rebuild, to protect the boards against dampness cover the humidity sensor, pressure sensor, temperature sensor, lux sensor and thermopile sensor with “Frog Tape Yellow” painters masking tape for delicate surfaces, cut to fit over just the sensitive areas of those components on the boards, leave a small tab pointing upwards from each piece of masking tape, then after connecting the wiring but before mounting in the enclosure spray two or three thin coats of “Plasti-Dip” aerosol liquid rubber over both sides of the boards, the wiring and connectors.

Provided the Plasti-Dip coats are thin then the coating won’t migrate along the pins and sockets of the connectors but will form a seal around the joints.

Applied thinly and once dry, Plasti-Dip in plain white or black colours has a very high electrical resistance and very low capacitive value and should not have any significant impact on the performance of the boards. Heat dissipation will not be a problem provided the coating is not too thickly applied.

For the white coloured Plasti-Dip the final coating thickness will just be sufficient when the white writing on the PCB’s can still be just about discerned through the coating, a single spray coat should have a mat appearance as it is applied, if it begins to look glossy and wet then the spray is being applied too thickly. 

After the Plasti-Dip is dry peel off the masking tape from the sensors by grabbing the tab with tweezers, the masking tape will easily tear away from the Plasti-Dip coating leaving the boards sealed and just the active sensors exposed.

Then mount the boards and wiring in the enclosure.

An alternative to Plasti-Dip is a regular conformal coating such as Ambersil Transparent Acrylic Resin in aerosol spray.

Acrylic conformal coatings are good for PCB protection, with higher electrical resistance, lower capacitance and better adhesion than Plasti-Dip but dry hard and rigid and so can not be used over plugs and sockets where they will act like glue and permanently bond them together, or used on wiring where it is brittle and liable to flake off as the wires are flexed during assembly.

For this type of project and the working environment Plasti-Dip would be the better option.

The next area to look at is reducing the volume of the enclosure.

The Dostmann enclosure has too great a volume and surface area for this type of project and provides an ample supply of fresh damp air to condense on the inside of the cover.

To reduce the formation of condensation on the cover apply a ~5mm layer of closed-cell neoprene rubber sheet to the inside of the cover, either self-stick or plain sheet glued in place, with minimal sized holes cut in the sheet for the thermopile and the lux sensor openings. Cork or polystyrene sheet could be used as an alternative to neoprene, so long as it bonds well to the domed cover.

The neoprene sheet will insulate the cover and greatly reduce condensation forming on the inside of the lid.

Reduce the air flow around the inside of the cover section of the Dostmann enclosure as much as possible so that the cover section with the thermopile and lux sensors are minimally ventilated, consider placing a bulkhead of thin sheet material between the cover section and the body section with just a half dozen 6mm diameter holes around the periphery of the sheet material for ventilation.

Glue a fine mesh of nylon or similar over the bulkhead vent holes if you want to keep bugs from setting up home around the warmth of the boards.

The bulkhead material could be rigid or flexible, as long as it is taught and can’t flex with air movements and come into contact with the  boards 

The heat from the PCB’s alone will provide some anti-condensation protection for the boards mounted in the cover section provided that the airflow around the cover is minimised and the inside of the skyward facing cover is insulated.

Optionally, If the enclosure is subject to extreme conditions consider adding a “rope” resistor chain around the inside of the cover lid to add some background heating, around 5W will be ample if the cover lid is insulated but you would need to regulate or switch it off during the day to avoid overheating.

The humidity/temperature/pressure sensor board can be placed down in the body of the Dostmann enclosure, away from the cover section, and in good airflow where condensation should not easily form.

Hope that the above gives you a few ideas.

Will.

[Len1257]

The TLS2591 doesn't need to be exposed to air anyway so use a lens facing the sky, siliconed to an enclosure that is sealed closed. Similar for the MLX90614 but of course mounted without a lens. The BME280 will fail if any moisture reaches the sensor well before corrosion becomes an issue. Folk have used goretex or other breathable but waterproof material behind breather vents to exclude moisture. 

The recommendation to use an open weather enclosure is flawed for the UK. I 3D printed my enclosures but plenty of weather sealed boxes are available from the likes of RS, Rapid and Farnell.

[TimBrandt]

Hi Will, thank you for a very comprehensive and useful answer.  I'll do as you suggest.  I can print up a small inter cover for the light/temp sensors with vent holes.  I used a resistor chain to heat an allsky enclosure but had not thought of one for the SQM. 

All the best

Tim 

[TimBrandt]

Thanks too, len.  I think if I mount the BME280 with some internal shielding I can keep effective airflow but divert any splashes from rain.  I'll try goretex if I'm still having problems.  I used a Wago Capsule Box for the allsky - I may look at the same again but with some through vents for the BME280.

Thanks again,

Tim

[kbrown]

Suppose you could put a sachet of silica gel in the enclosure too and replace / recharge it every now and then...

[TimBrandt]

Thanks all for your advice.  Rebuilt using Wago capsule box and lots of Plastidip.  So far so good.  As always, second time round is much easier.

Tim