🚀 go-pugleaf

"Don Pearce" <donald@pearce.uk.com> wrote in message 
news:a750c7dhudwb.1h2dtb3ggjv48.dlg@40tude.net...
> On Fri, 2 Sep 2005 15:27:47 +0800, Alan Rutlidge wrote:
>
>> "Richard Freeman" <bogoff@nospam.spam> wrote in message
>> news:4317f1be$1_1@news.iprimus.com.au...
>>>
>>> "Alan Rutlidge" <dont_spam_me_rutlidge@iinet.net.au> wrote in message
>>> news:4316b902$0$515$5a62ac22@per-qv1-newsreader-01.iinet.net.au...
>>>>
>>>> "Don Pearce" <donald@pearce.uk.com> wrote in message
>>>> news:1jlu9dz4m6nho$.9bip6dnzgr15$.dlg@40tude.net...
>>>>> On Thu, 1 Sep 2005 17:25:01 +1000, Phil Allison wrote:
>>>>>
>>>>>> "Alan Rutlidge"
>>>>>> "Arny Krueger"
>>>>>>>
>>>>>>>>
>>>>>>>> It takes significant distances for audio electrical signals to
>>>>>>>> demonstrate things like reflections and standing waves. For most
>>>>>>>> practical purposes, they never happen because by the time the audio
>>>>>>>> goes
>>>>>>>> that far in the modern world, its probably already been digitized.
>>>>>>>>
>>>>>>>>
>>>>>>> Digitising an audio signal does not eliminate propagation delay if 
>>>>>>> the
>>>>>>> signal is transmitted over a long distance.  To demonstrate this in
>>>>>>> combination with the problem of impedance mismatch - make an
>>>>>>> international
>>>>>>> telephone call to a third world country like India.  The echo you 
>>>>>>> hear
>>>>>>> is
>>>>>>> usually 100% attributable to poor impedance mismatch at the other 
>>>>>>> end
>>>>>>> of
>>>>>>> the line.  And guess what?  It's all occurring at audio frequencies
>>>>>>> (300Hz
>>>>>>> to 3400Hz).
>>>>>>
>>>>>>
>>>>>>  ** ROTFLMAO   !!!!
>>>>>>
>>>>>> Does this Rutmaniac, fucking idiot actually teach such monumental
>>>>>> **BULLSHIT**  to IT industry trainees  ????
>>>>>>
>>>>>> Only the dullest teenagers would fall for such utter shite !!!
>>>>>>
>>>>>>
>>>>>> FYI
>>>>>>
>>>>>> Telecommunications data is digital ( not audio) and travels to places
>>>>>> like
>>>>>> India via optical fibre and geo-stationery satellites  -  not 
>>>>>> thousands
>>>>>> of
>>>>>> miles of bloody co-ax !!
>>>>>>
>>>>>> What a complete MORON  !!!!
>>>>>>
>>>>>>
>>>>>>
>>>>>> ..........  Phil
>>>>>
>>>>> So Phil. It is digital right up to the phone at the far end, is it? 
>>>>> That
>>>>> phone doesn't have a hybrid transformer, does it? That transformer
>>>>> doesn't
>>>>> have a limited sidetone spec, does it? The match of phone to line is
>>>>> perfect is it? There is no signal reflection from that phone, is 
>>>>> there?
>>>>> That reflected power isn't digitised by the codec is it? That 
>>>>> digitised
>>>>> signal isn't audible at the sending end as an echo is it? And the fact
>>>>> that
>>>>> these signals might use satellites or optical fibre somehow magically
>>>>> makes
>>>>> all these reflected signals disappear, does it?
>>>>>
>>>>> Only only planet Phil does the world work this way.
>>>>>
>>>>> d
>>>>
>>>> Don,
>>>>
>>>> Fair go mate.(wink)  You don't honestly thinks a toaster repairman 
>>>> needs
>>>> to know much about impedance mismatch do you?  :P
>>>
>>> Well actually I am afraid that Hybrid leakage is not actually anything 
>>> to
>>> do with impedance mis-match but rather more to do with Gain of a
>>> Transmission System and the way a Hybrid works - The gain of the System
>>> can (obviously) be affected by Impedance etc however the echo is not the
>>> result of reflected signals due to Impedance mismatch but rather from 
>>> the
>>> fact that a hybrid is not perfect.
>>> The term signal reflection is often used to refer to hybrid leakage but
>>> (and here is the rub) it is not Impedance reflection.
>>>
>>> http://www.commsdesign.com/design_corner/showArticle.jhtml?articleID=16502325
>>>
>>> http://www.iec.org/online/tutorials/echo_cancel/topic02.html
>>>
>>
>> Perhaps the issue of reflected power in an impedance mismatched circuit 
>> has
>> gone a bit too far and turned the corner.
>>
>> Any energy not absorbed by a termination must go somewhere - agreed?
>>
>> When the source and load impedances are matched all the power developed 
>> by
>> the source will be transferred to the load without loss.  As soon as 
>> either
>> of the impedances vary relative to the other an impedance mismatch 
>> occurs.
>> Under such conditions not all of the power developed by the source will 
>> be
>> transferred to the load.   This results in a loss of power which is 
>> directly
>> related to the degree (ratio) of impedance mismatch.
>>
>> Reflection loss is the logarithmic ratio of the incident power to the 
>> load,
>> that is, the ratio between the power that would be absorbed under 
>> correctly
>> matched conditions, to the power actually absorbed by the load.
>>
>> Therefore :
>>
>> Loss due to reflection (dB) = 10 x log(base10) Incident Power / Reflected
>> Power
>>
>> or in terms of the impedances -
>>
>> Loss due to reflection (dB) = 10 x log(base10) (Z1 + Z2) sq / 4Z1 x Z2
>>
>> As an example, a 1 : 1 ratio results in 0dB loss due to reflection.
>> A 2 : 1 mismatch results in 0.5dB loss due to reflection.
>> A 4 : 1 mismatch results in approximately 1.9dB loss due to reflection.
>> An 8 : 1 mismatch results in approx. 4dB loss due to reflection; and so 
>> on.
>>
>> Impedance mismatch does matter.
>>
>> Cheers,
>> Alan
>
> We are not actually worried here about the power lost by reflection, but
> the fact that it causes an echo - particularly noticeable on an
> international line with long delays. In this regard, even quite small
> mismatches are important. A mismatch which results in just 0.1dB of signal
> loss will cause an echo at -20dB. This will be very audible.
>
> d

Absolutely no disagreement there Don about the fact that impedance mismatch 
contributes to echo.

Echo (due to reflection) is a function of the return loss of the circuit. 
The formula is -

Return Loss (dB) = 20 x log(base10) (Z1 + Z2) / (Z1 - Z2)

A 2 : 1 mismatch (which results in a power loss of 0.5dB) corresponds to a 
return loss of 9.2dB which is pretty poor.  A mismatch which results in only 
0.1dB loss due to reflection actually achieves a healthy return loss in 
excess of 40dB, but as we both know that good figure rapidly deteriorates as 
the mismatch ratio increases.

In a transformer hybrid circuit, the return loss between the two paths of 
the 4 wire transmission circuits will never go below about 7dB due to 
trans-hybrid losses even if the impedance ratio between the load on the 2 
wire circuit is infinitely high with respect to that of the network 
connected to the balancing network port.  In practice, most telecos build 
loss into their transmission networks to counteract the effects of 
instability due to impedance mismatches.  In Australia, the loss from 2 wire 
appearance at the MDF of the exchange to the same at the other end is 
designed at 6dB for each direction of voice transmission.  A nice idea, as 
the echo level is reduced by a factor of 2 times the loss of the link in the 
network, resulting in a minimum stability margin of at least 12dB even under 
the worst possible conditions.

Some other overseas networks aren't anywhere near as good as ours when it 
comes to echo performance.  In fact a well designed and impedance matched 
network requires little or no echo cancellation equipment, resulting in a 
clearer network to talk over and minimal VF data transmission / fax 
transmission problems.

Cheers,
Alan