| Forum Home | ||||
| PC World Chat | ||||
| Thread ID: 76153 | 2007-01-22 22:55:00 | Too far from Exchange? _Nah..........Too close | pctek (84) | PC World Chat |
| Post ID | Timestamp | Content | User | ||
| 518465 | 2007-01-25 01:52:00 | For analogue telephony "The objective off-hook impedance of any device, measured at the line terminals, is complex and is represented by a network consisting of a 370 ohm resistor in series with a parallel combination of a 620 ohm resistor and a 310 nF capacitor". It does matter for sidetone and echo. | PaulD (232) | ||
| 518466 | 2007-01-25 02:48:00 | I thought you'd previously posted that the problem at Wyndham was a Telecom Network one so why are Xnet chasing an individual line issue? I did. Because they have to play Telecoms game. All the posts about impedance, attenuation, ohms etc. You can stop now - my husband used to install and maintain exchanges for a living...he knows already. In fact this is getting boring now, I posted it so you could all have a luagh not take it seriously. Geez. |
pctek (84) | ||
| 518467 | 2007-01-25 03:39:00 | For analogue telephony "The objective off-hook impedance of any device, measured at the line terminals, is complex and is represented by a network consisting of a 370 ohm resistor in series with a parallel combination of a 620 ohm resistor and a 310 nF capacitor". It does matter for sidetone and echo.But that's for telephone devices, not the lines. The magnitude of that impedance is nearer 1000 ohms than 600 or 120. The line doesn't look anything like that either, and it's the line I was talking about. The lines are not matched. At voice frequencies, or ADSL frequencies. Perhaps they should be. ;) They would be if the network had been designed for transmitting data using RF carriers. But the netwrok was designed as a telephone network. :D | Graham L (2) | ||
| 518468 | 2007-01-25 04:27:00 | pctek, if you're posting for our amusement I think we all got the joke several posts ago. Graham L, whatever but that network represents the impedances you see when you sweep an average length telephone line at voice frequencies which is why BT and most other telephone companies moved to that or similar rather than the traditional 600. |
PaulD (232) | ||
| 518469 | 2007-01-25 04:39:00 | Oh dear. They moved to twisted pair cables because the overhead open wire lines cost to maintain and were too expensive to maintain. It just happens that the particular wire gauge, and the construction of the twisted pair cables gives a characteristic impedance of about 100-120 ohms. Particular wire gauge, and a particular spacing of overhead lines could give a characteristic impedance of around 600 ohms. A "average length line" is all very well. However, what is seen at the far end of the line will vary horrendously as the actual length of a line differs. If the load (the device) does not match the line characteristic impedance, the line acts as an impedance transformer. I'm not surprrised they see complex impedances. Telephones are not RF devices (except accidentally). ADSL modems are. My original point, which Terry will have seen, is that the attenuator circuit he gave will have the stated attenuation only if it was fed from and terminated in 600 ohms. It was a little bit of pedantry. ;) The aim of trying that would be so see if some reduction in the siganl might make an ADSL modem work better. The absolute value of the attenuation would not matter. |
Graham L (2) | ||
| 518470 | 2007-01-25 08:53:00 | Right, I was going to respond earlier, but pctek didn't seem to want to know, but seeing as we have now hijacked the thread lets continue . . . . . . . Graham is right, I just wanted a resistive balanced attenuator with a "ball park" attenuation . The calculations are somewhat meaningless because the line impedance is unknown . There is the 'standard' Telecom cable to the pole outside the property, most likely 0 . 4mm, ancient heavy twisted flex from the pole to the house , then 100 ohm cat 5 to the modem with a home-brewed gaseous arrestor and transorb unit HV protection unit interposed . So 600 ohm is as good as anything to start with as a basis . This is the attenuator with DC blocking: . imagef1 . net . nz/files/attenuator2 . jpg" target="_blank">www . imagef1 . net . nz The attenuation produced of 45dB total, ie line plus attenuator is fairly representative of what happens a few kilometres distant from the DSLAM . The modem worked perfectly well at this level, and the router connection speed was almost unchanged . 300 kHz is a common frequency at which to quote attenuation . Here is a commonly seen chart of ADSL spectrum: . imagef1 . net . nz/files/spectrum . jpg" target="_blank">www . imagef1 . net . nz As Graham implied the line attenuation does vary with frequency, 13 . 8dB/km is the quoted value at 300kHz for 0 . 4mm PIUT, but it rises to 26dB at 1 MHz as is shown in this table: . imagef1 . net . nz/files/cableattenuation . jpg" target="_blank">www . imagef1 . net . nz (PIUT=Paper Insulated Unit Twin) This table is from here: . aca . gov . au/acmainterwr/telcomm/industry_codes/codes/c559 . pdf" target="_blank">internet . aca . gov . au An interesting related document which also has some maths is: . acma . gov . au/acmainterwr/telcomm/industry_codes/codes/c559_2_2005 . pdf" target="_blank">www . acma . gov . au I don't 'believe' for one moment that being too close to a DSLAM would give miserable ADSL speeds, but belief belongs to religion, not science, so using an attenuator like the one above is one way of testing the veracity or otherwise of what the Telecom reps . are saying . It is so easy to do, especially for an ex- Telecom installer :) |
Terry Porritt (14) | ||
| 518471 | 2007-01-25 22:08:00 | I don't 'believe' for one moment that being too close to a DSLAM would give miserable ADSL speeds, but belief belongs to religion, not science, so using an attenuator like the one above is one way of testing the veracity or otherwise of what the Telecom reps. are saying.I was just lurking on this thread and hadn't intended to comment, but the reference to religion got me going. In my view the alleged comment from Telecom verges on Shamanism. If it were true, there would be a "bubble" zone of unreliable service around every dsl-enabled exchange in the country (or the world for that matter) and no service for anybody closer than 3-400 metres. I can't see the argument as being technically sustainable in any way shape or form. That is not to say that other issues involving line or equipment faults etc could not cause problems in high signal strength areas, especially if harmonic artefacts were being generated that could confuse cheap modems. Arses need kicking. Cheers Billy 8-{) |
Billy T (70) | ||
| 518472 | 2007-01-25 22:50:00 | On another forum pctek has posted the modem synch speeds which are the max possible for a Full/128k plan. Unless Xnet know that their other customers in her area are OK, any problems with slow speed are on the other side of the exchange (DSLAM). | PaulD (232) | ||
| 518473 | 2007-01-26 02:31:00 | All forms of computer networking rely on magic (which is just advanced religion -- magic miracles work) . The incantations do matter . So does the state of the moon, whether it's Tuesday, and the orientation of the rubber chicken . I've seen (I even had one once) 9600 bps modems which had adjustable attenuators to be set for best performance (on leased lines) . There was a test point to connect an oscilloscope for "eye display" . I have an idea that 57k modems do this level adjustment automatically as part of their initialisation . So I wouldn't totally reject the idea that it would be possible for "too good" a signal being a problem . However, I wouldn't expect it to be a common problem . Of course, Telecom hardware might be programmed to reflect vehement criticism of Telecom and Xtra on the uplink as noise on the downlink . :D |
Graham L (2) | ||
| 1 2 3 | |||||