I decided to have go at doing a better job of adding a common mode choke to the Mark 1 perhaps to make a Mark 4
. From reading references earlier in this thread it is disappointing that the Mark 3 choke is not bifilar as a bifilar choke should be able to do a much better job on the common mode with much less differential mode impact.
I have taken a ferrite toroid with Al value 12 micro Henry and wound my adsl lead round it. I have tried 10 turns, 7 turms and 5 turns and settled on 5 turns for a ~70Mb/s connection. For a lower sync speed more turns should be better. 5 turns gives me a sync speed loss of about 1Mb/s (similar to a Mark 2) more turns give a bigger reduction. I attach Hlog with and without the toroid and a phote of the toroid itself. In Hlog you can see the differential mode attenuation starting to bite. That should have the effect of stopping any higher frequencies saturating the modem input stage. The common mode attenuation can be calculated from the Al value and number of turns and should be enough to give good common mode rejection over the vdsl band.
Unlike the Mark2 and mark 3 this arrangement does have a helpful impact on my errors.
I have been using it for 3 days and the whilst the ES are down a bit the CRC and SES are down much more. I will need a week more to give a confident outcome but I like being able to control the filter with the number of turns