GPS Data Logger Software - BT747

There are different ways to influence the file size:

1. Fields selected (recorded, selected for output);
2. Filter Settings (=selected records);
3. File type.

1. In principle only the fields that are recorded in the logger can be written to a file - so the less you record, the less is written.  With BT747 you can record what you want (so that you can use it for filtering) and still select fields to write.  That can be done in 'Output Settings' in the 'File Output Fields and Items" subpanel.  According to your explication, you only need UTC Time, Latitude, Longitude and Height.   You can set the number of digits for Lat/Lon to 6 and for 'Height' to 1.  You do not need any of the other items, but you can add them as you like.
   In the 'Various' subpanel on 'Output Settings', you can also untick 'RecordNumberInfoInLog' - the positions will not be given a sequential number in that case.
2. In 'Filters' you can indicate what is a Track Point and what is a Way Point, and filter out some 'bad' positions.  Track Points and Way Points are technically pretty much the same.   The difference is mostly 'conceptual': a Track Point is a position to indicate the track, a Way Point indicates a special position that you would like to 'highlight'.
   In order to reduce the file size a maximum, do not 'define' Way Points and deselect everything in 'Log Reason' for Way Points.  Also untick 'No fix' - you only want actual positions.
   On the main panel ('Log operations'), you can set the start and end date.  That is another way to reduce data.
   You can also generate several files in stead of one - creating one file for each day or track - you can set this in the 'Various' subpanel on the 'Output Settings' panel.  
3. The 'native' log format for Google Earth is 'KML' or 'KMZ' and not 'GPX'.  When you use 'GPX', Google Earth has to do more work to import the data.  Therefore, use KML or KMZ.
   The difference between KMZ and KML is that a KMZ is a ZIP Archive of the KML that can also hold other data such as pictures, icons, ... .  Obviously, a KMZ is smaller than a KML.   Even if the expanded data is the same, I've observed that loading of a KMZ is faster.  That is essentially because disk access is slow and probably takes more time than the extra effort for decoding.

So, to summarize: select only the fields that you really want to write, limit the number of way points that you create, select 'kmz' as the output format.

Google Earth needs minimum: position (latitude & latitude) and time (not sure time is absolutely needed, but pretty usefull).

Track Point vs. Way Point.

Suppose I have a crossing in a road.  The center of the crossing is a position on the road.  When I consider the road, that center position will be a Track Point because I consider the road as a Track.  However, the same position is also a Way Point because I can use it to locate the crossing.  In that case, the Way Point would be a crossing type and could take the form of an icon representing a crossing on the map.

If every position on my track would be a way point then I would get a lot of icons to represent the track and not a line.  That is not very practicle.

So Way Points and Track Points are pretty much the same, the difference is in how I use them.  Because of that, they are encoded differently in GPX or KML for instance.  For a track point what interests me most is the positions, for a way point I want to know more - at least what the way point represents (a restaurant, a traffic light, a location where to turn, a target, ...).

You can deselect every tick box in 'Log reason' as indicated in a previous message to avoid way points.

Any position can be both a way point and a track point at the same time.  The default (on initial startup) is to have only buttons and 'user' positions as a track point.