PEMS at Location 4 (PEMS4), continued

All the while I've been looking at better ways to do things. One of the biggest is distribution of the sensor network. That also included power distribution for sensors so not to rely heavily on the USB bus from the Raspberry Pi. I built out the "Pi Station" from its rushed testing setup into something planned out, with separate 12VDC and 5VDC power supplies, RJ45 distribution, two separate sensor networks, along with test meters, TVSS protection (of the data bus), and fuses.

Small Pi Station
Figure 1. Small Pi Station
Click for larger picture
Expanded Pi Station
Figure 2. Expanded Pi Station
Click for larger picture

Something that I wanted to do was provide external power for the aspirated temperature sensor housing. For that I laid conduit out over the ground next to the fence, but for out in the yard I dug a trench and laid the conduit through there. Interestingly enough, I never had any wiring chewed on by the squirrels that populated the area.

Conduit in a trench
Figure 3. Conduit in a trench
Click for larger picture
Conduit all covered up with wires coming out
Figure 4. Conduit covered, wired pulled through
Click for larger picture

I later rethought sensor power requirements, and because there were two master controllers I decided to isolate the power supplies for each bus too. The 12VDC power supply in the shed provided 12VDC for the fan in the thermometer housing, but also to step down converters for 9 and 5VDC. These were placed in an external junction box, and the near year I used them they hadn't developed any issues that I'd seen.

Board with DC step down converters
Figure 5. DC Power Board

Junction Box wiring
Figure 6. Junction Box 1, all wired up

Problems with water

Nothing ever works as planned. I would deploy sensors and one day things would ground out, usually after a storm. Sometimes though it would all work like a champ and I'd go on happy, just to find out the WiFi network connection to the shed went down. Having the sensor networks isolated in the shed meant I could normally tell which one it was pretty quickly, if one was up but not the other then it was a bus issue. A combination of cheap RJ45 connectors, power sent on pins not used, and bad seals in some places caused problems with corrosion. The pictures below illustrate some of the ones I took photos of.

Corroded RJ45 connector and junction
Figure 5. Bad connector caused ground out of sensor bus. Had to cut out and replace both connector and junction.

Bad Capacitor
Figure 6. Capacitor failed due to water entry into "waterproof" module

Good Capacitor
Figures 7. Replaced the cap and figured out the failure point, water didn't get in again

Very bad corroded connector
Figure 8. This one got pretty bad too, but it was just wrapped in tape near the ground

NEXT: Still More!

Back to the Index Page

© 2020
created 12 Apr 2020