Patience is a virtue.

1. I made the classic out-by-one error here:

   sysread( FCS, my $data, $eData - $sData ) or die $!;
   [download]

   The start- and end-of data offsets are inclusive. Therefore to read the whole section you need $eData - $sData +1.

2. The delimiter for the TEXT section pairs in the second example is '|'.

   This is a regex meta-character, which means the split to construct the hash:

   my %text = split $delim, $text;
   [download]

   Screws up. The solution is to quotemeta the delimiter read from the first character of the TEXT section just in case:

   my %text = split quotemeta( $delim ), $text;
   [download]

3. In the second example, $P1R is set to 4096, for $DATATYPE:I & $P1B:16.

   This means that only the first 12 bits of the 16-bit values should be considered. And that once the data has been unpacked, it must be masked to 12 bits to ensure that you don't get out-of-range values from picking up "stray bits" in the upper 4-bits.

   It would be a bloody-minded data supplier that would pack 12-bit data in to 16-bit words without ensuring the up 4-bits are zeroed, but according to the specification he isn't obliged to do so. And their example demonstrates this.

   I've added thise code to deal with that:

   if( $text{ '$DATATYPE' } eq 'I' and defined $text{ '$P1R' } ) {
       my $mask = $text{ '$P1R' } - 1;
       $_ &= $mask for @data;
   }
   [download]

4. I also noticed that there is a value in the TEXT section $PAR that indicates how many values there are per event, so I've used this to display that number of values per line:

   my $par = $text{ '$PAR' };
   
   for( my $i = 0; $i < @data; $i += $par ) {
       print join ' ', @data[ $i .. $i + $par - 1 ];
   }
   [download]

5. You asked by .msg about the %tmpls hash.

   my %tmpls = (
       '1,2,3,4' => {
           I => { 16 => 'v*', 32 => 'V*' },
           F => { 32 => 'f<*' },
           D => { 64 => 'd<*' },
       },
       '4,3,2,1' => {
           I => { 16 => 'n*', 32 => 'N*' },
           F => { 32 => 'f>*' },
           D => { 64 => 'd>*' },
       },
   );
   [download]

   Three TEXT section key/value pairs identify the format of the binary data in the DATA section.

   1. $BYTEORD Identifies the byte ordering of the data as big or little endian using the text tags '1,2,3,4' or '4,3,2,1'.

      Other byteordering are apparently possible!

   2. $DATATYPE: Identifies the type of the data: {I}nteger; {F}loat ; {D}ouble.

      There is also an 'A' datatype I haven't attempted to cater for. Though it doesn't look too hard to do.

   3. $P1B: Identifies the binary size in bits of the data. 16/32/64 bits.

      This seems to be mostly for the 'I' datatype. 64-bit integers is a possibility. The datastructure makes it easy to cater for that.

   All of this information is in the spec! Other formats that I haven't catered for in my example are possible.

   I use the %tmpls structure to map those three values to an appropriate unpack template:

   my $tmpl = $tmpls{ $text{ '$BYTEORD' } }
                    { $text{ '$DATATYPE' } }
                    { $text{ '$P1B' } };
   [download]

6. There is a lot more useful information in the TEXT section hash.

   You should work through the specs, and work out what each of those key/value pairs is, and how you should be using them.

Anyway, here's a modified version that incorporates the corrections outlined above, that handles both supplied samples.

But it is still only example code intended to give you a starting point from which to progress!

#! perl -slw
use strict;
use Data::Dump qw[ pp ];

my %tmpls = (
    '1,2,3,4' => {
        I => { 16 => 'v*', 32 => 'V*' },
        F => { 32 => 'f<*' },
        D => { 64 => 'd<*' },
    },
    '4,3,2,1' => {
        I => { 16 => 'n*', 32 => 'N*' },
        F => { 32 => 'f>*' },
        D => { 64 => 'd>*' },
    },
);

open FCS, '<:raw',
    'c:\downloaded\FCSExtract-TestData\FloatData4031.fcs'
#    'c:\downloaded\FCSExtract-TestData\CC4_067_BM.fcs'
    or die $!;

sysread( FCS, my $header, 58 ) or die $!;

my( $id,
    $sText, $eText,
    $sData, $eData,
    $sAnalysis, $eAnalysis
) = unpack 'A6 x4 (A8)6', $header;

sysseek FCS, $sText, 0 or die $!;
sysread( FCS, my $text, $eText - $sText ) or die $!;

my $delim = substr $text, 0, 1, '';
my %text = split quotemeta( $delim ), $text;

pp \%text; <>;

sysseek FCS, $sData, 0 or die $!;
sysread( FCS, my $data, $eData - $sData +1 ) or die $!;

my $tmpl = $tmpls{ $text{ '$BYTEORD' } }{ $text{ '$DATATYPE' } }{ $tex
+t{ '$P1B' } };
die 'Can\'t handle format' unless $tmpl;

my @data = unpack $tmpl, $data;

if( $text{ '$DATATYPE' } eq 'I' and defined $text{ '$P1R' } ) {
    my $mask = $text{ '$P1R' } - 1;
    $_ &= $mask for @data;
}

my $par = $text{ '$PAR' };

for( my $i = 0; $i < @data; $i += $par ) {
    print join ' ', @data[ $i .. $i + $par - 1 ];
}
[download]

Thank you so much for your help I appreciate it very much. It now works for CC4_067_BM.fcs but still doesnt give the correct results for FloatData4031.fcs I have no idea how to fix that, apologizes for the inconvenience.