can power be transmitted through both ends of a transmission line..

How else would power be delivered from one location to another, if not through "both ends" of the transmission line?

Obviously what you said does not describe what you mean though. :-) Can you repeat the question in different words, and maybe give an example of what you mean?

I think the OP is asking about a multiple supply power grid.

It can be done if the supply phases are in sync.

Arlowe formulated on Sunday :

Yes, and I just realized I shouldn't have posted that....

My apologies...

Power can be and often is transmitted in either direction, from A to B or from B to A, if that is what you were asking.

Don Young

It can it at least two senses:

1) In a power "grid" power can be added to drawn from any point on the grid. 2) It's possible to send "power" along a transmission line from point A to B and at the same time sent "power" from B to A. This happens routinely on the telephone wires. ** Posted from **

Another example: cable Internet carries both upstream and downstream data traffic on a single coaxial cable. There are different frequency bands used for upstream and downstream, which makes it fairly easy to have bidirectional amplifiers along the cable.

Dave

Telephone is more "interesting" though! The data is sent both directions using the same bandwidth, and echo canceling technology is used to separate the two (when necessary).

As I understand it, the local loop manages to send signal in both directions by using nothing more than a special transformer at each end of the loop. It couples two signals going in different directions to a single balanced pair, with relatively low unwanted reflections at the interface. For a local loop, where the round-trip delay isn't audible, that's all that is needed for echo suppression. For long distance, the echo is delayed enough to be audible, and that's where additional echo cancelling or suppression is used.

Dave

On Tue, 19 Aug 2008 04:22:36 +0000 (UTC) Dave Martindale wrote: | snipped-for-privacy@apaflo.com (Floyd L. Davidson) writes: |> snipped-for-privacy@cs.ubc.ca (Dave Martindale) wrote: | |>>Another example: cable Internet carries both upstream and downstream |>>data traffic on a single coaxial cable. There are different frequency |>>bands used for upstream and downstream, which makes it fairly easy to |>>have bidirectional amplifiers along the cable. | |>Telephone is more "interesting" though! The data is |>sent both directions using the same bandwidth, and echo |>canceling technology is used to separate the two (when |>necessary). | | As I understand it, the local loop manages to send signal in both | directions by using nothing more than a special transformer at each end | of the loop. It couples two signals going in different directions to a | single balanced pair, with relatively low unwanted reflections at the | interface. For a local loop, where the round-trip delay isn't audible, | that's all that is needed for echo suppression. For long distance, the | echo is delayed enough to be audible, and that's where additional echo | cancelling or suppression is used.

I remember long ago building a balanced interface that let me inject audio into the telephone pair without getting any of that audio back to me. Now if I could only remember the exact circuit. It involved 2 transformers. It worked quite well. I ended up with a home make speaker phone system that was truly full duplex and had clear good audio all the way across the room. And I mean full duplex as in both parties could talk at the same time and hear each other at the same time. I haven't seen a commercial speaker phone system that can do that.

There is also a design I do remember that involved 3 or 4 correctly chosen resistors and depended on the line impedance, while attenuating the signal some. The input and output were at crossing diagonals in a square formed with the resistors.

Lots of digital technology, though, just does "ping pong" by only sending one way at a time, and alternating which way as needed. USB apparently does this.

More or less true, but misses the part that is "interesting", which is the ability to _totally_ separate the two signals. A "standard loop" using only a hybrid transformer doesn't separate the signals on the transmission line. It only prevents the transmitted signal from being fed directly to the receiver at the same time it is fed to the transmission line. It does nothing to prevent the transmitted signal from going down the transmission line, being reflected, and coming back down the line as in "incoming" signal that is sent to the receiver.

In essence, if a transmitter and receiver were paralleled across a transmission line, the transmit power would be equally split between the receiver and the transmission line. The hybrid prevents that. It still splits the power in two, and only half of it goes down the transmission line but the other half does not go to the receiver (it is dissipated in a resistor). And the echo return that comes back down the transmission line is fed directly to the receiver.

Adding echo cancellation separates the outgoing signal that is echoed back an an *incoming* signal on the transmission line, and removes that from the incoming data sent to the receiver.

The significance is not small either. For our ears, it makes almost no difference unless the delay between the outgoing signal and the incoming echo is relatively great, hence EC has traditionally been added to long distance calls that are greater than 1800 miles in length.

But for something like a modem, which uses phase modulation, it is necessary to remove even short delay echo. The trick there is to build EC directly into the modem and disable any EC that might be added by the Public Switched Telephone Network (PSTN).

I'm not sure "audible" is the correct term there. It certainly is audible, but it is also indistinquishable to the human ear from the desired "sidetone" signal supplied so that the speaker can hear themselves. With a small phase difference, it just adds to the sidetone signal. With a large enough phase difference it becomes exceedingly distracting.

Echo Suppression, basically an analog technology, is no longer used.

Good explaination, but in the telephone network there is a hybrid at both ends of the loop and the loop is terminated at both ends in a "compromise impedance", so a reduction in reflected signals on the loop does occur. Typically the echo return loss from a hybrid with comprise balance (the kind in the loop plant and telephones...) will result in a 10-20 dB echo return loss. The unbalance at the phone results in the sidetone signal, which people need to keep from shouting into the phone or thinking they aren't connected. For very long distance, the "magic" number for good service is about 40 dB of echo return loss, so there are echo cancelers installed in the network, typically on circuits that exceed 500 miles (via net loss echo control is no longer used, toll connecting trunks have 3 dB of loss and that is the only loss in the digital portion of the network). Unfortunately, the signal to quantizing noise ratio of A or mu-law encoding is only about 35 dB, so network echo cancelers include a "non-linear processor", which is effectively an echo suppressor (i.e. it inserts a large loss in the opposite diriection of detected speech transmission). As analog long haul facilities are no longer used, the classic echo supressor is no longer used. It is interesting to note that the first network echo cancelers (developed in 1979, in service

1/1/1980) were placed in analog trunks on a card that physically was the same as an echo suppressor card. The card contained two codecs in addition to the digital echo canceler chip.

Interestingly enough, if the the "echo canceler" is placed at the CO line card before A or mu-law encoding, significantly higher echo return losses can be achieved. Many modern codecs use a 16 bit linear encoding followed by a conversion to A or mu-law, so if the echo canceler is placed in the codec prior to the conversion, echo return losses approaching 60 dB can be obtained. Agere used to make such a device, but I don't know who "owns" it now or even if it still is in production. BTW, it was recommended that the non-linear processor be disabled when that chip was used...

In any event, the adaption must be turned off if a modem call is in progress otherwise it screws up the training of the modem echo canceler. Modem answer tone is 2100 Hz and is used to disable all network echo control devices (except the fixed hybrids, of course).

Eric Tappert

If one were to make an analogy of a power line and telephone lines, I think a long transmission line with a bunch of party-line drops hanging off of it would be more like an electrical utility rather than thinking of it as a loop-start POTS line or a 2 wire hybrid line or a 4 wire E&M line.

So long as the handset can couple an AC signal to the line, any other handsets can hear or talk If an end user has a crank ringdown generator he can ring someone elses bell. If the listener is to far away no signal makes it to that end but that does not stop others down the line from hearing people close to them. This is more like how the grid works IMHO.

peace dawg