ADSL works over wet string (2017)
revk.uk
revk.uk
"Only four properties really affect the performance of most digital transmission structures. The "big four" transmission-line properties are impedance, delay, high-frequency loss, and crosstalk." Dr Johnson then goes to describe these properties in barbed wire.
0: https://www.sigcon.com/Pubs/edn/SoGoodBarbedWire.htm
P.S. Yes, this is the Dr. Howard Johnson of the famed "High-Speed Digital Design" book.
[1] https://gizmodo.com/barbed-wire-fences-were-an-early-diy-tel...
https://en.m.wikipedia.org/wiki/Party_line_(telephony)
A party line! Our cabin still has a grandfathered one with the distinctive ring. I think we pay 1/4 of the regular rate so we keep it despite everyone having a cell phone now.
I don’t think there’s anyone else on the line.
Got a funny spam call with a recording for some miracle septic tank treatment. It asked me to press 1 for more information and I can’t do that on a rotary phone. Scammers don’t understand their target market. (We also don’t have septic system, just an outhouse.)
Here we go - 1990: https://www.washingtonpost.com/archive/national/1990/07/13/p...
[0] https://en.m.wikipedia.org/wiki/Interactive_voice_response
If you have absolute pitch, and have a friend with absolute pitch (and I mean absolute frequency really not pitch), and are both freaks who can sing perfect sine waves... you could sing the dual tone for number 1 (697 Hz + 1209 Hz).
Since at this point we are talking about mutant alien phreaks, you should also be able to do overtone singing and squeeze out a 2 or a 3 column on your own which are 10% flat and sharp from the 2nd order harmonic of the first row 697 Hz.
Happy phreaking
Whistle, don't sing. Try it with a realtime analyzer which shows the fundamental being far more prominent than any harmonics above it, which means it's close to a sine wave. Then sing various vowels which will show as a series of harmonics whose amplitude ratios change per vowel but are never reduced nearly as much as when whistling.
Was the bandwidth that narrow at the resonant frequency? Intuitively, I wouldn't have expected that "wet string" has a high Q. :)
If I had been your student, I would have had great fun with it.
And the bandwidth of a resonant circuit depends on its Q. Hence my question about a wet string’s Q, which might well be different from a guitar string’s Q.
https://www.allaboutcircuits.com/uploads/articles/series-res...
I chose to just use the single (good) line.
So, yes, I can confirm this does work in the field… but with about as much practically as you’d expect.
Is this just a North America thing or an everywhere thing?
Some time later I saw a Pacific Bell truck and crew trenching the street. (We had underground utilities.) I excitedly asked them, "Are you running fiber?" "No, just more copper."
I guess they saw the demand for lots of copper lines per house and decided to meet it!
The underground phone lines did lead to some excitement one time. Their water sealing wasn't very good, and after a heavy rain the line would get noisy and clicky. During one of these episodes, we got a loud knock on the door: "San Jose Police! Open up!"
I ran to the door and asked what was going on. One of the officers said, "We got a 911 call from this address. The dispatcher couldn't hear anyone on the line, and no one answered when they called back. So we're required to come out and investigate the situation."
I replied, "That's odd, neither of us called 911. Oh... I think I know what may have happened. Come in and I'll show you."
We went to the kitchen and I put the wall phone (remember those?) on speaker and had the officers listen. It was clicking up a storm!
Then I asked them if they remembered the old rotary phones (they did). I explained how the phone dial worked by breaking and making the circuit N times for each digit you were dialing. And it must have been that in that flurry of clicks that the circuit was broken nine times, then once, then once again.
During the Australian NBN^^ rollout. Where I used to live was on the border of a FTTC and a coax/DOCSIS one. My street had no pre-existing cable TV cable so you'd think no brainer to lay fibre. Nope, brand new coax. And because the street had no coax that meant no homes had an existing cable so they had to trench every driveway to connect it to the street. It would have been an absolute no brainer to run fibre then use the existing phone lines up driveways to their homes. Cheaper, and an easy upgrade path to full fibre. This was 2019.
^^Australia built a national broadband network. It was originally designed to be all fibre (except for remote places over satellite etc), but politics happened and it was changed to a multi technology mix which included DOCSIS, fibre to the node then vdsl for the last km or whatever, or fibre to the curb and vdsl up your driveway. And the government entity building it NBN was buying up old copper phone lines and coax.
German article on the matter:
https://netzpolitik.org/2018/danke-helmut-kohl-kabelfernsehe...
Standard neoliberal bullshit coupled with a lack of imagination as to how things like local-loop unbundling might be used to create a competitive market. The UK is now in a fairly good position re broadband, but it could be so much better.
But even in North America, I don't think every home had two pair all the way to the central office. They might not even all have two pair all the way to the crossbox [1]. It was not uncommon to order a 2nd phone line, and not be able to get it, because there weren't any available pairs on your pole. Especially if your neighborhood was established well before the explosion of the internet and fax in the mid to late 90s.
It made a lot of sense when I ran a dial-up BBS, or for homes that were heavy modem users. Otherwise, once high-speed internet was an option, 2nd lines made less and less sense.
10 years ago there was a huge amount of surplus Cisco 1700s with SHDSL line cards (apparently coming from Czech government's project to connect every school to internet in early 00's) and we had huge spool of flat phone cable that was left over from earlier project. So we had the bright idea to wrap ethernet trafic in AAL5, pass that over SHDSL and use that as an LAN for anime convention. Interesting observation from that it matters whether the cable is coiled or un-coiled. We built and tested the whole network in a lab (with coiled cables), it worked well, the G.991bis bonded links synced up at 6Mbps with two pairs and everything was good. We labeled everything and we built the exact same thing (including same cables) at the venue and the links will not go above 1.5Mbps and were frequently losing sync. Disabling the second bonded pair caused it to work reliably at 2Mbps. (Back then, we did not need that much of bandwidth, it was essentially for few SCCP phones, some IP tunelled serial ports and ssh)
Well, next year we bought two boxes of Cat5 cable and switched to native ethernet (and today the backbone spans are 10G fiber, as it also carries video streams).
I'm sure DOCSIS is great for what it is, and in fact it's extremely impressive what it can do with existing cable lines, but the second biggest player here (Videotron) basically milked it dry. Again it's weird to praise them but Bell invested in wiring up the entire city and suburbs and I can get 3gbps symmetrical GPON FTTH for the price that 150mbps used to go for not even 3 years ago. I blame DOCSIS in a way because it made players with existing lines extremely complacent
https://www.revk.uk/2017/12/please-upgrade-me-to-adsl-over-w...
Due to them, most ISPs in big cities in Brazil are switching to fiber to the home.
Fun fact: Switzerland operates the only significant G.fast deployment in the world. The technology was such a flop that Huawei stopped making the hardware.
Nope. In Germany, M-net (Munich) and NetCologne operate massive g.fast networks, and if you don't want Huawei, go for Adtran's 516 lineup on the ISP side and AVM on the CPE side.
I didn't know about network equipment from other manufacturers, not sure why we're stuck with Huawei here, but it's a dead end. The plan is to migrate as many G.fast subscribers as possible to FTTH in the next 5 years, and eventually to mmWave 5G in a few corner cases.
By the way most of the fibre needed to connect the FTTC/VDSL/GFast cabinets was already in the ground and they just needed to splice into it, as it was there for exchange to exchange backhaul, and leased lines. It's the final ~500 metres that's the expensive part of installing FTTP, as that has to be new and there is probably not a lead in duct that is clear so means climbing poles, clearing duct, or new trenching.
Had they considered an infrastructure upgrade to carrier pigeons, cans on a string, or smoke signals?
But the latency and loss rate would make it all worthless.
https://www.reddit.com/r/vintagecomputing/comments/9of84w/ne...
Rings true for me. In my younger days I was "privileged" to catch an early morning flight to San Jose with a stack of tapes that I then drove to Palo Alto. All because a RAID array had gone kaput and tapes in a carry-on provided better bandwidth than the upstream from the office where our backups were.
Before this we didn't even get DSL because the lines were so degraded.
This reminds me of that.
Why not? Is it dangerous, ou would it just make it stop working?
He used to be hostmaster@cam.ac.uk, amongst other things. Why he’s farming internet points on HN I’ve no idea!
- Comments bots (like on other threads), usually to promote a crypto-thing.
- Submissions bots to farm karma (likely one of them here).
- Upvotes bots (taboo).
Recent karma is necessary for upvoting other posts, as there is a system which makes you can't upvote if you didn't get upvoted yourself.
If you can upvote some posts on HackerNews (as it would be on Reddit) you can help people make a lot of money, but again, you need recent karma.
I’ve got a lot of good stuff in my link log https://dotat.at/:/ so I thought it might be fun to share a few retro classics. There has been some good discussion, so it seems to be worth it.