On Tue, 18 Jun 2013 15:22:56 +0100, The Natural Philosopher
<t...@invalid.invalid> wrote:
>On 18/06/13 08:22, Jasen Betts wrote:
>> On 2013-06-17, Windmill <spam-n...@Onetel.net.uk.invalid> wrote:
>>
>>>> Most of this might already exist in something like Asterisk.
>>> Is there such a thing as an off-the-shelf Class D amplifier? (I know
>>> there are inexpensive ICs). Maybe one could be used together with one
>>> of the Pi's output pins to synthesize a ring voltage and pattern. (Step
>>> up the voltage with a small transformer if required.)
>> ring voltage is tricky, the phone ringer is a capacitive load the the
>> required voltage is higher than is convenient and bipolar.
>>
>> only a small current is needed but the low frequency means makes
>> finding a suitable transformer harder, because as reducing the frequency
>> also reduces the saturation voltage by the same factor.
>
>nah. You dont do it that way. You drive a HF inverter using a small
>ferrite core, rectify and smooth te outpout, but not too much, and
>simply chop the input to it.
>
>> you want about 100V out
It's been a good 30 years since I last worked in a strowger exchange
but I do clearly recall that the line polarity on an incoming call is
reversed with respect to that when the line is idle or engaged by an
outgoing call, along with the reason why.
The two legs of the balanced phone line are referred to as the 'A'
and 'B' legs of the line. When idle, galvanically speaking, the B leg
is connected to the negative of the exchange battery via one of the
two windings of the A relay in the calling cct (typically a linefinder
for residential customers and uniselector for business customers). The
A leg is connected to the exchange earth (positive connection of the
exch battery) via the other winding on the A relay .
For outgoing calls, the polarity remains the same as at idle. The 50v
supply allows for plenty of volt drop due to line resistance since the
phone only needs about 10% of this (it's current that's important,
circa 30mA or so for a carbon mic telephone handset) and, afaicr,
about 12 to 15 mA for a modern electronic phone.
On an incoming call, the Final Selector (FS) supplies the operating
DC current and the AC ringing current to ring the bell or tone sounder
of the called telephone. The ring current generator supply in the
exchange is a 75v RMS at 17.67Hz single phase supply[1] interrupted at
the appropriate ringing cadence (normally, the supply was generated by
a DC to AC motor/generator machine which also drove the cams which
generated the ring cadence, providing 3 different feeds which balanced
the generator loading in a timeshare fashion).
The reason for the polarity reversal was on account the ringing
current generator was an unbalanced source of AC which used a common
ground return. The ringing current would be applied to the B leg[2]
and the necessary -50v bias supply to detect when the phone had gone
'Off Hook' to answer the call and power the mic cct would be fed over
the A leg.
This meant that the relatively high voltage ac (peaks of 100v) didn't
have to be superimposed on top of a further 50v, neatly serving two
requirements, namely health and safety issues and minimising
complexity.
>IIRC is 70V DC that goes up and down when ringing..
If by that you mean a 70v DC chopped at 17Hz or so, that's one
possibility. Another one is to rapidly reverse the 50v at 15 to 17Hz
as ISTR doing with a home made setup in the early 70s using a fast
reversing relay.
>
>But actually driving a phone is a complext task.
It isn't, really.
>
>Better to buy a box that does it all.
In most cases that's tue, simply to avoid spending time that could
otherwise be used to earn the money to buy several such 'boxes'
>> perhaps a 50Hz 240V to 12V transformer run in reverse with
>> 5VAC at 20Hz on the 12V winding.
You could generate 25Hz from a cct locked to the 50Hz mains and a 5v
peak to peak should generate an 80v peak to peak output without
saturation ofthe transformer core.
>>
>> for 3ma out and with a transformer ratio of 20:1 you want a
>> transformer with a 60mA secondary.
You seem to be out by about a factor of 10. I'd be looking at a 25mA
output, hence a half amp 'secondary' rating.
[1] 16.67Hz (17Hz) was standard for public telephone exchanges but
other frequencies were used in PABXes, typically 25Hz which I believe
could be readily synthesised from the UK PSU's 50Hz mains supply -
it's quite likely the yanks and countries using the 60Hz PSU standard
might well have used 30Hz for the same reason. It's also possible that
a lower AC voltage may have been used in PABX equipment for the ringer
supply since they'd normally only be expected to service much shorter
line lengths to the extension phones within the site they were
designed to serve.
[2] On normal phone lines, the A leg would provide the return path for
the ringing current (the 250v DC rated 1.8 microfarad capacitor within
the phone (main if extension phones were involved or else the master
socket on the later and current plug in phone system) provided the
isolation from the -50v battery connection via the FS at the exchange
during the ringing phase prior to the called subscriber taking the
phone offhook to answer the call.
Of course, it was possible to arrange for the ringing current return
to be via the subscriber's local earth connection. This permitted a
means of providing shared service in parts of the exchange's coverage
area that experienced demand for customer service that had exceeded
the telephone planner's expectations.
A single phone line could provide automatic shared service to two
customers on a time shared basis. Shared service customers had to
agree to an etiquate with regard to the use of the line. Each customer
being expected to listen for an existing call by their shared party
before pressing the call button[3], apologising, if need be for
interrupting an existing call, optionally ascertaining how long they
might have to wait or else explain the need to make an emergency call
right then and there.
Incoming calls would automatically ring the appropriate customer's
telephone by virtue of the use of the local earth return cct at each
customers' premises and the fact that the Y customer's line would have
the A and B legs reversed with respect to that of the X customer such
that the 'hot' leg for the ringer connection would be on the B leg for
the X subscriber whilst the Y subscriber would use the A leg.
The "P" (Private) wires used by the exchange equipment associated
with each customer's number and calling equipment are jumpered
together on the IDF to allow the exchange equipment to correctly
return a busy signal whenever the line is engaged regardless of which
of the shared numbers is being called.
An unshared number would just simply have the single P wire jumpered
to link the calling equipment to the associated FS outlet's P wire
connection so that a 'line engaged tone' could be returned to the
caller in the event the line was busy due to an established incoming
or outgoing call.
Statistically speaking, this usually worked ok most of the time
(shared service was only offered to the low usage rate residential
customer) but it was possible for an incoming call to be accidentally
intercepted by the other customer trying to make an outgoing call.
When this happened, the caller would be advised of the shared line
situation and either asked to attempt another call or try later if the
calling customer who had answered by accident had an urgent need to
make their call. However, there was nothing to stop an obliging shared
partner from asking the caller to hang on whilst he alerted his
sharing neighbour of the incoming call.
The GPO (as it was when I joined the civil service those many years
ago) only offered shared service if both parties were aware of and
willing to accept the limitations and pitfalls of a shared service
party line.
The usual difficulty lay with finding an existing customer willing to
accept the reduction in their quarterly bill as sufficient inducement
to downgrade to a shared service line. In some cases, iirc, some
existing customers were chosen on the basis of minimum call usage and
offered the downgrade to shared service when the additional subscriber
was also an existing subscriber who had moved from another part of the
exchange's catchment area.
In these cases, the customer being asked to downgrade to a shared
service line could make strenuous objections which resulted in the GPO
testing the waters with other likely candidates fed from the same DP
until they found one that had the least objections or a better rental
reduction offer resolved the issue.
Shared service was deprecated a while before I transferred from
exchange maintainance to district maintainance duty so I very rarely
found myself embroiled in any shared service acrimony.
I believe another form of shared service, or party line working, was
practiced in the US which took the form of different ringer burst
counts to identify which of anywhere from 2 to 4 sharing customers the
call was intended for. In this case, the issue of who was responsible
for picking up the tab on outgoing calls could only have been resolved
by an operator so must have been a very early form of shared service
offered in remote rural areas that still relied on manual switchboard
working. Whether this form of party line operation was ever practiced
here in the UK, I couldn't say.
[3] The Call Button was required to allow the exchange equipment to
detect which of the two parties would be footing the bill for the
outgoing call. The mechanism relied upon the exch equipment detecting
which of the two legs of the line the earth calling signal appeared
on. As you can imagine, this system was open to abuse by a
knowledgable and crafty customer, as well as due to errors of
maintainance and repairwork which could introduce a line reversal.
These days, there are modern technology solutions for delivering
'shared service' transparently to high demand 'hot spots' not yet
served by sufficient line capacity - essentially, technology related
to that used by ADSL broadband.
It's been over two decades since I quit BT plus another five years on
top of that since I last dealt with any residential shared service
lines. I don't imagine that, after all this time, there'd still be any
'legacy' shared service lines still in service today.
--
Regards, J B Good