Filtering noise on tip ring line

[drpappy]

I have a floating tip ring line that is fed into a transformer on my telephone device. The output of the transformer is referenced to ground. I can filter the noise from the output side with an active bandpass filter. However...

My question is: How do I put the filter(s) on the floating side without messing up the line balance?

 

[Skogsgurra]

You could either use a passive filter or an active one with symmetric differential input and an insulated (low capacitance) PSU.

You need to careful about idle current and return loss as well. It is not an easy thing to do if you haven't done it before. The phone companies are kind of picky when it comes to adding non authorised equipment to the line.

Even if you put a filter on the transformer secondary, you need to make sure that return loss is OK. Primary impedance usually gets upset when you terminate the transformer other than it was designed for.

Gunnar Englund

http://www.gke.org

 

[drpappy]

Thanks. I was a little weary of shunting the tip and ring lines to ground with caps, thinking about ground loops.

I wonder if just bridging from tip to ring using capacitors or inductors (for 60Hz noise) would work? The signal is FSK. I know the 60Hz is common mode, and should cancel, but it's there on the output.

 

[Skogsgurra]

Normal telephone band-width is somewhere between 300 Hz and 3400 Hz. So a capacitor that takes 60 Hz away will kill your FSK signal as well. An inductor may help, but it would represent a path for the DC current that probably upsets the station relay (assuming that we are talking about a PSTN aka POTS).

Why do you think that 60 Hz is a problem? Shouldn't be, considering the 300 - 3400 Hz bandwidth.

Gunnar Englund

http://www.gke.org

 

[VEBill]

Nominal telephone VOICE bandwidth may be as stated, but don't forget about (for example) the AC ring voltage which can be 20 to 40 Hz and perhaps 90 VAC. In other words, think about all the other signals riding on the phone line besides just the voice before you start adding filters.

 

[Skogsgurra]

Yes, that's absolutely right. We have, for instance, a 25 Hz 110 V ring signal. It is not a nice experience to "hold the line" when the ring signal arrives.

Gunnar Englund

http://www.gke.org

 

[MacGyverS2000]

I remember many moons ago as a kid hanging out at a friend's house on the back porch (screened in). We were tasked with washing down the astroturf, so out came the garden hose and off came the shoes. Now I was smart enough to stay away from electrical outlets, but what kid would think to stay away from a phone jack. It was a mild tingle that we kept shooting the phone jack for, but it was years before I knew enough about POTS to understand where the tingling came from.

Sorry, just a side story this reminded me of...

Dan - Owner

http://www.Hi-TecDesigns.com

 

[drpappy]

It's one modem to another using a dedicated twisted pair, and is on all the time. I think the higher freq noise is riding in on 60Hz. This is because you can attenuate it just enough with an additional transformer (with less bandwidth, no DC; i.e. cheaper), and it seems to help, or at least move the noise. However, I don't think that's a consistent enough solution.

 

[BrianR]

If the 60Hz is a differential signal on the balanced line before the transformer, then I don't think you can get rid of it very easily. The transformer should get rid of common mode interference as you said. If however it is induced through that ground connection on the secondary side then that's where you start looking. A LPF is not an option so you need to look closely at balancing techniques.

 

[drpappy]

Thanks. I did try lifting the drain wire to the gnd, but no help. In about a week, I'm going to try DC blocking caps like skogsgurra talked about on a different thread. Although, I don't know how they will impact CMRR.