Timestamp offset

The timestamp from the Solys2 is logged (which I believed is based on a GPS received, so it is highly accurate).

```python
df['Solys_time_1'] = pd.to_datetime(
df['Solys_gps_time'].str[:17], format='%Y %j %H %M %S', errors='coerce')

# This format only exists in January 2015
df['Solys_time_2'] = pd.to_datetime(
df['Solys_gps_time'].str[-19:], format='%H:%M:%S %m/%d/%Y', errors='coerce')

(df.index - df['Solys_time_1']).dt.total_seconds().resample('1h').mean().plot()
```


Start by running the pandas timestamp parsing shown in the issue and comparing the resulting Solys_time_1 values with df.index. Determine the source and extent of the offset across the two timestamp formats; the issue is complete when the discrepancy is explained and the timestamp handling has a reproducible correction or documented expected behavior.

Written by the indexing model from the issue text.