I still don't understand if these are examples of what you are getting from a regex that "doesn't work" or examples of what you want to extract from the given strings. In any event, here's an approach that produces the given output. I don't say it's the most efficient or elegant.

c:\@Work\Perl\monks>perl -wMstrict -e
"my @test = qw(
   Sample1Repeat1A Sample2Repeat2 Sample3Repeat 4SampleRepeat
   4SampleRepeat4  5SampleRepeat5D
   );
 ;;
 for my $s (@test) {
   printf qq{'$s' -> };
   if (my ($p_r, $d1, $p_mc, $p_new_mc) = $s =~ m{
     \b (.+?) (?: (\d*) (?: ([[:upper:]]?) ([[:upper:]]?) )? )? \b
     }xms) {
     printf qq{'$_' } for $p_r, $d1, $p_mc, $p_new_mc;
     print qq{\n};
     }
   else {
     print qq{no match \n};
     }
   }
"
'Sample1Repeat1A' -> 'Sample1Repeat' '1' 'A' ''
'Sample2Repeat2' -> 'Sample2Repeat' '2' '' ''
'Sample3Repeat' -> 'Sample3Repeat' '' '' ''
'4SampleRepeat' -> '4SampleRepeat' '' '' ''
'4SampleRepeat4' -> '4SampleRepeat' '4' '' ''
'5SampleRepeat5D' -> '5SampleRepeat' '5' 'D' ''
[download]

Update: sjain: I just saw your note above about not having access to Data::Dump, so I've changed the example code to a version that uses neither that module nor Data::Dumper, the output of which doesn't look so good in this instance.

Further Update: I tried this regex with strings  RS RC1XY RW12QW1X and while the latter two parse ok,  RS does not, so AnonyMonk's approach below looks like a better one.