Telephony isolation

Hi all,

I am using an MT8870 chip to decode DTMF signals from a POTS line. To have an electrical isolation between my circuit and the telephone line, I'm using the recommended 1:1 600 ohms transformer. This works just fine, but the transformer is quite big for my liking and I wonder if it is possible to replace with something else. An opto-coupler with the emitter side across the line comes to mind, provided the voltage is correctly limited to avoid destroying the device. However, I am not sure this would work as the frequency and level of the audio signal might not be enough. Does anyone have any experience in this area?

Any opinions very welcome.

Cheers Olivier

Reply to
OBones
Loading thread data ...

I'm not sure about the optocouples, the right circuit just might work. But most of the transformers I've seen for this use are about 3/4" by

1/2" by 1/4" or so.

formatting link

Luhan

Reply to
Luhan

Well, I'd be glad to hear about one, and I looked around but did not find anything conclusive.

Well, yes, those I have are about that size too, but the height is a bit too much, and most of all, it does not allow to put anything in between it. With optos, I could arrange the circuit a bit differently.

Thanks Olivier

Reply to
OBones

Use a transformer with a smaller housing. These are available (Pulse Engineering), but it may need some asking around to get them.

--
Reply to nico@nctdevpuntnl (punt=.)
Bedrijven en winkels vindt U op www.adresboekje.nl
Reply to
Nico Coesel

Yes, what you need is a couple of four of those linear optocouplers and then you stick some opamps around them that make things behave like a transformer and your problem is solved.

DNA

Reply to
Genome

If you need a one-way audio path only, that is, from the phone line to your DTMF decoder then you may drive the optocoupler LED directly from the line through a suitable resistor. You can add a diode bridge to make the circuit insensitive to line polarity. Then you get your audio output from the output transistor of the optocoupler using a collector feedback resistor (you may start with 100K) and get the signal form the collector. You may have to fiddle a bit to get correct currents and signal levels.

On the other hand, tiny (dry) 600 ohm transformers are available.

Tom

Reply to
T. Atkin

Yes, a one way audio path is all I need.

Would something like that be ok?

Vcc | .-. | | | | '-' ___ .-------------. | +-+---|___|----. .---+--- MT8870 A A | | |/ | Phone -+ | | V -> -| | line -(-+ | - |> | A A | | | | +-+------------' '----. '-------------' | Opto | Gnd

(created by AACircuit v1.28.6 beta 04/19/05

formatting link

Thanks for any help. Olivier

Reply to
OBones

This won't work. First of all, there is 48V on the line so the LED will always be on or off. Secondly (with a capacitor in series), the amplitude of the DTMF tones is probably not sufficient to drive the LED. Third, you'll need some filtering / protection against the ringer voltage (75V AC)

--
Reply to nico@nctdevpuntnl (punt=.)
Bedrijven en winkels vindt U op www.adresboekje.nl
Reply to
Nico Coesel

have

but

possible

device.

the

your

line

circuit

optocoupler

the

If that's for a one off, I'd give a try to an opto that have the base connection available. That way he can servo the collector DC bias point into the linear region, (almost) whatever the phone DC line voltage is.

Of course that's a bit simpler if he have a neg supply on the MT8870 side.

The other way, if you plan on making lots of these, is to decode the dtmf on the phone line side, and cross the barrier with 2 small caps (33/47pF) like it's now done in those under $10 modems.

The one I have in hands right now uses a Silicon Labs Si3017-KS.

--
Thanks,
Fred.
Reply to
Fred Bartoli

It is for a one off.

How would I do that? I'm sorry, but I'm a bit new to the theory behind the linear region of transistors.

Meaning? It is supplied with +5V if that's what you mean. If it's related to the DTMF signal input, it has In+ and In- inputs, which can be used in differential or single ended mode, according to schematics in the datasheet.

So those would blow up in case of a problem?

Yes, but that's an all integrated modem chip, I'm just looking for a DTMG to logic combination conversion, which is just what the 8870 is doing.

Reply to
OBones

region,

Bias it so that Vce = VCC/2.

side.

In order to keep your DTMF signal across the opto you "just" don't want the transistor neither saturated nor blocked. Unfortunately, the line voltage variation and opto CTR huge tolerance probably don't allow a simple direct biasing. This means you allow enough current into the led to saturate the transistor under the worst case condition (lowest line, lowest CTR, led aging), then have a servo extract some base current to bring you the right bias point. Extracting base current is just easier with a negative supply.

Something like this:

VCC + | ___ .-. -|___|-. | | | | | V '-' - `\\ | || | \\ +-----+-----||-----> DTMF out -------' `> | | || OPTO |/ .-. .--| | | | |> | | | | '-' | === | || | GND +------||-. | | |\\ || | | '--|-\\ | | | >---+ | VCC/2 >-|+/ | \\| |/ | |------------------'

Reply to
Fred Bartoli

Thanks for your help, it sounds feasible and interesting.

Cheers

Reply to
OBones

The correct telephone circuit type opto-couplers are available with AC input dual LEDs so no bridge rectifier is needed, and telecoms type optocouplers also have more robust LEDs which can withstand upto 3x the peak overcurrent. If only one is needed a good place to look is scrap modem cards.

Reply to
ian field

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.