Maker Pro
Maker Pro

Phone line interface: Replacing a 1:1 transformer with an optocoupler

O

OBones

Jan 1, 1970
0
Hi all,

I'm designing a system that uses a MT8870 DTMF decoder to detect tones
on a phone line and the typical phone line interface is done via a 1:1
600ohm "telephony" transformer, like this :



-. ,--------o------ To MT8870
Line )|( |
)|( z
-' '-. A
| |
Gnd |
V
z
|
Gnd

The two zeners are here to ensure the voltage does not go above Vcc
applied to MT8870.
My problem is that the transformers for these applications are bulky and
take a lot of space. Hence I was wondering if I could replace the
transformer by a diode bridge and optocoupler pair like this:

Vcc
|
___ .----------. |
+-+----|___|---|-. .-|-'
A A | | |/ |
Line -+ | | V -> | |
-(-+ | - |> |
A A | | | |
+-+------------|-' '-|-- To MT8870
'----------'

My main concern is the frequency that the optocoupler can support. I
have a bunch of 4N25 here and considering the max rise time and fall
time with a security margin, I'd say it can transfer signals up to at
least 30kHz which to me is enough for this application as DTMF signals
only go up to 2kHz.
Can anyone confirm that my calculations are right, and that the
arrangement of the transistor side of the 4N25 is fine?

Thanks a lot for your help.
Cheers
Olivier
 
R

Ross Herbert

Jan 1, 1970
0
Hi all,

I'm designing a system that uses a MT8870 DTMF decoder to detect tones
on a phone line and the typical phone line interface is done via a 1:1
600ohm "telephony" transformer, like this :



-. ,--------o------ To MT8870
Line )|( |
)|( z
-' '-. A
| |
Gnd |
V
z
|
Gnd

The two zeners are here to ensure the voltage does not go above Vcc
applied to MT8870.
My problem is that the transformers for these applications are bulky and
take a lot of space. Hence I was wondering if I could replace the
transformer by a diode bridge and optocoupler pair like this:

Vcc
|
___ .----------. |
+-+----|___|---|-. .-|-'
A A | | |/ |
Line -+ | | V -> | |
-(-+ | - |> |
A A | | | |
+-+------------|-' '-|-- To MT8870
'----------'

My main concern is the frequency that the optocoupler can support. I
have a bunch of 4N25 here and considering the max rise time and fall
time with a security margin, I'd say it can transfer signals up to at
least 30kHz which to me is enough for this application as DTMF signals
only go up to 2kHz.
Can anyone confirm that my calculations are right, and that the
arrangement of the transistor side of the 4N25 is fine?

Thanks a lot for your help.
Cheers
Olivier

Olivier,
Provided that you don't have any DC current flowing in a 600:600
transformer they don't need to be big and clumsy. Since you say you
only want to "detect dtmf" then you should be able to arrange your
transformer with a blocking capacitor in series with the primary
across the line.

A 220nF 250Vdc film cap and a Tamura MG-21 (600:600) miniature audio
transformer which has a 3dB freq range from 100Hz - 100kHz would be
suitable. The transformer measures only 8mmx10mmx11mm.
http://www.tamura-ss.co.jp/en/electronics/trance03/pdf/mg_mx.pdf

I dug up a circuit for a dtmf monitor here
http://www.glolab.com/freeinfo/kdial.pdf and it uses a 100nF cap in
series with primary across the line.
 
M

mpm

Jan 1, 1970
0
Hi all,

I'm designing a system that uses a MT8870 DTMF decoder...
           -. ,--------o------ To MT8870
     Line   )|(        |
            )|(        z
           -' '-.      A
                |      |
               Gnd     |
                       V
                       z
                       |
                      Gnd

Oliver - Two things:
First, is the MT8870 still a valid part?? I thought these things were
dinosaurs!! (Hard to find, not recommended for new designs, etc...)
Unless this is a one-off design? I could be wrong though.

Next, How do I use the text above? (The text that describes your
circuit.)
Is there some program that interprets this becuase I could not make
sense of it, even when I tried to re-print it in Courier font, etc..
This issue arises from time to time and I am unable to help people out
because I can't figure out their circuit design.

Oh, one more thing: CML Microcircuits (www.cmlmicro.com) is a good
source for all things "IC Telecom related". The CMX605 Digital to
Analog (POTS) Signalling Interface with DTMF Encoder/Decoder might be
just the thing for you. (In case the MT8870 really has gone belly-
up.)?
I don't recall if the 605 requires a transformer for interfacing..? I
don't think so??

-mpm
 
A

Anthony Fremont

Jan 1, 1970
0
mpm said:
Oliver - Two things:
First, is the MT8870 still a valid part?? I thought these things were
dinosaurs!! (Hard to find, not recommended for new designs, etc...)
Unless this is a one-off design? I could be wrong though.

http://www.futurlec.com/ICSFZarlink.shtml
mt8870 - $1.50 each
Next, How do I use the text above? (The text that describes your
circuit.)
Is there some program that interprets this becuase I could not make
sense of it, even when I tried to re-print it in Courier font, etc..
This issue arises from time to time and I am unable to help people out
because I can't figure out their circuit design.

Oh, one more thing: CML Microcircuits (www.cmlmicro.com) is a good
source for all things "IC Telecom related". The CMX605 Digital to
Analog (POTS) Signalling Interface with DTMF Encoder/Decoder might be
just the thing for you. (In case the MT8870 really has gone belly-
up.)?
I don't recall if the 605 requires a transformer for interfacing..? I
don't think so??

I sure didn't have any luck with cmlmicro. I wanted some (c)mx614 modem
chips once and tried sampling, no response at all. I tried buying, no luck
there either. Do you know where to get their stuff in small quantities?
 
O

OBones

Jan 1, 1970
0
mpm said:
Oliver - Two things:
First, is the MT8870 still a valid part?? I thought these things were
dinosaurs!! (Hard to find, not recommended for new designs, etc...)
Unless this is a one-off design? I could be wrong though.

This is a one off, or a two off if I fry the first one ;-)
And that circuit is easy to find around here, for hobbyists that is.

Next, How do I use the text above? (The text that describes your
circuit.)
Is there some program that interprets this becuase I could not make
sense of it, even when I tried to re-print it in Courier font, etc..
This issue arises from time to time and I am unable to help people out
because I can't figure out their circuit design.

Nothing much, it should just be pure ASCII text. I can see it just fine,
and others can as well.

Oh, one more thing: CML Microcircuits (www.cmlmicro.com) is a good
source for all things "IC Telecom related". The CMX605 Digital to
Analog (POTS) Signalling Interface with DTMF Encoder/Decoder might be
just the thing for you. (In case the MT8870 really has gone belly-
up.)?
I don't recall if the 605 requires a transformer for interfacing..? I
don't think so??

Well, it better have some sort of protection as the phone line is about
100 V when ringing and 48 when talking, while the circuit usually is 5V
only. On top of that, I'd rather have some sort of isolation if
lightning goes into the line...
 
O

OBones

Jan 1, 1970
0
Ross said:
Olivier,
Provided that you don't have any DC current flowing in a 600:600
transformer they don't need to be big and clumsy. Since you say you
only want to "detect dtmf" then you should be able to arrange your
transformer with a blocking capacitor in series with the primary
across the line.

Well, a regular phone line has a DC component, if I'm correct.

A 220nF 250Vdc film cap and a Tamura MG-21 (600:600) miniature audio
transformer which has a 3dB freq range from 100Hz - 100kHz would be
suitable. The transformer measures only 8mmx10mmx11mm.
http://www.tamura-ss.co.jp/en/electronics/trance03/pdf/mg_mx.pdf

I trust you on this, but how do I access this for a one off series, when
I generally take the parts from the local hobbyist shop?
I'd rather use the optocoupler isolation as it uses readily available
parts that even come up in SMT packages, which are definitely smaller
and easier on the layout than a through hole part.

I dug up a circuit for a dtmf monitor here
http://www.glolab.com/freeinfo/kdial.pdf and it uses a 100nF cap in
series with primary across the line.

Thanks for that.
 
J

Jan Panteltje

Jan 1, 1970
0
Well, a regular phone line has a DC component, if I'm correct.



I trust you on this, but how do I access this for a one off series, when
I generally take the parts from the local hobbyist shop?
I'd rather use the optocoupler isolation as it uses readily available
parts that even come up in SMT packages, which are definitely smaller
and easier on the layout than a through hole part.

I have used optocoupler for audio from phone lines.
Firts quastion you must ask then is : Do you have a second isolated supply?
In my case I had thta, so I could drive the LED of the opto cooupler
with an opamp current source (linearity).
Just picked the voltage of the line.


Uin opamp
-||-- +
out ---------
|
--- - |
| \ / / LED
| ---
---------------------|
|
[ ]
|
----

In case a switchmode is present the extra suply is just some extra turns on
a potcore.


At the reception side (different opamp!!) I had the photo transistor
drive into the inverting input ... low impedance, good high frequency response,
good linearity.

In the example the OP (you?) gave the LED already draws quite a bit of current.
I am not sure about the linearity of that circuit.
 
A

Anthony Fremont

Jan 1, 1970
0
OBones said:
Ross Herbert wrote:

Well, a regular phone line has a DC component, if I'm correct.

It does, but you don't have to pass DC thru the transformer to get audio off
the line. In fact, this is a good way to monitor a line without placing
much load on it.
http://www.tkk.fi/Misc/Electronics/circuits/tele1.gif

Of course this all relies on another phone (or something) to allow DC to
pass on the line to bring the phone line to off hook status. There's allot
of decent information about different ways to interface to the line right
here:
http://www.tkk.fi/Misc/Electronics/circuits/teleinterface.html
I trust you on this, but how do I access this for a one off series,
when I generally take the parts from the local hobbyist shop?
I'd rather use the optocoupler isolation as it uses readily available
parts that even come up in SMT packages, which are definitely smaller
and easier on the layout than a through hole part.

Let us know how that works out, I'm very interested. I've never seen it
done that way, but it may work fine. You're going to have to place a pretty
severe load on the line to do it though IMO. I prefer the DC blocked
transformer with clamping diodes myself. Watch out for those ringing
voltages.
 
M

Michael Black

Jan 1, 1970
0
OBones said:
My problem is that the transformers for these applications are bulky and
take a lot of space. Hence I was wondering if I could replace the
transformer by a diode bridge and optocoupler pair like this:
If you'd asked in the right place, instead of adding to the clutter of
this cesspool, you'd get a very simple answer.

Pull a junk modem out of a junk computer or off a shelf somewhere, and
use the parts from it to make the coupler. They all have the needed
transformer and they are in no way bulky. The rest of the components
right down to any needed optocouplers would be there too.

Cordless phones would also offer up a suitable transformer in
their base unit.

Likely answering machines, too.

Michael
 
D

Don Bowey

Jan 1, 1970
0
This is a one off, or a two off if I fry the first one ;-)
And that circuit is easy to find around here, for hobbyists that is.



Nothing much, it should just be pure ASCII text. I can see it just fine,
and others can as well.

Not very "pure."

For some reason your schematic's spaces appear as question marks (?).

Does this look better? I did not fix the bottom line, but left it as it
was.
 
R

Ross Herbert

Jan 1, 1970
0
Well, a regular phone line has a DC component, if I'm correct.

Of course... However, since you are only "detecting dtmf" you don't
need to pass the DC current on the line through the transformer
primary. You will simply be "sniffing" the dtmf which is ac signal
level so that is why you couple the primary via a cap.

Unless you are also carrying out line looping (on-hold) functions on
the telco side of the transformer, then you don't need a large
transformer. The reason the ones you have seen/used are so large is
because they have to handle a fairly large dc current in the primary
because they are intended for line loop and hold functions on the
primary side.
I trust you on this, but how do I access this for a one off series, when
I generally take the parts from the local hobbyist shop?
I'd rather use the optocoupler isolation as it uses readily available
parts that even come up in SMT packages, which are definitely smaller
and easier on the layout than a through hole part.

Mouser have the Tamura MET-31
http://www.mouser.com/search/ProductDetail.aspx?R=MET-31virtualkey55420000virtualkey838-MET-31
(US market) and is identical to the Japanese MG-21 except for case
color. They want a hefty price though ($13 per 1). I can post you an
MG-21 if you don't have any success with your opto experiment.
 
O

OBones

Jan 1, 1970
0
Michael said:
If you'd asked in the right place, instead of adding to the clutter of
this cesspool, you'd get a very simple answer.

Hum... I consider this a "design" question, hence the reason I put it in
sci.electronics.design. But if you have a better idea, please let me know.
 
O

OBones

Jan 1, 1970
0
Don said:
Not very "pure."

For some reason your schematic's spaces appear as question marks (?).

Does this look better? I did not fix the bottom line, but left it as it
was.

I saw the question marks in mpm answer as well, but I don't know where
they come from, I only see blank spaces in my original message.
And I used AACircuit to draw it, so I'm quite puzzled.
 
O

OBones

Jan 1, 1970
0
Ross said:
Of course... However, since you are only "detecting dtmf" you don't
need to pass the DC current on the line through the transformer
primary. You will simply be "sniffing" the dtmf which is ac signal
level so that is why you couple the primary via a cap.

Indeed. And I'm only sniffing the dtmf as there is a regular phone
handset on the line that will actually put the signal there.

Unless you are also carrying out line looping (on-hold) functions on
the telco side of the transformer, then you don't need a large
transformer. The reason the ones you have seen/used are so large is
because they have to handle a fairly large dc current in the primary
because they are intended for line loop and hold functions on the
primary side.

Mind you, by fairly large, I'm talking about 28mm square and 15mm height.

Mouser have the Tamura MET-31
http://www.mouser.com/search/ProductDetail.aspx?R=MET-31virtualkey55420000virtualkey838-MET-31
(US market) and is identical to the Japanese MG-21 except for case
color. They want a hefty price though ($13 per 1). I can post you an
MG-21 if you don't have any success with your opto experiment.

Thing is, I'm not residing in the US, but I will definitely go ahead
with the opto experiment.
 
O

OBones

Jan 1, 1970
0
Anthony said:
It does, but you don't have to pass DC thru the transformer to get audio off
the line. In fact, this is a good way to monitor a line without placing
much load on it.
http://www.tkk.fi/Misc/Electronics/circuits/tele1.gif
Interesting.


Of course this all relies on another phone (or something) to allow DC to
pass on the line to bring the phone line to off hook status.

There is a phone on the line, yes, so I'm only monitoring what the phone
is putting on the line.

Let us know how that works out, I'm very interested. I've never seen it
done that way, but it may work fine. You're going to have to place a pretty
severe load on the line to do it though IMO. I prefer the DC blocked
transformer with clamping diodes myself. Watch out for those ringing
voltages.

Well, I'm already using optos to detect line usage and ring "trains", my
only interrogation is on the bandwidth that the optos can let through.
 
M

mpm

Jan 1, 1970
0
I saw the question marks in mpm answer as well, but I don't know where
they come from, I only see blank spaces in my original message.
And I used AACircuit to draw it, so I'm quite puzzled.

It must be something screwy on this computer, I guess?
(I have no idea where to even start looking - XP strikes again!!)

What is "AACircuit"? Shareware?

-mpm
 
O

OBones

Jan 1, 1970
0
mpm said:
It must be something screwy on this computer, I guess?
(I have no idea where to even start looking - XP strikes again!!)

What is "AACircuit"? Shareware?

Andy's Ascii-Circuit
http://www.tech-chat.de/download.html

Website is in German, but the software is in English. Very nice to have
it when drawing schematics to be displayed in text only forums.
 
A

Anthony Fremont

Jan 1, 1970
0
OBones said:
Hum... I consider this a "design" question, hence the reason I put it
in sci.electronics.design. But if you have a better idea, please let
me know.

Just ignore him, it was a design question. He runs around playing net-kop
insulting people, encourages the soc.men and a.u.k. trolls, and then
complains that it's a cesspool in here. Can anyone say hypocrisy.
 
J

jasen

Jan 1, 1970
0
Hi all,

Vcc
|
___ .----------. |
+-+----|___|---|-. .-|-'
A A | | |/ |
Line -+ | | V -> | |
-(-+ | - |> |
A A | | | |
+-+------------|-' '-|-- To MT8870
'----------'

should work.

If you put the resistor before the bridge you can use 1N914
diodes for the bridge.

but ringing voltage can be over 100V (peak) andf if you shunt it you may
stop other devices from ringing.

it may be best to keep the bridge and put some sort
of current-limit in series with the resistor.

While we're discussing different approaches, how about a line-powered
blocking oscilator built on one side of an ethernet transformer
and an AM (or fm?) detector on the other side :)



Bye.
Jasen
 
J

jasen

Jan 1, 1970
0
If you'd asked in the right place, instead of adding to the clutter of
this cesspool, you'd get a very simple answer.

Pull a junk modem out of a junk computer or off a shelf somewhere, and
use the parts from it to make the coupler. They all have the needed
transformer and they are in no way bulky. The rest of the components
right down to any needed optocouplers would be there too.

yeah... some modems will even decode DTMF if you send the right commands,
the only modification needed would be a series capacitor to stop it from
seizing the line,

Bye.
Jasen
 
Top