I'd like to edit my post however it is not possible. Therefore I send longer post as reply.
I've noticed on my ADSL2+ circuit unsusual Hlog.
There are many undulations. I've written a lot about it on forum.kitz.co.uk. They suggestes perhaps I have got circuit consisting of many sections with different gauges. However as I know this type of circuit is very widespread but I don't see so unregular Hlog on other circuits.
I've done much progress towards explaining this Hlog however on forum.kitz.co.uk moderators forbidden to continue writing about my research results. Therefore I can't defend my ideas against their counter arguments.
I understand many people are not interested with my tenacious searching of answer but I think it is not reason to forbid me writing. Uninterested persons don't have to read my posts. Censorship is harm to potential interested person.
1. What was I explained?
I've discovered what causes through at tone ~55. It is caused by measurement errors. On the other circuit I've noticed small dip on the same tone.
Furthermore, some day I had resynchronization and I and by mistake I was connected with target SNR margin 15 dB for donwstream and 17 db for upstream and other power output profile (18.9/12.4 dBm instead usual 18.6/12.1 dbM). By mistake because as I reconnected again I got again usual 6 dB SNR margin with usual power output. Then I had slightly more regular Hlog (weaker undulations).
This is a evidence of Hlog measurement errors (for example on tone ~55). It is noticable when I place two graphs on one coordinate system.
One forum.kitz.co.uk user (burakkucat) has not agreed with my explanation. He emphasized limited accuracy of Hlog caused by granuality.
However I think it weights in favour of my theory. Precisions limits cause that dip at ~110 tone is not noticable on other circuit Hlog. This Hlog have significantly smaller but the same measurement errors. Even my circuit had smaller than usual measurement errors at one day.
This is visible when I attach two graphs on one coordinate system.
2. What haven't I explained yet?
I've still not disovered what causes tone ~210 dip. It is not smaller on anomalous synchronization with more accurate Hlog.
However one day I had event which would be helpful. I've noticed sudden reconnection and change of attenuation values at the time of pavement resurfacing works. It was 23.5/20.2 dB instead usual 25.0/14.9 dB (usual on D-Link DSL-2740B). Probably more users were reconnected because I had SNR margin about 4 dB immediately after reconnection (instead 6 dB). Next day after that day I also had reconnection during pavement resurfacing work and attenuation came back to usual values. On day with unusual attenuation day Hlog was significantly different per some tones. I think some shock caused some ipairtment on some joint reflected signal differently than usually.
Following there are usual and onedayhlog on one coordinate system.
I suspect dip moved from tone ~210 to tone ~150.
Other interesting observations which could help explain cause of dip is that as I've shortened cables in my home installation I've noticed that as I expected attenuation per the most of tones was reduced however on some tones attenuation increased.
1. Red graph is Hlog with 12m long cable connected to master socket. Green graph is Hlog with 1m long cable connected to master socket. Attenuation with 1m long cable was slightly lower per tones 33 to around 180. It was also lower per tones 200 to 400. However, it was slightly higher per tones 180 to 240 and on the end of graph.
2. Red graph is 1m long cable connected to master socket. Green graph is Hlog when I've cut down 20-30m cable from box to master socket and moved master socket. Attenuation on shorter circuit is slightly lower per tones 33 to 180 and per tones 350 to 511. However it is higher per tones 190 to 240 and per tones 280 to 320.
3. Red graph is Hlog with 12m long cable connected to master socket before cutting down. Green graph is circuit after cutting down 20-30 m cable to master socket. Attenuation on very short circuit is lower with exception per tones 180 to 240. On these tones is higher attenuation.
It is interesting becuase it seems to me that longer cable does not cause signal is less attenuated on any tone than it was attenuated when cable is shorter. Perhaps my impairtment causing Hlog undulations is related with this unusual observation.
Edited by konrado5 (Fri 03-Aug-18 18:23:56)