As far as I can see with what passes for logic in my brain, Wipe Cluster Tips only applies to secure file deletion.
Many thanks, your logical answer has rescued my sanity.
For decades I was satisfied that I learnt all I needed to know when I used 8086 code in DOS 3.3? to format the tracks of a floppy disc.
I knew that special NON-rewritable identity blocks were each followed by a gap that allowed time for a Western Digital FD1771 to decide if the upcoming sector was what it wanted,
and upon getting what it wanted it then had time to switch into the appropriate mode to either read that data sector,
or instead to write a new data sector in which case it had to start erasing a few bytes of leading space because it could not avoid the possibility that the timing was a bit "off".
I understood that the same was true for my 20 MB HDD but I never risked trying to format that.
I chose to believe that NTFS was merely a "rip-off" from Unix of a Journalling System "overlay" that organised the use of data sectors but NOT how they were transferred to/from the media.
I accept that everything gets faster (and sometimes better) as technology advances,
but an improvement on the FD1771 technology that can preserve the first 500 bytes whilst over-writing the last 12 bytes of a 512 byte sector is an affront to my intellect.
It requires belief in a technology advance from several kilobytes per 200 mSec rotation of a floppy disc to much larger data tracks at much higher speeds on a modern HDD,
together with starting to write to the precise end of the 500th byte in a 512 byte sector,
instead of simply erasing a few bytes and then starting to write a new sector within a few dozen bytes of the previous location.
Incidentally, I seem to recall that groups of independent bits are NOT written as individual items on a track,
but those groups of bits become "encrypted" into patterns on the track,
and it would be more than ambitious to over-write the the last half of a pattern and hope that the first half of the group of bits can be preserved.
I guess the only way to deal with the cluster tip of a live file is to read the entire file and write back a new version that pads out the end of the file with zeroes fill the final cluster,
and that would "scare the willies out of me".
I wonder if, and how, Heidi can actually wipe the cluster tip of a live file.
I searched and found others were WANTING this capability in two results from
http://www.heidi.ie