Re: Getting matching transformer from telephone

[Ross Herbert]

On Mon, 05 Jan 2009 18:25:18 +0000, Eeyore

:
:
:Ross Herbert wrote:
:
:> :Ross Herbert wrote:
:> :
:> :> Since your PC is mains powered and it may not have the required isolation
:> :> between the mains side and the sound card input you can do your own thing
:> :> using an approved 600:600 transformer with 3kV isolation rating to
interface
:> the
:> :> telephone line to the sound card input.
:> :
:> :He DOES NOT need a 600 ohm transformer since the input impedance of the
sound
:> :card is not 600 ohms <sigh> !
:>
:> Since the application is merely detecting signal "voltage" it hardly matters
:> that the secondary impedance of the transformer is 600 ohms and the input
:> impedance of the sound card is more like 10Kohms. The only reason one tries
to
:> match impedances is where one needs to maximise "power transfer" and that
:> doesn't apply in this case.
:
:Untrue. A non-optimally loaded audio transformer will not have a flat frequency
:response. Nor is it about power transfer.
:
:Graham

With regard to a POTS line the VF bandwidth is some 300 - 3400Hz - hardly hi-fi
- so optimal flat frequency response is not an issue.

The fact that the secondary impedance of the 600 ohm transformer does not match
the input impedance of the sound card is totally unimportant in this
application.

 

[krw]

On Sun, 04 Jan 2009 20:36:09 -0900, [email protected] (Floyd L. Davidson) wrote:

:
:> voice frequency circuits were all 300 - 3400Hz in my day.
:
:The PSTN is specified from 400 to 2800 Hz, with 24 dB SNR.
:
:Individual channels on various carrier systems, and some
rivate line voice circuits are specified with more
:bandwidth.
:

In Australia PSTN is specified for 300 - 3400 Hz bandwidth.

How well do V.92 modems work or is the bandwidth really 4K (8K line
cards, and such)?

 

[John Livingston]

On Sun, 04 Jan 2009 20:36:09 -0900, [email protected] (Floyd L. Davidson) wrote:

:
:> voice frequency circuits were all 300 - 3400Hz in my day.
:
:The PSTN is specified from 400 to 2800 Hz, with 24 dB SNR.
:
:Individual channels on various carrier systems, and some
rivate line voice circuits are specified with more
:bandwidth.
:

In Australia PSTN is specified for 300 - 3400 Hz bandwidth.

Also in the UK. Floyd seems to be referring to the US Bell standards.
The CCITT standards (which apply in most other parts of the world) are
different.

John

 

[John Livingston]

You mean ITU-T standards, which is what both the UK and
Australia use. (It hasn't been CCITT for a long time.)

You are citing the bandwidth for an individual *channel*.
It is not possible to provide that sort of bandwidth on
every local loop, and therefore significantly less is
required for a complete connection or for any individual
loop.

Are we discussing the same thing, I wonder ? The "Local Loop" is the
exchange - subscriber path, and is not bandwidth limited other than by
the basic line parameters. Hence the reason a local loop can (typically)
support 8Mbit ADSL.

Moreover, the 400-2800 Hz is not what you'd expect. It
is actually specified as up to 3 dB of rolloff per local
loop on any given connection. Plus up to 8 dB of
rolloff in the network switching fabric. Added up, that
means that both at 400 Hz and at 2800 Hz it would be
within specifications to have a connection that has 14
dB of rolloff compared to 1008 Hz.

Which standard do you quote from ?
Could you cut and paste some text from this standard in support of this
statement ?

Indeed, you will find that all of the ITU modem
standards (v.32, v.34, etc.) specify performance for a
400-2800 Hz connection as the minimal "voice grade"
specification.

This may well be the case as the designers will work to the worst
(international) case.

John

 

[Eeyore]

The PSTN is specified from 400 to 2800 Hz, with 24 dB SNR.

By WHOM ?

Graham

 

[Eeyore]

Are we discussing the same thing, I wonder ? The "Local Loop" is the
exchange - subscriber path, and is not bandwidth limited other than by
the basic line parameters. Hence the reason a local loop can (typically)
support 8Mbit ADSL.

Or up to 24 Mbps ADSL2+ !

Graham

 

[Eeyore]

I doubt it.

You're just plain wrong about everything.

"Most of the current telephone systems are still restricted to the historically
motivated limitation of the bandwidth from 0.3 to 3.4 kHz."
http://portal.acm.org/citation.cfm?id=1142855&dl=&coll=

http://advancingphysics.iop.org/previous/wb/teacher/BandwidthW6.pdf

Graham

 

[Eeyore]

:Ross Herbert wrote:
:> :Ross Herbert wrote:
:> :
:> :> Since your PC is mains powered and it may not have the required isolation
:> :> between the mains side and the sound card input you can do your own thing
:> :> using an approved 600:600 transformer with 3kV isolation rating to
interface the telephone line to the sound card input.
:> :
:> :He DOES NOT need a 600 ohm transformer since the input impedance of the
sound card is not 600 ohms <sigh> !
:>
:> Since the application is merely detecting signal "voltage" it hardly matters
:> that the secondary impedance of the transformer is 600 ohms and the input
:> impedance of the sound card is more like 10Kohms. The only reason one tries
to match impedances is where one needs to maximise "power transfer" and that
:> doesn't apply in this case.
:
:Untrue. A non-optimally loaded audio transformer will not have a flat frequency
:response. Nor is it about power transfer.
:
:Graham

With regard to a POTS line the VF bandwidth is some 300 - 3400Hz - hardly hi-fi
- so optimal flat frequency response is not an issue.

The fact that the secondary impedance of the 600 ohm transformer does not match
the input impedance of the sound card is totally unimportant in this
application.

So your attitude is "it's so bad it doesn't matter messing it up even more".

A proper 600:600 transformer will be more expensive than a 10k:10k one too.

Graham

 

[Palindrome]

PCPaul wrote:

I know
loads of UK service personnel. And they are unanimously nervous of being
posted to places where they have to fight alongside US forces. That ain't
right.

Did you ask them what they felt like about fighting alongside French,
German or Italian forces? I'd bet a pineapple upside down cake that they
would have been even more nervous... However, let's face it, the odds
are that if anyone is supplying someone to fight alongside Brits - it is
going to be the US and so the possibility becomes a realistic and not
just a theoretical one.

Of course a Brit is at greater danger from a Yank than another Brit -
the odds are that the Yank actually has a gun/missile/helicopter/torpedo
that /actually works/. Or, more to the point, actually *has* one,
without having to go and scrounge it from another Yank.

There is greater risk, of course. The Yank may not be aware of the
British Army's traditions. Such as sending out soldiers with battlefield
communications that don't work. In aircraft where spilt fuel has to be
mopped off the deck after each IFR. With body armour that they have to
take in turns wearing...

I don't know many UK service personnel. Those I do have said that, when
they have been in a fix, they have prayed for US air support to turn up.
Not because it will be any better than Brit air support - but because
they are praying for something that could actually happen and not just
hoping for a miracle.


US casualties in Iraq = 4147 US population =304 million
Rate = 13 per million

UK casulaties in Iraq = 178 UK population = 60 million
Rate = 3 per million

Yep, that sounds a good reason to feel nervous about fighting alongside
the US military - they go into harm's way.

 

[tony sayer]

Or up to 24 Mbps ADSL2+ !

Graham

When you live -in- an exchange that is;!...

 

[Palindrome]

Michael A. Terrell wrote:

BTW, I seem to be meeting a lot of young women who recently moved
into Central Florida from the UK. One was complaining about her accent,
and upset that her sister had already lost most of hers. She said she
was from north of London and I found the accent interesting.

"North of London", in terms of British accents, covers just about all of
them. I don't think anyone ever loses an accent - at most it is
temporarily mislaid. They tend to return to the surface when
circumstances dictate, with or without any deliberate intention.

I'd suggest that accents, even regional accents, really become
interesting when used in a different language. Which, from what I
understand of Florida, could easily have applied in the case you
mention..

 

[Ross Herbert]

On Tue, 06 Jan 2009 04:53:17 -0900, [email protected] (Floyd L. Davidson) wrote:

:>On Sun, 04 Jan 2009 20:36:09 -0900, [email protected] (Floyd L. Davidson)
wrote:
:>
:>:
:>:> voice frequency circuits were all 300 - 3400Hz in my day.
:>:
:>:The PSTN is specified from 400 to 2800 Hz, with 24 dB SNR.
:>:
:>:Individual channels on various carrier systems, and some
:>rivate line voice circuits are specified with more
:>:bandwidth.
:>:
:>
:>In Australia PSTN is specified for 300 - 3400 Hz bandwidth.
:
:I doubt it.


Well let's give some examples...

When I was involved in junction commissioning (unloaded copper inter-exchange VF
junctions) with Telstra, transmission measurements were carried out over the
300-3400Hz range. This hasn't altered.

When looking at the transmission characteristics of an international telephone
exchange (ITU-T Recommendations)the only frequency range mentioned is
300-3400Hz.
http://www.itu.int/rec/dologin_pub.asp?lang=e&id=T-REC-Q.45-198410-I!!PDF-E&type=items
http://www.itu.int/rec/dologin_pub.asp?lang=e&id=T-REC-Q.45bis-198811-I!!PDF-E&type=items

And the following document recommends all channel terminal equipment be lined up
using 300-3400Hz.
http://www.itu.int/rec/dologin_pub.asp?lang=e&id=T-REC-G.120-199812-I!!PDF-E&type=items

Note that ITU-T G235 (3KHz spacing) is supesrseded and is no longer recommended
for international connections.

As for the specification relating to customer equipment connected to the PSTN
the frequency range used for testing is 100Hz - 4KHz.
http://www.commsalliance.com.au/documents/standards/S004:2008

In other documents from this website the definition of VF telephony or Voiceband
is 300-3400Hz.

There probably a number of other publications if I had the time to research them
but suffice to say that in Australia the VF telephony channel bandwidth is
specified as per ITU recommendations ie. 300-3400Hz.

 

[krw]

Thank you Sue. I knew there had to be someone in the UK with common
sense.

BTW, I seem to be meeting a lot of young women who recently moved
into Central Florida from the UK. One was complaining about her accent,
and upset that her sister had already lost most of hers. She said she
was from north of London and I found the accent interesting.

My SIL is a Brit. In the 35+ years she's been here she hasn't lost
her accent, though has lost her ignorance.

 

[John Livingston]

On Tue, 06 Jan 2009 04:53:17 -0900, [email protected] (Floyd L. Davidson) wrote:

:>On Sun, 04 Jan 2009 20:36:09 -0900, [email protected] (Floyd L. Davidson)
wrote:
:>
:>:
:>:> voice frequency circuits were all 300 - 3400Hz in my day.
:>:
:>:The PSTN is specified from 400 to 2800 Hz, with 24 dB SNR.
:>:
:>:Individual channels on various carrier systems, and some
:>rivate line voice circuits are specified with more
:>:bandwidth.
:>:
:>
:>In Australia PSTN is specified for 300 - 3400 Hz bandwidth.
:
:I doubt it.


Well let's give some examples...

When I was involved in junction commissioning (unloaded copper inter-exchange VF
junctions) with Telstra, transmission measurements were carried out over the
300-3400Hz range. This hasn't altered.

When looking at the transmission characteristics of an international telephone
exchange (ITU-T Recommendations)the only frequency range mentioned is
300-3400Hz.
http://www.itu.int/rec/dologin_pub.asp?lang=e&id=T-REC-Q.45-198410-I!!PDF-E&type=items
http://www.itu.int/rec/dologin_pub.asp?lang=e&id=T-REC-Q.45bis-198811-I!!PDF-E&type=items

And the following document recommends all channel terminal equipment be lined up
using 300-3400Hz.
http://www.itu.int/rec/dologin_pub.asp?lang=e&id=T-REC-G.120-199812-I!!PDF-E&type=items

Note that ITU-T G235 (3KHz spacing) is supesrseded and is no longer recommended
for international connections.

As for the specification relating to customer equipment connected to the PSTN
the frequency range used for testing is 100Hz - 4KHz.
http://www.commsalliance.com.au/documents/standards/S004:2008

In other documents from this website the definition of VF telephony or Voiceband
is 300-3400Hz.

There probably a number of other publications if I had the time to research them
but suffice to say that in Australia the VF telephony channel bandwidth is
specified as per ITU recommendations ie. 300-3400Hz.

Good, accurate stuff, Ross.

I really don't know where Floyd L. Davidson is coming from. He is grimly
sticking to an assertion the the UK and Australian PSTN is specified as
"400 to 2800 KHz", but has been unable to quote his source in terms of
unambiguous specifications.

I can only assume - from his location - that he has some experience as a
technician in military comms systems, and has some view of the US Bell
system. Hence the belief in "400-2800". As has been fully demonstrated
by several others, this is a profound misunderstanding of the
international PSTN.

This thread has wandered far and wide from the OP's question, but it's
getting a bit too far off the subject for me. It's also degenerated into
a classic battle of the Trolls.

If Floyd wishes to make further wild and inaccurate statements, perhaps
he would like to open a new thread ?

John

 

[Ross Herbert]

On Wed, 07 Jan 2009 06:52:48 -0900, [email protected] (Floyd L. Davidson) wrote:

:>
:>Good, accurate stuff, Ross.
:
:All of it referenced the specifications for individual
:channels on various facilities. None of it had to do
:with the overall minimum allowed specification for an
:end to end connection via the PSTN.

Hold it, that's a dumb statement.

The ITU recommendation for international VF channel bandwidth is 300-3400Hz, so
why would it be necessary to "allow a minimum" bandwidth of 400-2800Hz? If the
300-3400Hz applies to ALL channels used in an end-to-end international link,
then it follows that the overall bandwidth is 300-3400Hz.

Obviously the ITU spec is greater than the so-called "minimum" of 400-2800Hz, so
it is not necessary to "allow a minimum" bandwidth of less than this.

The fact that some administrations may not have adopted the ITU recommendation
and continue to use 400-2800Hz simply means that they are not prepared to
upgrade their equipment and are therefore behind the times. Any communication
carried over such links will mean the overall bandwidth is degraded even if some
sections do conform to the ITU recommendation of 300-3400Hz.

:
:>I really don't know where Floyd L. Davidson is coming from.
:
:More than three decades in the long distance
:telecommuncations industry (but not with the Bell
:System).

And I have 37 years of Telco experience from 56 - 93 in both metro and long-line
installations. In all those years we used 300-3400Hz.

 

[Eric Tappert]

On Wed, 07 Jan 2009 06:52:48 -0900, [email protected] (Floyd L. Davidson) wrote:

:>
:>Good, accurate stuff, Ross.
:
:All of it referenced the specifications for individual
:channels on various facilities. None of it had to do
:with the overall minimum allowed specification for an
:end to end connection via the PSTN.

Hold it, that's a dumb statement.

The ITU recommendation for international VF channel bandwidth is 300-3400Hz, so
why would it be necessary to "allow a minimum" bandwidth of 400-2800Hz? If the
300-3400Hz applies to ALL channels used in an end-to-end international link,
then it follows that the overall bandwidth is 300-3400Hz.

Obviously the ITU spec is greater than the so-called "minimum" of 400-2800Hz, so
it is not necessary to "allow a minimum" bandwidth of less than this.

The fact that some administrations may not have adopted the ITU recommendation
and continue to use 400-2800Hz simply means that they are not prepared to
upgrade their equipment and are therefore behind the times. Any communication
carried over such links will mean the overall bandwidth is degraded even if some
sections do conform to the ITU recommendation of 300-3400Hz.

:
:>I really don't know where Floyd L. Davidson is coming from.
:
:More than three decades in the long distance
:telecommuncations industry (but not with the Bell
:System).

And I have 37 years of Telco experience from 56 - 93 in both metro and long-line
installations. In all those years we used 300-3400Hz.

The smaller bandwidth was used on undersea cables to increase capacity
(16 channels in the same bandwidth as 12 standard channels). Such
usage has gone away with the introduction of fiber optic cables. TASI
(Time Assignment Speech Interpolation) is still used to about double
the capacity of modern undersea cables. The operating principle of
TASI is to disconnect a channel with no voice detected and give the
circuit to an active user. This game is played continuously and
allows the operator to about double the capacity of a link. The
downside is the speech detector tends to clip the first syllable a
bit...

ET

 

[Ross Herbert]

On Tue, 06 Jan 2009 19:47:17 +0000, Eeyore

:
:
:Ross Herbert wrote:

:> With regard to a POTS line the VF bandwidth is some 300 - 3400Hz - hardly
hi-fi
:> - so optimal flat frequency response is not an issue.
:>
:> The fact that the secondary impedance of the 600 ohm transformer does not
match
:> the input impedance of the sound card is totally unimportant in this
:> application.
:
:So your attitude is "it's so bad it doesn't matter messing it up even more".

Not at all. Any practical measurement of the degradation of a 300-3400Hz signal
using the 600 ohm transformer would be insignificant. If your assumption that
audio degradation when using a 600:600 transformer was valid then why is there
no available audio transformer, with a 3.5Kv isolation rating, to match a 600
ohm telephone line to a high impedance (eg. 10K ohm) input? The answer has to
be, "it's not necessary".

:
:A proper 600:600 transformer will be more expensive than a 10k:10k one too.
:

That may be true but how many readily available 10K:10K audio transformers have
a 3.5Kv isolation rating?

In order to conform to the required isolation rating of 3.5Kv between the mains
powered PC and the telephone line, the only transformer which is readily
available, and complies, is an approved 600:600 transformer.

 

[tony sayer]

On Wed, 07 Jan 2009 06:52:48 -0900, [email protected] (Floyd L. Davidson) wrote:

:>
:>Good, accurate stuff, Ross.
:
:All of it referenced the specifications for individual
:channels on various facilities. None of it had to do
:with the overall minimum allowed specification for an
:end to end connection via the PSTN.

Hold it, that's a dumb statement.

The ITU recommendation for international VF channel bandwidth is 300-3400Hz, so
why would it be necessary to "allow a minimum" bandwidth of 400-2800Hz? If the
300-3400Hz applies to ALL channels used in an end-to-end international link,
then it follows that the overall bandwidth is 300-3400Hz.

Ummmm... those figures are not -quite- that meaningful unless we have
some sort of reference level ..

i.e. 300 to 3400 +/- ? dB ...

 

[Ross Herbert]

:
:What do you suppose happens to the bandwidth of a signal as it
:is passed through successive audio devices that are band
:limited to 300-3400 Hz?
:
:Would you expect to get 300-3400Hz +- ndB response in an end
:to end loopback test when you use white noise as the test
:signal? (n =ITU spec)
:
:Going from a handset to a PBX to MUX to a CO into the network
:and back if each device has a 300-3400Hz response the looped
:back signal should be more band limited than the first device
:in the signal chain with 300-3400Hz response originating test
:signal by quite some bit.
:

I suggest you do some research on the various ITU documents, eg.
http://www.itu.int/rec/T-REC-P/e
http://www.itu.int/rec/T-REC-O/e

or pick any relevant section from the main page
http://www.itu.int/ITU-T/publications/recs.html

 

[Eeyore]

ITU-T standards and Bellcore standards.

Bell don't count as they're not International.

Since everyone else says ITU-T say 300-3400Hz I suspect we're seeing
another case of you being stuck in a time warp.

Graham