Well the nearest exchange with a Zen presents that I have found is Stirling which is about 40 miles away or Waverley Exchange in Edinburgh at 50 miles. Unless they now have POP in the Perth Exchange, which is 8 miles by road, or Dundee, 24 miles away as the crow flies, that the details have not been updated.

Will give it till the end of the month and see if it improves. If not and they can't put it back the way it was, I may give Talk Talk Business a try as it says they have presence in my local exchange.

Did anyone solve this. My connection was fine on ADSL2+ with TTB now moved onto VDSL2 and my single thread downloads are half of what multi thread downloads are?

See speedtest in signature.

The issue discussed in this thread is specific to Zen's backhaul network, so only affects Zen and those companies using wholesale backhaul from Zen (IDnet, I believe).

Some other providers have had issues, and suggests there may be some common bit of kit/firmware upgrade/config in core networks thats doing this, but given the majority of speed testers mask the issue people when told to test connection, get good figures

That is what I was thinking Andrew there must be something common to this. I know different ISPs but similar Openreach equipment.

My downstream single thread is half of sync although Uno are detecting more than acceptable errors on the line.

In reply to a post by MrSaffron:

Some other providers have had issues, and suggests there may be some common bit of kit/firmware upgrade/config in core networks thats doing this, but given the majority of speed testers mask the issue people when told to test connection, get good figures

Very true. Zen have understandably been very coy about what the issue was, as well as the topology of their network and the link providers. Of course, this sort of thing is commercially sensitive, so it is understandable that they do not want to comment. It believe that most if not all of the backhaul links in their network are provided by Vodafone (Cable and Wireless), so it's possible that the issue is common to other providers using Vodafone circuits or identical kit / equipment to Zen.

As you say, the way that most speed testers only give multi-threaded throughput figures masks the issue. Thank you for all you and the rest of the thinkbroadband team do in providing your speed test; it is a trusted and invaluable aid to debugging.

I am with IDNet and for quite a while I have had serious single thread speed issues. As things stand IDNet are planning to move me to "an alternate part of the back haul network from the exchange into our network", though quite what that means in practical terms I have no idea.

I know IDNet do sometimes make use Zen backhaul but that isn't available in my exchange and I have so far been unsuccessful in getting an answer to the question of who's backhaul I am on. In fact the question just gets ignored! TalkTalk backhaul is available but I don't know whether IDNet use them.

Anyway, the change is booked for 11th April so it will be interesting to see what happens.

In reply to a post by Adrian:

Anyway, the change is booked for 11th April so it will be interesting to see what happens.

Please do post back to let us know. I too have had this issue with Zen for past 12 months and gave up trying to get them to fix and have all but accepted nothing single threaded gets full benefit of the connection. For various reasons I'm going to have a second line installed soon and I'm going to try Talk Talk Business on that and directly compare against Zen.

I am on TTB with single thread half of sync. Let us know how you get on with a second line.

Andrews & Arnold had some issues on some TalkTalk lines where the evening single thread speed plummeted, but being A&A they persisted with it in conjunction with TT and finally found the problem.

This is how a first class ISP handles customer problems and if mine with IDNet isn't resolved to my satisfaction I may consider moving back to A&A despite the higher cost. However, so far I am happy with how IDNET has dealt with the problem, but the proof is in the pudding.

AAISP Fault log