Type 3 Hard Anodizing Salt Spray Test?

The ASTM B117 spec is for neutral salt spray corrosion testing.  There are also 2 other accelerated salt spray tests (B287 Acetic Acid and B368 Copper-Accelerated Acetic Acid aka CASS) and 1 slower test (the ASTM number eludes me for now, but it uses a Water Fog).  Section X2.2 of the B117 spec points out that the B117 spec is not really designed for decorative chrome (nickel-chrome or copper-nickel-chrome) nor for treated aluminum (chromated, anodized, etc), and recommends that the B287 or B368 tests be used instead.  However, B117 is the most widely available test, so it is still used in those applications.  I, myself, use the B117 test for my nickel-chrome testing.

Note that the B117 spec does NOT actually include any reference numbers on the amount of hours spent in the chamber.  It is up to the customer to specify the number of hours.  You may be calling for 336 hours, while another shop may want more or less.  You can also choose to use the accelerated tests, but there is not concrete formula that will allow you to compare your B117 results to the B368 or B287.  If you are sending out parts for 336 hours of B117, and they are coming back withing 48 hours, then your testing vendor is not doing their job.  They may be substituting the accelerated tests.  Check your spec to see if you are allowed to use either of the other 2 tests.  I suspect that they are using the CASS test (B368).  Also, be sure that you are specifying a panel rating to your testing company.  Getting results back that say "pass" or "fail" mean nothing without the ability to say what is passing and what is failing.  See ASTM B537 on the panel rating process.  If your spec says no visible corrosion after X hours, then that would be a panel rating of 10.

I'm not sure that we can answer your question on how frequently to test your parts.  You could use a simple AQL program for each lot, or come up with something of your own.  If you are trying to meet an automotive, MIL, or AS spec, it may say how frequently to test your parts.  Personally, I have mutiple lots of parts produced every day, and I sample on a daily basis.  If you are only trying to check your parts for your own knowledge, I'd start by testing frequently, then backing off if the results continually come back positive.