Perhaps you can try this
CODE
(?=^[a-zA-Z0-9]{6,8}$)[a-zA-Z][a-zA-Z0-9]{0,6}\d[a-zA-Z0-9]{0,6}
Let's look at this part
CODE
^[a-zA-Z0-9]{6,8}$
This looks for an alphanumeric sequence of characters length 6 to 8. Carat ^ to match the start, dollar sign to match the end.
Then it's placed in (?= ... ) for a lookahead match. It'll take more to explain it. Just go with it...
Then you have [a-zA-Z] to match your condition of "must start with alphabet". The next part might be a little complicated to explain. Let's go with the 2 extreme examples first:
"a1bcdefg" and "abcdefg1", representing alpha-numeral-rest_are_alphanumeric and alpha-rest_are_alphanumeric-numeral.
\d matches 1 numeral (so you'll have at least one number somewhere)
You need a starting alphabet, and at least one numeral. That leaves at most 6 more characters, hence the {0,6} part.
So, the first part, the lookahead fulfills your length requirement (between 6 and 8 characters). The second part fulfills your "start with alpha, and at least 1 numeral" requirement. Combine them and you have done it!
For more information, check this
regex lookahead article.
FYI, \w matches alphabets, numerals and the underscore character, so it might not be what you want.
Hope this helps!