Apply corrections to rev. 0 board

[platisd]

As an INS I want to work as intended by my creators so to make them happy.

HW rev. 0 has the following problems:

• Batteries slightly too close to the WiFi Module, should be moved couple of millimeters to the sides.
• Understand why WiFi module receives too low voltage and implement a solution.

DoD: Above problems are fixed and new boards are ordered.

[platisd]

These are my current findings in regards to this problem:

• I could not get the Wemos D1 mini to power up directly from 2 AAA batteries (or the equivalent voltage) in any consistent nor reliable manner. I could however get it to power up directly from 3 batteries connected to the 5V pin or the 3.3V pin. I could get the Wemos to power up directly from the two batteries hooked up to the 3.3V pin.
• Using two 2N7000 MOSFETs to handle the current (it won't work with just the one we currently have!) I was able to power it up reliably with voltages from 4.0V to 4.5V connected to the 3.3V pin. This is probably not a very good idea, but seems to work. Now whether this will destroy the ESP chip at some point is a different question. :stuck_out_tongue:
• Two MOSFETs will be also able to drive 5V to the 5V pin and power the WeMos up reliably.
• Three MOSFETs were also able to drive the Wemos with 4.5V connected to the 5V pin. However there seemed to be problems when I lowered the voltage to around 4V.
• I was reading contradicting reviews on the Wemos power consumption while in deep sleep (i.e. in the occasion we'd decide not to use a power controller chip). Some were very optimistic (e.g. 150uA in deep sleep which would allow the batteries to last for at least a couple of months) while others gave a pretty bad picture (e.g. 20mA). I did not have time to make my own experiments, so this might worth investigating more.

So, overall there seem to be various options available on how to power the WeMoS up. The overall problem is choosing the appropriate power source at a voltage that is as close to the specs as possible AND at the same time will keep being able to power the WeMos even if the voltage falls as time passes (and also cheap).

My best suggestion is 3 AAA batteries wired up to the 3.3V pin driven by 2 2N7000 MOSFETs. At full battery power, the voltage that actually reaches the chip is 4V. According to the specs, the maximum voltage is 3.6 voltage. We seem to be over it by 0.4V in other words, but since it seems to work and the plan is to keep the Wemos turned off for most of the time, I suggest that we go for that. It is a risk, but I honestly cannot think of a better solution that will not cost us a lot and will have other problems (I am looking at you Lithum batteries).

Let's make a decision tomorrow and order the new boards by tomorrow night!