Tie phone lines

Rewire the transformers as a hybrid

martin

Thanks Martin, I did do a circuit with LTSpice using two hybrids and two amplifiers (to compensate for inductive losses). Problem with that is the receive to transmit isolation (and vice vers) changes quite a bit with frequency. Because of the nature of the transformer hybrids they allow for a bit of feedback (sidetone) which is unwanted for my application. I tried making a hybrid electronically, using two op amps and the existing two transformers, however I encountered similar results to the passive transformer hybrid circuit. I'm basically looking for an easier solution, design time is short. Thanks for your input.

Do the transformers have a split primary and/or secondary?

Do you have a diagram you can post, and provide a link to it? I'm not familiar with the Girator and what they do. What do you mean "they seize the line at the proper impedance?" Depending on how it is connected, it could introduce additional losses. They may also mess up the impedance.

Each cap is only about 15 Ohms... Not much loss there.

I would keep them isolated.

We really need to see schematics.

Did you achieve a balanced condition it the two hybrids? Unless something is incorrect they should achieve, perhaps, 18 dB of balance across the voice-band. What did you use as a balance network? How much gain did you insert in each path?

Both Pri & Sec are center tapped, but not seperated.

The girator circuit allows me to couple the line at a desired dc load without affecting the impeadance, I believe I set that around 20ma. It offers a high impeadance to the line to allow the ac component to pass. Therefore the impeadance is the series cap and transf primary (close to 600?). The transf cannot connect to the line directly being it is of the "dry" type, and would saturate by any line current.

I have never uploaded a schematic to the news group. Do I send it here? I created it in "Eagle" (.sch). Are there file size limitations?

I did not use a passive hybrid because the losses were to great. I created an electronic hybrid, that seems to be the way that all the telephone manufactures are going with. The physical size of two passive hybrids are not an option. The elect hybrid reacts the same, only I can add amplification easily. I have two amps in a spice model now, both at a gain of 3x. The seperation is good with 500HZ into one line, and 1KZ into the second. I can achive around 22 db seperation! I can live with that, but trying other freqs yielded seperation results as bad as 3 db! I am using 300-3000 HZ as my goal.

I've looked at IC's designed for telephones. One option is to use two MC34014's back to back crossing the rec & trans lines. The problem with that is no one uses transformers anymore! All of these chips are designed around a bridge rectifier on the line. If I use that, my isolation goes away! I tried to find out how these conference phones tie the lines, but came up dry. Do they just parallel the lines I wonder! I don't have a conf phone to play with, might be worth for me to get one.

You can put it on alt.binaries.schematics.electronic as a message attachment, or post it on a web page somewhere and post the link here.

I suggest you take a screen-shot of the schematic (.jpg for example) and post it rather than the .sch file. I can open about 200 file types and .sch isn't among them.

Thanks

A single hybrid, as I recall, should have less than 4dB loss in the transmission direction. If you use a "compromise" balance network of 900 Ohms R in series with a 2 Mfd NP cap, you should get good response between

What do you use to balance them?

Likely caused by the balance networks.

I downloaded the datasheet for the MC34014 and need time to study it. It's line interface is....... Strange. Do you encounter any hum or crosstalk problems?

I'm out of touch with modern (after about 1995) sets. In the "olden" days, when you could count on each line being technically identical within a group of lines, it was possible to just tie them together, tip to tip, ring to ring. Can't do that today without isolation.

Don, I posted a partial shot of my schematic. The circuit was to big to put the entire thing into a screen shot. I also posted a copy of the spice hybrid I've been modeling. Thanks...

Scada,

The simple transformer or opamp based "hybrid" circuit (aka transhybrid) is useful for two reasons:

Just adding this hybrid won't buy you anything (for your particular problem). You need to, now, add some gain in the path that you're having trouble with.

Bob

Martin, Do you know of a supplier for the LL6702? I've been unsuccessfull in trying to locate anyone who sells hybrid transformers anymore. Any links would be appreciated.

I found a circuit on line that is equivilant to a telco 22 repeater. This fits my needs exactly! It uses a readily available hybrid transformer, TM-028 available from mouser. Thanks to all for your replys!

Shame on you, you didn't give a link to th cct

martin

Thanks again!

The type 22 will allow you to insert gain in both directions, but what is your intention for passing DC supervisory signals between the POTS and VOIP interfaces?

I didn't find inter-channel signaling conversion capability in the 34014 hardware.

Don

Good question! I thought I would insert an opto isolator (led-to NPN) led in series with the "dummy line" resistor (which I will use a pot - for line imp matching). That way I can monitor line current for the hang up "wink" pulse. That will work on the VOIP end, but another problem is the POTS line. Where I'm using it, they don't use "wink" pulse for hang up. I think I'm going to look for 20 seconds of silence, then hang up.

Anyway, the repeater offers new chalanges. I think I will need two independent bi polar DC supplies to individually power each op amp. Otherwise I will ground out the hybrid! Maybe I can steal power from each line, I'll have to chew on that. It's not the easy solution I was hoping for, but it's the best I have so far.

Steve...

Thinking more about your question, I did not explain myself very well. I do not need to extend the DC supervisory. The only signal I will need to use is the above mentioned "wink" hang up pulse. My device will disconnet each line (see relays in my schem.) I have a PIC chip that does the logic.

snip

Ah, I had forgotten about that one, so obvious

martin

It sounds like you are going to use the Lundahl transformers or an equivalent. It looks good.

If this is a "for sale" product in the US or it will be put in service commercially, you should look at CFR 47, Part 68 or the appropriate ANSI Standard for the Telco interface you are meeting. You may need to Register the equipment.

I'm cruising the internet for info about the VOIP station interface. I'm still in the dark, but did find some high-level insight about what's going on. If you're interested:

Good luck