Mystery solved. Based on two successful ECI/r connections and two more ECI/r failures the ECI/r seems to be sensitive to which dsl line is A or B. The twisted pair cable that the Hg612 and HG622 were very happy with has the pairs swapped round one end to the other. Using a "no swap cable" the ECI/r syncs as it ought to - just fine. My collection of cables are mostly straight thro but a few are swapped?? I assume the swap is the issue - the only alternative (unlikely I think) is that swapped cable has a plug that the ECI/r does not like. i have had things on and off too many times today to want to risk furthers tests.
Is it a wrong to have a dsl cable which swaps round the pairs??
Current stats without capping don't show much difference. All with 80/20 sync.
HG612 038 BLOB snrm 7.4 attainable 83800
HG622 033 Blob snrm 6.7 attainable 85400
ECI/r n/a attainable 83900
I notice that the srnm's don't seem that well related to the attainables between the two HG's.