Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[burakkucat]

I would be interested to know if number (3) resolves the problem . . . before progressing to number (4).

[Weaver]

More news: total strangeness, and an answer to (3).

5) When modem #2 was power-cycled, it triggered an immediate hollow-curve phenomenon in modem #4 !!  I can see the hollow curve now. There was a burst of #4 FECs and CRCs at the exact power-cycle time followed by a linear rise over the next 90 mins of the observation period. The downstream #4 SNRM plunged vertically from 3.0 dB to 0.1 dB. BT forced a line #4 re-sync after nearly 5 hr at 0.1 dB downstream SNRM and correspondingly high error rates.

Modem #1 was fine; why one modem was cross-triggered and not the other is unknown.

6) Later this evening, Janet pulled the DSL line out of the afflicted line #4. This cured the hollow curve phenomenon (HCP). So that gives us an answer to another of our questions.

So seeing that pulling out the mains can cause HCP as well as curing it, is it a dangerous idea? A cure that we had discussed might aggravate the disease?

I believe that HCP was completely unknown before the great lightning strike. ( - is that correct ? ) A search through the forums might be very helpful but difficult.

We urgently need to swap out these modems to get some answers. I’ll see if I can arrange a modem swap-out tomorrow. I have a mountain of modems in my store. For some reason Janet calls my equipment store-chest or chests "Bert’s Barrow" - because we don’t know anyone called Bert and it doesn’t look like a barrow, so I’m stumped.

[burakkucat]

Thinking scientifically, I have to ask the question: Do the experiments produce repeatable results? 

[Weaver]

I agree. I don’t think that some of the results are repeatable, yet to me it doesn’t matter in one sense, because a positive result proves that something is possible, and that is very interesting to me. For example, the difference between modem #1 and modem #4 behaviour is, I would expect, just due to random stuff.

Am I right in thinking that the HCP is a recent thing, completely unknown before 2020 ? Have I ever reported it, and have any other kitizens seen it ever? To qualify, someone has to show their bits-vs-tones or SNR-vs-tones (aka ‘vs bins’), and they only have to have a small hollow. Do not confuse the required shape with that caused by a pilot tone, or very narrow band interference from eg radio stations. The HCP only needs to be of modest width, say 15 tones, in the very early stages, but will be developing rapidly to cover tones 40-80 or 40-85 in the case of ADSL downstream. I have absolutely no idea if such a thing as VDSL2 HCP exists or is even possible.

[burakkucat]

. . . to me it doesn’t matter in one sense, because a positive result proves that something is possible, and that is very interesting to me.

In that case, I must say "be aware of chasing your own tail" . . . for I have been in that situation.

[Weaver]

Heard and duly warned. It depends on what predicates you’re interested in, no?

[burakkucat]

It depends on what predicates you’re interested in, no?

Absolutely, yes.

[Weaver]

Have told AA that we now have a fix / workaround. I assume that they’re very relieved. I’m hopeful that swapping modems out altogether will make this whole issue go away, but that is just my optimism plus a desire for a plausible rational answer.

Luckily I have a huge stash of VMG 1312-B10A modems. AA once sold me one, whose config I used as the basis for all the others. I intend to label all the current four modems as “working:suspect / lightning: HCP: 2020 - 2021-12” and configure fresh ones out of my stash.

[Weaver]

I am now told by my dearest that the source of the name is that there was originally a jute shopping bag which had "Bert’s Barrow" printed on it. Then it became full and so we had to get a large storage trunk made of willow which sits on the landing upstairs. This is the main equipment store. But original name stuck firmly. I never knew it’s meaning until today.

[Weaver]

After, what, about four weeks, there has been another outbreak of hollow curve disease, on line #2 this time. Noticed this afternoon by Janet, who got an automated email from A&A saying line 2 had gone down a couple of times. Immediately cured by power cycling the modem. This was with very short off-time. The DSL line was plugged into the modem the whole time during this, not disconnected.

[burakkucat]

Noted.

[Weaver]

And I’m ashamed to say that I still haven’t replaced the current modems with new ones out of my store, my assistant having been so very poorly, the project has been postponed until better times.

[burakkucat]

Would it, perhaps, be a more interesting experiment to replace just one of the modems? Keep notes over a, say, three month period of which circuits that show the "hollow-curve phenomenon".

If the circuit with the replaced modem still shows the phenomenon then it cannot be due to equipment malfunction.

[Weaver]

Good point. Will do so. And NB: Recall that this never happened before 2020-01, indeed was completely unknown to science.

[Weaver]

How long has line 2 been good for? I note that line #2 or is it line #4 that is the most recent line to have suffered from the disease? When was it ? - 2021-11-11 approx?

Note that unlike before, switching off line #2 has not triggered the disease in a different line.