Hi

About a week ago my connection started dropping several times a day. It came out of nowhere — nothing changed in my flat, no new devices, no rewiring. Before that the line ran stable for a long time.

Router: Speedport Smart 4 Typ A, firmware 010139.5.0.001.0

Location: Berlin

Tariff: [please add yours]

Between 29.08.2026 14:28 and 30.08.2026 09:37 (about 19 hours) the router logged 16 sync losses. Each outage lasts roughly two minutes before the line comes back.

Message (R006-1) "DSL antwortet nicht mehr (erfolgreiche DSL-Synchronisierung wurde verloren)":

29.08.  15:14:46

29.08.  15:17:01

29.08.  16:01:27

29.08.  18:27:33

29.08.  18:37:00

30.08.  00:39:06

30.08.  02:54:47

30.08.  03:39:13

30.08.  04:24:57

30.08.  05:10:53

30.08.  05:20:08

30.08.  08:53:56

30.08.  09:18:46

30.08.  09:24:05

30.08.  09:27:09

30.08.  09:30:45

What strikes me most is how much the sync rate varies on every retrain. Message (R007) "DSL ist verfügbar":

30.08.  05:12:46   Down 204415 / Up 23358 kbit/s

30.08.  05:21:58   Down 153156 / Up 23054 kbit/s

30.08.  08:55:50   Down 128378 / Up 18395 kbit/s

30.08.  09:20:45   Down 204413 / Up 23358 kbit/s

30.08.  09:29:11   Down 169713 / Up 23313 kbit/s

That's 128 to 204 Mbit/s downstream on the same line within a few hours.

Before several of the drops the router first logs unanswered LCP echo requests:

30.08.  00:38:18  (R107)    Die LCP Echo Anfrage wurde von der netzseitigen Schnittstelle nicht beantwortet.

30.08.  00:38:48  (R107)    Die LCP Echo Anfrage wurde von der netzseitigen Schnittstelle nicht beantwortet.

30.08.  00:39:06  (R106-1)  PADT Paket wurde vom HG gesendet. Grund: Vorher gesendete LCP Echo Anfragen wurden nicht beantwortet.

Line values from the hidden engineer menu:

SNR margin: 7.4 dB down / 8.4 dB up

Attainable data rate: [please add]

My questions:

1. Does 7.4 dB explain this on its own, or does the pattern look more like impulse noise or a physical line fault to you?

2. Is there any way to find out whether something changed on my port in the last week — vectoring group, line profile, DLM ? The sudden onset with no change on my side is what puzzles me.

3. Is G.INP normally active on a Super-Vectoring line like this, and is it worth asking for?

4. Would a lower, capped profile with a higher target SNR margin be the sensible thing to request? I'd much rather have a stable 175 than an unstable 250.

Happy to post more log data or line values if that helps. Thanks in advance for any pointers.