Update 19.05.18 - done redesigning the board. Bigger but more comfortable to use!

I did another iteration on the board design after I thought about what pissed me off the most when assembling the last (first) unit.
It clearly was soldering all those connections from the cables coming in from outside!

Another thing I didn't like was the batteries weren't accessible from the top. So here's the final design:

Above: I switched to Wage spring loaded clamps. These work perfect for small to medium diameter wires! No more soldering strip headers to sensor cables! Yes, they're more expensive but so is the bigger board... Now you can use external V in or one 18650 li ion cell.

Above: back of the board. Now with two EEPROMs which have address-select solder pads so you can add more than one for additional space. You can now choose to supply your SD directly from VCC or by an TPS27082 high side switch which can be turned on/off by your microcontroller.

Update 15.05.18

I've been adding features and made a new board:

Board will be 60x40mm. I'm not yet happy with how it looks and components are placed. I guess another layouting round is due. Just to do things proper :) I mean there even is wasted space... 

All features of the list below have been added anyway.

What's missing in V1.1

Hardware

• proper, easily accessible µSD socket
• button that can fire an interrupt (important for interaction, because the logger is sleeping most of the time = no digital pin polling)
• way to get the data out without opening the watertight case
• external EEPROM of 512kBit to 2Mbit = less SD writes

Software

• writing all logs from EEPROM to SD with the press of a button (atm I have to wait till the EEPROM is full (and then written to SD) or I lose 6,25 hours of data)
• sleeker state-machine. I mean it works but is quite confusing at times...
• even less global variables
• some more error-checking

What's coming in V1.2

Hardware

• µSD socket by Attend (already here)
• breaking out one interrupt pin for a button/switch (one is taken by the RTC already, the other one is used for SD select in V1.1 = waste of interrupt pin)
• adding an HM-11 BLE bluetooth module to transfer logfiles to my phone while near the hive (Datasheet). Planned to work like this:
  • HM-11 always switched off
  • pressing button for 2s: BLE module activated, waiting for send command
  • send transfer command by phone user (via bluetooth serial app)
  • transfering logs line by line (already tested)
  • shutting down module again
    • transmitting power is settable by AT command
    • module uses less than 10mA -> pin powered!
    • no way to change/set data by expernal user

Conclusion

The next board will be slightly bigger and totally re-designed to fit my new needs. I want to have it ready and tested before the bad weather period comes (november...) so I've got plenty time for tests and stuff. Plan is to log not only one hive through winter...

Another idea would be to include no SD card at all and just use an EEPROM. But this would need the user to get the data off the logger every few weeks (512kbit: 16 days with 40kb logs, every 15 minutes or about a month with 1Mbit EEPROMs)...