Telephone Interface

[longsoughtfor]

I am talking to a client who has asked for a telephone feed to monitor the session remotely. I figure I can take the mono out from my board and send it to the phone line and have seen a couple of sites with simple transformer coupled interfaces but my question is: Has anyone here had experince building a circuit to do this?

Thanks
Kevin.

 

[PRR]

> Has anyone here had experince building a circuit to do this?

Last time I did that was before CarterPhone, and we HAD to use a telco-supplied interface box.

If I recall, you need a transformer that will carry quite a large DC current, to fool the switchboard's line relay into thinking "the phone is off-hook" so it will keep the line open. There may be a resistance spec too, because a "too low" line voltage implies a fault (shorted line) and monitoring gear may disconnect and flag for repair.

So it is really a special-purpose transformer.

It *might* work to use a modem transformer. These are clearly made to sit on phone lines. The specs look bad, 300-3KHz, but this is often at -1dB so the usable bandwidth is better than most phone gear. I do think you want the biggest core you can find, not one of these micro trannies found on cheap modems. You can buy transformers from DigiKey, I think, and Radio Shack used to stock one that didn't look bad.

The maximum legal level used to be +8dBm, traditionally measured with a VU-like meter (so peak levels might be +18dBm), set to establish tolerable cross-talk in your neighbor's lines. These days another limit will be the digital interface on the other end of your phone line, which will clip at some point.

Since studio standard is now +4dBm, I would use a 600:600 modem transformer, with switch or plug to break the line to "hang up". Drive the primary with any handy output rated to drive 600Ω, and keep the meter "low" like -6VU unless the other end reports that it isn't loud enough. To set-up the call, either they call you and you connect, or you Y a telephone on the same line, call them, and then engage your feed.

 

[CJ]

You could also do an internet feed, but I do not know the specifics.

 

[NewYorkDave]

http://electronicdave.myhosting.net/miscimages/audiototelco.gif

 

[BYacey]

Further to this, Does anyone have a schematic for an electronic hybrid for on-air interviews over the telephone line? I understand a hybrid transformer can be used for this, but I think using op-amps to null the feedback is probably a more elegant way of doing this and is less dependant on line loading and other impedance variables.

 

[PRR]

> an electronic hybrid for on-air interviews over the telephone line? I understand a hybrid transformer can be used for this, but I think using op-amps to null the feedback is probably a more elegant way of doing this and is less dependant on line loading and other impedance variables.

You need a transformer anyway; phone lines are nasty things.

Hybrids were perfected long before opamps appeared.

At these impedances, simply counting the turns gives about 0.1% balance, better than you can get with opamps without precision parts or hand-balancing. Anyway the ultimate balance depends how well your dummy load matches the actual phone line.

All four ports are fully-floating: you can wire them balanced, unbalanced, or even on separate building-grounds if you need to.

If your other gear can work at line level, a hybrid transformer needs no power supply as active parts do.

If you have the money, buy a box. Radio call-in hybrids are highly developed for real-world problems, probably using active parts these days.

But for low-buck DIY, I doubt you can beat or even equal a simple passive hybrid with a bagful of chips.

 

[SmG]

[quote author="BYacey"]Further to this, Does anyone have a schematic for an electronic hybrid for on-air interviews over the telephone line? I understand a hybrid transformer can be used for this, but I think using op-amps to null the feedback is probably a more elegant way of doing this and is less dependant on line loading and other impedance variables.[/quote]
Yes, I have an old (but pretty good) Alice schematic - I'll scan it later on today and post a link to it. If anybody else has any more, I'd be interested to see them too.

 

[jrmintz]

[quote author="NewYorkDave"]http://electronicdave.myhosting.net/miscimages/audiototelco.gif[/quote]

Hi Dave!

:thumb: :thumb:

 

[SmG]

Okay - sorry for delay - the schematic, whilst it has good content, doesn't have good quality - and since I didn't think you'd like the scan very much, I've cleaned it up. It's now a pdf, and you can download it from here. If you want more details, I have one of these sitting on the bench here, waiting to be built into a standalone box. As PRR mentioned, the transformer T1 (listed as DEG477H) is actually marked IR477H KTD1, and is quite substantial. T2 (marked with the RS part number) is actually one of those OEP transformers, with turns ratios marked as 1+1:2+2, so we all know what that is. There's nothing else controversial about this (apart from the complexity of it) - everything is available, although I'm not quite sure where you get Vactrols these days.

 

[SmG]

Another one you can look at is in the schematic for is the D&R Airmix manual on page 50. D&R have thoughtfully made this available here. This schematic is, on the face of it, a little more simple but the way they've done the hybrids is actually more complicated, because they've got a 2-line switcher built into it. The hybrid balancing arrangements are similar, though.

In fact, D&R have got a lot of manuals available on their site - well worth a look there if you are into mixer design. The download page is here.

 

[NewYorkDave]

I've worked with a number of hybrids and I have to admit that, in this case, performance usually IS directly proportional to circuit complexity. The best hybrids I've used are digital hybrids, such as that made by Gentner, until recently the leading US manufacturer of such devices (and now a victim of corporate acquisition... R.I.P).

The chief figure of merit in a hybrid is the degree of isolation between send and receive audio and is termed "trans-hybrid loss." A passive hybrid connected to a phone line cannot achieve high degrees of trans-hybrid loss across the frequency range of interest because the phone line impedance is not flat across the band. A good hybrid transformer with nulling components adjusted carefully can get you into the region of "good enough" on a good day.

For an example of a simple hybrid utilizing a transformer and op-amps, there was a serviceable but not stellar hybrid built into the Gentner Autocoupler.
Manual with schematic and theory of operation (337kB PDF)
At a previous job of mine, we used one of these to interface our television station production intercom to a similar intercom system at a news bureau several miles away via POTS (plain old telephone service). It did an OK job but required careful adjustment at initial setup. For on-air call-in use, though, we used higher-class Gentner hybrids.

The hybrid circuit in the Autocoupler is simple enough, though, to be attempted by the casual home experimenter. I can't say the same for the more sophisticated, higher-performance hybrids out there.

Also, it is instructive to study the two-wire to four-wire interface circuits in analog production intercom systems such as RTS "TW" and Clearcom "Party Line" products.

As PRR says, anything connected to the "nasty" Telco should indeed be transformer-isolated, even if you're doing the bulk of the work of the hybrid with active circuits.

Here's just a couple of examples of commercial hybrid transformers. There are many different models out there from several manufacturers.

http://rocky.digikey.com/WebLib/Parallax%20Pwr%20Comp(Magnetek)/Web%20Data/TY-302P.PDF

http://www.sowter.co.uk/specs/8321.htm

 

[SmG]

Lundahl also make a hybrid transformer - the LL6072. It's interesting to note that nowadays, most of these transformers are physically rather smaller than they used to be. The one I mentioned above marked DEG477H is a pretty substantial item - it has a 3/4" pile of 1.5" x 1.25" laminations, and I have absolutely no idea who made it.

It's also notable that it's not used as a hybrid - the hybrid arrangements in that circuit are purely electronic. It's only acting as an iso transformer. All the circuits we've come up with so far don't use the classic two-transformer system that PRR was describing (the one on the Lundahl data sheet), but rely on some form of electronic cancellation. Presumably this is either better, or cheaper. I've done very little experimenting with the alternatives, so I'm not sure what's actually 'best' - if anything this bad can be described as such - but I've noticed that the newer units that use active digital cancellation seem to work rather better. Now, one of those would be a neat DIY project. (In my dreams...)

 

[PRR]

> All the circuits we've come up with so far don't use the classic two-transformer system that PRR was describing (the one on the Lundahl data sheet), but rely on some form of electronic cancellation.

A hybrid can be one transformer, and the Gentner uses a simple passive hybrid, no electronic cancellation. It does not look like the classic 1-core and 2-core hybrids because they had no need of floating windings on the send and receive ports, and even wind up with the dummy-line (C9 R35) grounded.

Some of those others do have mind-boggling electronic cancelations.

Dave' from-the-trenches input is interesting, and does explain why the quality of radio call-in shows is so much better than it used to be. But I still see these choices:

1) use standard hybrid: cheap and may be good enough

2) devote the rest of your life to hybrid studies and experimentation

3) buy a box, even if it is painful, if you need "best" cancellation.