Bus Shelter Install Notes
=========================
DHD20160202

Install of first LoRa sensor module (and temporary GSM module).

SUCCESSFUL TEST OF FULL INSTALLATION METHOD AND LORA RADIO TRANSMISSION.

Installation crew: Damon HD & Deniz E.

Sensors:

 1) 8aad GSM (temporary)
 2) e0dd GSM (temporary, aborted)
 3) b9f9 LoRa

Locations are recorded elsewhere, by ID and install date.

All eaves panels installs, none plain eaves panels.

Observations during install:

  * Two-step ladder provides sufficient height for working.


  * (3) Line of sight from DigiCat building
        and thus probably LoRa base station.

  * (3) TX poor from eaves void on first install attempt,
        visually seems to be almost complete Faraday cage!

  * (3) Sensor unit test from upstairs at KX station behind target stop
        yielded zero successful transmissions to the LoRa base station.

  * (3) Eaves panel above powered video ad panel so bolt access difficult
        and temperature readings expected to be elevated,
        but not enough noise heard to prevent voice pickup.

  * (3) Tested used of cleaner and hook+eye fastener to top of void,
        successfully.
        (Policy: fluffy side to shelter, hooks to unit/battery.)

  * (3) TX on-site test before install 11:05, after first install 11:46,
        after re-install 13:47.

  * (3) TX ~50% or better success,
        temperature readings close to Met Office observations nearby.


  * (1) One of eaves voids had photosensor and non-standard screws.
        Still plenty of space in void with photo sensor,
        and this void was used.

  * (1) Shelter is under large metal bridge.

  * (1) External battery was given own hook-and-eye fastener.

  * (1) TX offsite test before install 07:37, on-site pre-install 14:35,
        after install 15:02.

  * (1) Good TX success, apparently close to 100% though probably blocking.

  * (1) Will probably adjust to 'discard oldest' TX queue model
        to keep stats as fresh as possible
        and timestamp (applied at receiver) as accurate as possible.


  * (2) Timetable void not correct type/size for install of our box 3.

  * (2) Missing bolt from one (vanilla) eaves panel,
        photosensor and bent internal flaps on other
        but still enough space for us to use.

  * (2) TX off-site text 19:00 previous evening, TX on-site preinstall 15:29,
        only on TX heard after install.
        On retrieving the unit it was no longer seeing its GSM card.

        (Subsequent investigation shows one signal connection severed,
        possibly by mechanical action of USB connector during install.)

  * (2) Hook-and-eye fastening detached from top of void, ie failed,
        when we went to examine/reset unit after extended TX silence.