>If I run the test in a once, there are some duplicate IDs (2 or 3) in 2000 records
>
>However, if I run the test in 10 separate sections of 200 records per time,
>there is no duplicate ID
>
>(zapping the table and starting from zero)
>
>I think it's more a matter of memory problem rather than a fail in the lock () , unlock process
>
>What do you think ?
>
>Both testes type performed using 3 different forms adding data to the table at the same time.

If you are getting a single duplication you can't trust it.
Cetin