Battery Failing In the middle of the night

The data looks to me like the battery voltage dropped in July, someone replaced it with a charged battery on 8/4 and data was continuous, but then on 8/25 someone replaced the battery with the almost-dead one from earlier in the month.  And if you’re using the sample LTE logging sketch like what’s found here: https://github.com/EnviroDIY/ModularSensors/blob/master/examples/DRWI_LTE/DRWI_LTE.ino , you’ll see in line 423 of the code that the station won’t transmit data if the battery voltage is lower than 3.55v.  So your station should still be recording data onto the memory card (down to 3.4v as set in line 419), but won’t transmit unless the battery is above 3.55v.  If you replace the battery with a fully-charged one, then you shouldn’t get gaps in your transmissions.

Also seems to me the core issue is you lost charging /solar about the end of May.

Possibly the panels got over grown, or shaded out.

I have a separate LiIon charger ~ https://www.adafruit.com/product/1905 – but a spare mayfly also works.

What’s the battery size (maHr)  and solar panel (W)?.  Also which port do you plug the solar panel into?

I’m wondering though, you are also sampling every 5minutes, do you need it to be that frequent? Would 15minutes work?

Interesting, it does seem to be a sudden V drop. Looking at the raw readings on the downloaded csv it is also oscillating from 4.1V to 3.6V a couple of times, which seems very unlikely in real life. Time snip at UST 2021-08-25 14:00
<table width=”344″>
<tbody>
<tr>
<td width=”140″>time</td>
<td width=”140″> time diff</td>
<td width=”64″>V</td>
</tr>
<tr>
<td>2021-08-25 14:00</td>
<td>0:05</td>
<td>4.094</td>
</tr>
<tr>
<td>2021-08-25 14:05</td>
<td>0:05</td>
<td>3.563</td>
</tr>
<tr>
<td>2021-08-25 14:10</td>
<td>0:05</td>
<td>4.094</td>
</tr>
<tr>
<td>2021-08-25 14:15</td>
<td>0:05</td>
<td>4.094</td>
</tr>
<tr>
<td>2021-08-25 14:20</td>
<td>0:05</td>
<td>3.563</td>
</tr>
<tr>
<td>2021-08-25 14:25</td>
<td>0:05</td>
<td>3.563</td>
</tr>
<tr>
<td>2021-08-25 14:30</td>
<td>0:05</td>
<td>4.154</td>
</tr>
<tr>
<td>2021-08-25 14:35</td>
<td>0:05</td>
<td>4.139</td>
</tr>
<tr>
<td>2021-08-25 14:40</td>
<td>0:05</td>
<td>3.563</td>
</tr>
<tr>
<td>2021-08-25 14:45</td>
<td>0:05</td>
<td>3.563</td>
</tr>
<tr>
<td>2021-08-25 14:50</td>
<td>0:05</td>
<td>3.563</td>
</tr>
</tbody>
</table>

We’ve found that loggers with the Digi cell boards on them have a tendency to act a little squirrely sometimes and give either erroneous readings or cause excessive power draw.  In you case, I’d start with putting a fully-charged 4400 mAh battery on your station and see how it performs for a few weeks.  If that doesn’t solve it, investigate the solar panel, making sure the wires are okay, and putting a voltmeter on the panel output (with no load attached) on a sunny day.  We’ve seen panels fail in the field but look physically fine, but they either don’t generate the proper voltage (6v) or enough current.  If neither of those solutions fix the problem, then try replacing the cell board, or at least removing the cell board from the logger for a week or two and see what the data on the card says.  A Mayfly logger without a cell board should last several 3-6 months without a solar panel with 5 minute sampling of a CTD.  So seeing what a couple weeks of non-cellular logger activity looks like can give you a clue as to what the issue might be.  But to start with, just put a different, fully-charged battery on and go from there.