I'm in the same boat here being that it was in fact, a NIB, 1st dive reg set. Yeah, I know the assembly guys are skilled and very fast and proficient at what they do, but these parts should still have been 100% inspected at final assembly. Sadly everything typically hinges on that final functional test - it either passes or fails. A pass infers every previous operation was a 100% complete success. A failure unfortunately gives the false indication that your test is working and weeding out ALL the bad parts. You might not catch the latent defects - those will show themselves down the road but you should be rejecting all the gross errors such as misassembled parts. Not always the case.
Unfortunately, since the reg 'fixed itself' there probably won't be any hard evidence of what caused the lock up. Whatever caused it, cleared itself, and is now downstream - where it will likely get lost when it gets disassembled and cleaned. I really hope it gets found and we can get the failure mode here. Other than a gross assembly error, I really didn't think you could get a stage to fail in this manner without some significant effort