Kitz Forum
Broadband Related => FTTC and FTTP Issues => Topic started by: DaveC on July 07, 2018, 11:29:18 PM
-
Hi,
I have two AAISP VDSL lines, one via TT, one via BT. Equipment is identical - two HG612s linked to a Firebrick FB2700.
I've had the TT line for a few years, and the BT line was previously ADSL, converted to VDSL a couple of weeks ago.
The TT line shows max attainable around 80500/27000, and line rate 80000/19999. speedtest-cli and iperf3 show upstream throughput reaching about 18.5Mbits/s.
The BT line is syncing slightly slower at around 76500/25000, and line rate 77100/19999. However, speedtest-cli and iperf3 are only managing about 16.5Mbits/s upstream.
RTT to A&A is about 5.5ms for the TT line, and about 6.5ms for the BT line.
What could cause the slower upload on the BT line?
Here is the (very full) output of the various xdslcmd info options - https://vps.dchapman.com/2018-07-07__23-20-06.txt
Thanks.
-
Have you swapped the LAN ports on the Firebrick?
Have you swapped the 2 modems over between the 2 lines?
I would try that and see if the BTw upload still under performs then you have completely ruled out your hardware.
-
Here is the (very full) output of the various xdslcmd info options - https://vps.dchapman.com/2018-07-07__23-20-06.txt
Having taken a copy of the above file I have produced a montage of the usual four snapshot plots (Bit Loading, SNR, QLN & Hlog v sub-carrier index), attached below. I wonder if your BT circuit just has more cross-talk than the TT circuit? The SNR and QLN plots certainly hint at some cross-talk.
-
@j0hn,
I tried swapping the hardware between lines, yes, and that made no difference.
@burakkucat,
Thank you very much for doing that. I would run dslstats, but as I only have a Mac and headless Linux boxes, it's not convenient. I've started logging the stats myself into an SQL database, but haven't yet written any UI to display/graph them.
I have to admit to not understanding the graphs (more reading needed), but I guess that if the cause is crosstalk, then that means there's nothing I can do about it?
-
I have to admit to not understanding the graphs (more reading needed), but I guess that if the cause is crosstalk, then that means there's nothing I can do about it?
I get a hint that there appears to be some cross-talk involvement but I would like to read other members' views based on their interpretations of the four plots.
-
I don't think crosstalk would explain the difference in throughput. Both lines have the same sync speed, usually crosstalk would affect the sync speed.
-
Hi
I could be wrong sorry, but the hg612 which is connected to the bt line needs the QoS turning off on the hg612
This should then allow the upload speed to reach 18.5
I mention the above as it has been discussed many years ago on kitz
Many thanks
John
-
Hi
I could be wrong sorry, but the hg612 which is connected to the bt line needs the QoS turning off on the hg612
This should then allow the upload speed to reach 18.5
I mention the above as it has been discussed many years ago on kitz
Many thanks
John
They ruled out modem configuration by swapping the modems between the 2 lines.
You need to get on to AAISP about the BT line.
If both lines sync at the same upstream rate of 20MB they should achieve very similar throughput.
If you really did swap each modem between the 2 different lines (they should hot swap, no config changes needed) then there is an issue on the BT line somewhere.
Have you also double checked the AAISP control panel to make sure no limits have been set?
-
Hi
@john many thanks, but TT may not use QoS but BT Fttc does have an impact
It still worth checking as QoS disabled does give 18.5mb on BT fttc lines. If enabled, 16.5mb and that matches the current through put
It’s just a thought and you maybe correct
Many thanks
John
-
@john many thanks, but TT may not use QoS but BT Fttc does have an impact
Talktalk would have no control of that
The upstream QoS setting on the HG612 is 100% controlled by the modem and would slow the lines upstream regardless of the backhaul provider.
If it was the HG612 at fault then the slow upstream throughput would have swapped between the 2 lines when the modems were swapped.
QoS setting on the HG612 was the 1st thing that came to mind which is why I suggested swapping the modems between the lines. This rules out the modem completely, not just config but hardware too.
-
QoS setting on the HG612 was the 1st thing that came to mind which is why I suggested swapping the modems between the lines. This rules out the modem completely, not just config but hardware too.
Turs out you were right. QoS was still enabled on the BT line's HG612 (I was previously using it for a backup ADSL line, so didn't worry too much about it), and disabling that brings the upstream in line with the TT line. The TT line's HG612 had QoS disabled already.
I think when I swapped the modems I must have only looked at the sync rates (my initial concern was that the BT line was syncing lower than the TT line), and I couldn't have actually done any speed tests on the upstream - my bad.
So thanks for the pointer.