Kitz Forum

Broadband Related => ADSL Issues => Topic started by: Weaver on July 28, 2018, 03:19:16 AM

Title: Slight packet loss, coinciding with lightning strikes on the mainland
Post by: Weaver on July 28, 2018, 03:19:16 AM
From 21:30 to 23:40 GMT last night, around the time when there were lightning strikes on the mainland plus a couple at the very northernmost tip of the Western Isles, I noticed there has been some sporadic 1% - 2% packet loss, at times when there was just a bit of sporadic web browsing activity, no serious sustained load.

I am running at a really really low downstream SNRM these days, only around 1.5 dB and I am wondering if this is enough normally but the distant lightning just pushes it over the edge of what it can cope with? Almost all of the strikes were a long long way away, and only one during the whole period showed up as an event on my hardware lightning detector. The information about their occurrence came from a live service on the internet, driven by other peoples’ hardware detectors hooked up to some servers.

Have other people seen the same thing with distant lightning?
Title: Re: Slight packet loss, coinciding with lightning strikes on the mainland
Post by: banger on July 28, 2018, 03:37:26 AM
I was in my car on the drive changing the battery and setting up my radio again when I put the MW band on it was off station and I could hear my VDSL signal on the radio but also distant lightning strikes over the VDSL tones. There was no local strikes as it was bright sunshine but currently DLM has set my SNRm at 3.3db. I would say 1.5db is a bit low and it is quite possible in my considered opinion for some packet loss to occur even with distant strikes.
Title: Re: Slight packet loss, coinciding with lightning strikes on the mainland
Post by: Weaver on July 28, 2018, 03:50:09 AM
Banger, I am thinking along the same lines as you.
Title: Re: Slight packet loss, coinciding with lightning strikes on the mainland
Post by: banger on July 28, 2018, 03:59:07 AM
Yes I think we are in agreement.
Title: Re: Slight packet loss, coinciding with lightning strikes on the mainland
Post by: aesmith on July 28, 2018, 12:48:40 PM
Have other people seen the same thing with distant lightning?
For some reason DSLstats stopped saving snapshots again, so I don't have detailed statistics.  However the A&A graph shows packet loss (a term I've heard abused so much it sets my teeth on edge by the way) leading up to the disconnections and eventual failure ..

Title: Re: Slight packet loss, coinciding with lightning strikes on the mainland
Post by: Weaver on July 28, 2018, 06:19:40 PM
See the period marked 22:30 Fri - 00:40 BST

https://flic.kr/p/26Pj7ay
Title: Re: Slight packet loss, coinciding with lightning strikes on the mainland
Post by: burakkucat on July 28, 2018, 06:29:19 PM
I wouldn't worry about it. Leave things alone and let them do their own thing. You don't have the DLM process, waiting to pounce, that G.993.2 service users' of which need to be so cautious.
Title: Re: Slight packet loss, coinciding with lightning strikes on the mainland
Post by: Chrysalis on July 29, 2018, 11:53:45 PM
weaver I guess you got very clean external noise in your area and good quality copper to be able to run long line ADSL at a 1.5db margin. :)

Interesting about the hardware lightning detector.
Title: Re: Slight packet loss, coinciding with lightning strikes on the mainland
Post by: johnson on July 30, 2018, 12:12:39 AM
Interesting about the hardware lightning detector.

Indeed! I spent some of time on blitzortung.org after someone suggested it earlier in the year when we had those few weeks with many large electrical storms.

Do you have one of their system blue/red/green kits? What sort of antenna do you have, one of those squarish orthogonal ones? Do you upload your data to servers that make those great maps with all the traces to detectors?

Interesting stuff anyway, really cool that there is a home grown network of these with accurate enough data to correlate and map strikes.
Title: Re: Slight packet loss, coinciding with lightning strikes on the mainland
Post by: Weaver on July 30, 2018, 04:28:00 AM
No, I don't own their hardware, but if I were well enough I would love to do so. I have a commercial SkyScan unit, which has extremely badly designed software - dangerously so - because if you do not configure it correctly, or if you ever get a power cut, the defaults are dangerous in that it just doesn't warn you. This caused £900 of damage when the device was asleep on the job. But now I have it by my bed so I can see it and check that it is running, not asleep and I also have the Blitzortunglive app running on two iPads, and that warns me if their detectors pick up an event inside a circle that I can choose on a map. They do have sensors in Scotland but none near me. Even so, the range of these things is very very impressive as I have just proved to myself, so I think that I can probably trust their data feed. My own unit picked up spikes that were an enormous distance away, >100 mi because it was set to max range, which was probably a bad idea. Blitzortunglive then gave me the position of the strike straight away, otherwise I would not have had any idea what it was or even if it was some kind of red herring. Electrical disturbances in the house can trigger my lightning unit, for example, light switches. Turning the sensitivity down to something reasonable like 50 mi might be an idea. The thing is, I am thinking about the nightmare case where a storm is approaching and the only strike that happens is one at extreme range, then nothing until it is right upon us, so in that case I need to be woken up to look at the map, never mind the fact that the strike is too far away to be a problem in itself.

It would be a nice software feature if the Blitzortunglive app could track the storms, differentiate their positions and then see if the velocity vectors brought them into your chosen circle, giving a time to intercept plus radius of approach. I would write it myself. I would need to work out how to get hold of the live data.