[evlatests] correlator fault?

[Ken Sowinski]

On Thu, 2 Aug 2007, Claire Chandler wrote:

> I have seen a bi-stable phase state, with phases flipping 180 degrees on
> alternate integrations, on baseline 5-23.  It occurs in multiple datasets
> recently, at all frequency bands.  I think it must be a correlator
> problem.

There is definitely a problem with 5-23 IF A right now.
I have not examined many baselines, but 1-23 and 1-5 both
look normal.  It is likely a sampler problem, but I can't
explain why other baselines to antenna 5 are not affected.

Looking at 'rawish' data from the correlator it looks like
5Acos*5Acos is sometimes not very stable.  Let's replace
the sampler and see what happens.