Whats the best way to isolate A2D on a telephone line?

[DavidP]

A/D on telephone line

What would be the best way to isolate an A2D (like ADC0803) connected to a telephone line in order to measure DC voltage and AC voltage (while ringging) on it. With AC it is easier as a 600R can be connected between the line and the other side of the transformer may be grounded, but with DC....

Thanks

David

 

[arturt134]

Re: A/D on telephone line

If you need only to make a fact: is ringing voltage or not, use opto isolation and a capacitor to block dc voltage.
If you need to know what exactly voltage is, use AD on line side (powered from the line), then opto isolation use between AD and uC.

 

[papyaki]

Re: A/D on telephone line

Hi,

Just an idea to mesure dc voltage without isolation. Why not use a sample and hold capacitor with relay. On the sample phase, the capacitor is connected to the line an remain dc voltage. On hold phase, the capacitor is switched to the AD inputs, which do the measurment.

 

[C-Man]

Re: A/D on telephone line

DavidP what exactly do you want to do?

Normally i am only interested if the phone is ringing or not, this can be detected with a simple optocoupler, resistor and series capacitor (have done that many times).

And why do you need to know the line voltage?
Do you for example want to detect if somebody picked up a telephone?

If yes a good (but expensive) solution is a DAA for example from Mitel (I use MH88435-P for a speaking remote controller).

You can check out the specs at the mitel website ...

best regards

 

[papyaki]

Re: A/D on telephone line

Hi,

DC value and polarity gives you the state of the line and the state of a call, eg. "caller off-hook/on-hook", "called off-hook/on-hook".

 

[Pazam]

Re: A/D on telephone line

Dear DavidP,

If you want to know is the hook is on or off,
you can use a simply optocoupler as a switch.

If you want to measure the actual voltage,
you can use a simple phototransistor based
optocoupler (NPN for example).
Look for a graph in the data sheet of
CURRENT TRANSFER RATIO vs. FORWARD CURRENT.
This graph will describe the ration between
the LED forward current and the transistor
gain.
Find a linear zone and use it.
As the LED forward current will change, the
current through the phototransistor will
change. Use a small resistor at the collector
(or emitter) to measure this current.

Be aware that calibration might be needed per
circuit.

It should work

Warning:
Be careful when you deal with the DC componenet.

Ususally none of the lines has any relation to the
common AC network ground.

This means that when you deal with the telephone
lines, you have to well isolate grounds and power
supplies.

Regards,

Pazam

 

[DavidP]

Re: A/D on telephone line

I thank all memebers trying to help me. Some of the ideas are really usefull.

The measurements are needed for PABX switches. I'm improving my automatic multi port tester. For this reason I must measure actual voltage on the line. With hook on I have to check that the voltage is not less then 30Vdc and not more then 53Vdc. with hook off the voltage should be 12Vdc and not less then 8Vdc. In addition the ringing voltage
is to be measured to be around 70Vac. Measurements are processed by an A/D connected to a nice 89S8252. Next step is to check for dial tone presence. This is not so difficult as i can check the audio signal at the earphone, and not directly on the line. More complicated is to detect presence of unwanted noises on the line. Since it is an automatic device there is no posibility to listen to the signal. The signal is to be tested by the micro.

I'll be glad to hear forther inputs on those subjects.

 

[Pazam]

Re: A/D on telephone line

Dear DavidP

Now I understand your needs better.

If you're building a test bench you should consider using a
differential probe. It would be most accurate and simple to use.
The only down side is the price of the probe.

If you're looking for a cheaper solution, I think my idea of using
an optocoupler to measure the DC voltage should suite your needs.


Recognizing signals and noise should be easy if you use FFT.
You can use the processor to sample the telephone line and then
send the data to a PC for further analysis.

Regards,

Pazam