MOPS would like a refstack for the STS fields.

The previous STS.refstack.20110613,20110615 wasn't usable as it was destroyed by magic.

STS.20120510 was made using a similar number of inputs from the best exposures from 2011. Not pretty but hopefully sufficient.

Details to add:

• FWHM of test stacks (STS.reftest.20120510.c#f#) like done with MD refstacks
• unconv vs convolved stacks and issues found
• test diffims with recent STS data

7 projections and their skycell overlap.

Approximate pattern, E left, N up:

 1 2     
3 4 5   
 6 7