New Fine Offset fun

[Charlie]

My station invented some strange temperatures and rain yesterday. It then flatlined most parameters, then fixed itself a couple hours later. Then this morning, the temperature started ramping down. I changed all batteries everywhere, but upon re-powering everything, the outside temperature gave random low readings... starting with negative numbers, then ramping up to about 7 or 8 degrees. Now it is showing blank (_ _ _) but other outside readings are about what is expected.

Thoughts? This is a North American 915 MHz unit. Time to dig for bad solder joints, or throw it in the scrap pile? It's only 8 months old...

[Charlie]

Well, for those that follow such things, I found the leads on the thermister corroded. I removed it, cleaned them up and resoldered with good quality solder - presto I'm working again! The inside of my unit was very clean, with only the odd solder spash. It looked like there may have been a conformal coating applied.
The fun is over for another day - back to work for me.

[steve]

Great result, Charlie, and useful info for anyone else whose station starts behaving oddly. Well, more oddly than usual, anyway.

[wd40]

Do you think it was a cold solder joint to begin with?

[Charlie]

No, this was not a cold joint. The joint looked fine (on top anyway). I have seen oxidation creep up leads under solder before, and this was the case here. I'm surprised these stations last as long as they do... the construction is certainly not exterior grade. The "chip on board" construction method is far from reliable as well. It's usually only used where the life expectancy is months, not years.

[Gina]

Another good reason to put it in a decent Stevenson Screen - keeps the rain off as well as the sunshine. The original screen doesn't afford much protection of either sort!

[wd40]

Gina I was hesitant when I ordered the bigger solar shield but know I am glad I did. I had not thought of the rain problem thinking the instrument was more water proof than resistant.

[Gina]

True! The circuit boards used in the outdoor sensors have no more damp protection than the console - only the microchips are encapsulated. Other equipment I have that is designed for use outdoors (in a weather resistant housing) have a waterproof coating over the entire boards. The RJ11 connectors are not of outdoor spec either. It's all very much "down to a price" but can be given a longer life with suitable care and extra weather protection.

[Charlie]

After resoldering, and allowing things to operate for an hour or so, I decided to look over the board very carefully with a magnifier before putting the case back together. When I did, I found a hairline crack in the material covering the processor chip. I know the purpose of the material is only to protect the chip from exposure to the elements, so a hairline crack is a very bad thing. I can’t even imagine how it could have happened based on my understanding of the manufacturing process.

Anyway, I decided to cover it with something to re-seal it. My old standby “five minute epoxy” tubes were empty so I used a thread locker. It is supposed to be dry in 3 hours max. At about 4 hours, I replaced the batteries, and everything worked. Then random readings began. I saw everything from temperatures in the high 50’s to -30’s. Wind was dead calm and I’m not sure if any direction was being reported. Pressure and humidity seemed fine.

This randomness went on for many hours, and I went away late afternoon assuming I had ruined the transmitter unit.
I woke up this morning to find good data had been recorded for at least the last 12 hours. I’ve remounted everything and I’m keeping fingers crossed that the fix is permanent.

I mention this experience because others may have cracks in the protective material and this could explain occasional wild temperature readings, particularly on humid days. I’m debating coating the whole board with 5 minute epoxy once I replenish my supply

[sanramonrover]

As to weatherproofing the PCB and components - just took apart a back-up beeper from a vehicle. The kind mounted to the licence plate holder. Gets logs of road grit, water etc. The PCB was completely dirty and grimy but still worked perfectly after about 20 years. And I wondered why.

Turned out the whole thing had been dipped in a wax - looks like a mixture of bees wax and paraffin.


Maybe a do it yourself remedy idea for the weather station PCBs?

[Ned]

The temp and humidity sensors would need masking before any blanket coating was applied