Standard Copper POTS vs. T1 delivered POTS

[Brian Bowers]

A couple of months ago I asked about ideas for constructing a variable regulated DC power supply. My problem (or so I thought at the time) was interaction between some consumer Caller ID units. Here was the scenario - 3 studio listener POTS lines were delivered via copper, the 4th in the hunt group was delivered via T1 POTS. I was also trying to power the consumer Caller ID units using a plug in wall wart type power supply. The Caller ID units usually ran each with 3 AAA batteries for a power of 4.5 volts. The Caller ID units wanted as close to this exact voltage, otherwise anything more would light all of the LCD screen characters.

After discovering that it would cost me close to $40 for all of the parts to construct a halfway decent safe regulated power supply, excluding my time to construct it, and order and obtain the parts, I (or the company) splurged for an $80 two amp regulated power supply (with metering!) from our local electronics retailer. With this power supply, all Caller ID units worked perfectly...with the first 3 copper POTS lines connected. The units fail to work, or lock up, when the 4th T1 POTS line is connected. I was like what? How can this be. But it be alright. I swapped the T1 POTS in the building with a copper pair POTS, and had TDS rework the hunt. All Caller ID units work just fine with this hunt group of all copper POTS.

Not really sure who was to blame here, the Fanstel Caller ID units, or the mixture of both the copper POTS and T1 POTS, or all of the above, along with the external power supply. When using the combination of copper POTS and T1 POTS, and batteries in all of the Caller ID units, everything worked just fine.

 

[littlejohn]

Well the dfference is, the copper gets batery and supervision from the swotch, while the T1 gets it from the T1 CSU, whether yours or the Telco's. That shouldn't matter. Maybe the locally sourced talk battery isn'tclean, or oisn't 48 volts or something. Or it's got hash on the line from the CSU. Interesting circumstance, sounds as if you've worked around it at any rate.

 

[RolfTaylor]

Swap tip and ring on the T1 derived line.

Hard to explain without graphics, but one side of a pots line is (partially) grounded and my guess is that is interacting with the external power supply.

You might try lifting the safety ground on the external power supply, though I do not recommend that as a permanent scenario.

Keep us posted.

 

[Brian Bowers]

To me the easier solution was to get rid of the T1 POTS in the hunt, and replace it with a copper POTS, once I figured out that that would solve my problem. That did solve the problem, but it took a lot of experimenting to figure out what was going. Our Telecom TDS wasn't able to help in the problem solving either. They asked me to get back to them regarding what I found, so they could pass it on to their team. Truthfully, I rarely pay any attention to Tip and Ring on POTS lines. Looking at two wires, I don't even know how to decipher one from the other. How do you do that, and when does it become critical?

 

[bilco]

Not the answer you are lookin for but we do not use caller ID on listener lines that are used with contests, lest we be accused of fraud. A caller might be a pain in the ass and the Jock may just skip that number when they see it. Hasn't happened, and won't since we don't use caller id on those lines, but in this litigous society it's a toreal possibility. With no caller ID, they can't accuse us of that.

 

[RolfTaylor]

Red=Ring=Negative

Remember that Telco -48 to ground.

Polarity can matter in a number of cases. It is almost always worth trying to swap them!

It used to be phones were balanced and it did not matter. Then, DTMF touch tone was added it mattered, only in terms of whether the tone generator functioned (apparently the lack of a rectifier bridge was deliberate as they did not want to lose the voltage drop of two diodes, so that the sets would work on very long loops).

These days the "termination" is separate for the audio and the DC current. In the older designs the transformer had to be able to deal with as much as 100 mA of loop current without saturating. Nowadays there is a circuit that sinks the line current separately and the audio is still balanced with a transformer (a much smaller one since it does not have to deal with DC. Smaller transformer saves space and money.

In the case of your external power supply, it probably has one side referenced to ground. Since the CID box was meant to be battery powered there was no attempt to isolate the "signal return/ground" from the battery and thus you wound up inadvertently grounding it when you connected the power supply..

The most common case where it matters is where you are using a single hybrid to "button mash" together multiple lines. If tip and ring are not the same on both lines, you are effectively shorting them (albeit a "short" over rather long conductors back to the CO). If tip and ring are correct on the lines, this approach, where the lines are simply tied together, works just fine.

However, button mashing lines that come from different sources (your case being a common example) means that they won't "mash" properly no matter the polarity of the lines.

Sources of "difference" can range from different CO, CO line with PBX port, or either of the previous to a derived line (e.g. off of a T1 or other carrier).

No back to your installation. Your power supply is still partially grounding one side of the lines. This means the common mode rejection ratio of the lines has been decreased. If you experience noise on these lines, it will probably be due to the caller ID boxes/power supply combo and will go away if you remove the power supply.

If that is the case, you could go back to a floating power supply by using a separate D battery set for each unit. Presumably D batteries would last long enough not to be too much of a maintenance problem (change when you reset the clocks).

Cheers

Rolf

PS- the telcos no longer care about Tip and Ring, but that doesn't mean they shouldn't!!

To me the easier solution was to get rid of the T1 POTS in the hunt, and replace it with a copper POTS, once I figured out that that would solve my problem. That did solve the problem, but it took a lot of experimenting to figure out what was going. Our Telecom TDS wasn't able to help in the problem solving either. They asked me to get back to them regarding what I found, so they could pass it on to their team. Truthfully, I rarely pay any attention to Tip and Ring on POTS lines. Looking at two wires, I don't even know how to decipher one from the other. How do you do that, and when does it become critical?

 

[littlejohn]

Again, I has learnt something. Thanks,Rolf.

 

[RolfTaylor]

Again, I has learnt something. Thanks,Rolf.

Aw gee, John, thanks for the thanks! I guess if I can teach you something new I should continue!!

Illustrated below, in its ascii-art splendor is the line transformer at a central office:

Balanced audio (from a DAC in modern COs) is input across primary A and B.

The secondary is not center tapped; rather, there are two secondary windings end-to-end. Thus the audio portion is still balanced despite the fact that the positive side of the power supply is grounded. Yes that is correct, it is a minus 48 VDC supply NOT a +48 VDC supply.

A
0
( | | ( - TIP
) | | )
( | | (
) | | )
( | | 0 - Ground
) | | Central Office Line Transformer
( | | 0 - -48VDC
) | | )
( | | (
) | | )
( | | ( -RING
0
B

So, in the case of two lines from the same CO are tied together (mashed) all is well if TIP is to TIP and RING is to RING. If you tied them together TIP to RING and RING to TIP, you have created (as I mentioned before) a "short circuit" though fortunately over a very long pair of wires.

However, if the lines are from "different" sources, there is a ground loop when you do this.

My hypothesis is that the design of the caller ID boxes assume a floating power supply (e.g. battery) and therefore one side of the line is connected (indirectly no doubt) to the battery. As soon as you connect the power supply leads together from multiple caller ID boxes you are going to have problems, even if the supply itself is floating. If it is not, then you have two problems.

Now, one might ask why the same battery supply at the CO can be fed to multiple lines without the audio being coupled as well. The answer to that is a choke in series with the -48 VDC, one for each line.

Regarding the DC termination method I mentioned before. This technique is has greatly reduced the size and cost of CPE line interfaces. Compare the line card on a Telos TWOx12 versus an Nx12 to see the difference.

 

[Kmagrill]

Interesting thread.

The telco line side of the caller ID box should be transformer coupled so that the polarity of the power supply cannot affect the line regardless of the phone line's polarity. If the box is FCC part 68 registered, that certainly has to be the case. If it is not FCC registered and is not compliant with the part 68 rules, then it's a safety violation because there's a path from the phone line to the AC power via the external supply. This should never happen because it can screw up the phone line and in event of a malfunction on your end, is potentially hazardous to people working on the phone lines. For example, imagine that you accidently connect 240Vac to the caller ID box power input. Aside from blowing up the box, there should never be any way for that to leak onto the actual phone line.

 

[littlejohn]

There are many aftermarket instruments out there which have no transformers in them. No idea if they're Part 68 compliant, but big retailers sell them daily.
All the copper coming into my facilities have protection just before the demarc, even the stuff which originates at the SLIC in the basement. Or the light pipe frame.

 

[Kmagrill]

There are many aftermarket instruments out there which have no transformers in them. No idea if they're Part 68 compliant, but big retailers sell them daily.....

Well, I wasn't nearly as clear as I should have been. There are other ways to isolate the telco line side from the user interface side of a product besides just transformers. The main requirements, among many, are that the product provide at least 1,500V of isoaltion between the user and the phone line. Also the device cannot unbalance the phone line. It's possible that one of these other techniques was used, but if the product is capable of unbalancing a line or shorting it out (other than by direct current coupling of two phone lines together), then there's likely to be a problem.

 

[Brian Bowers]

Just so you know, the Caller ID boxes I have been referring to are Fanstel model B72S. They say they comply with standards in Part 68 of the FCC Rules, and the product is registered with the FCC.

 

[Jasonce66]

These were FCC compliant caller ID boxes before they were modified. The easy way to get rid of the problem is to power each box from its own supply. For this I would have used a cheap radio shack 6 volt wall wart, put a 1k ohm resistor in series with the positive output and used a 4.5 volt zener diode to get the exact voltage. At this point, if you had wall warts from a discount shelf that were 12v, you could do the same thing, just adjust the resistor value accordingly, so the zener doesn't heat up too much - the ones from Radio Shack are so cheap anymore... A 6 volt wall wart, unloaded will yield about 10 or so volts, a 12 volt one will make about 16 to 20 volts, so a 2.2k resistor might work fine. Maybe some from All Electronics or Marlin P. Jones? Any amount of current will run the units, as they typically run for months on a set of AAA batteries, so if you read 4.5 on the zener, you are good. The wall wart will provide you with ground isolation, and if you use separate supplies for each id unit, you won't get any interaction from them.
Your bench supply will be nice for testing something else - I usually have to bring one from home to work to test something, so to have one on the bench is nice. For a wall wart, you want at least 600V isolation from both sides of the DC lines feeding the caller ID unit to maintain FCC compliance (maybe more?)... the bench supply will probably connect the negative post to AC ground, which is not good for phone use. The wall warts are perfect for this as they typically have 600 to 1200V hi-pot capability to be UL listed. Connecting all of the power supply lines for the units together will certainly make them interact with each other on the phone lines, and could cause ring voltage from one line to "leak" over to another.

Another option is to buy a caller ID that is phone line powered...

http://www.dealextreme.com/p/fsk-dtmf-phone-line-powered-caller-id-display-with-call-history-12729


Polarity issues are commonplace in broadcasting. Most hum issues on multi-line hybrid switches, like a Telos, are caused by a reversed line.

Good luck.

 

[SirRoxalot]

Since tip/ring IS an issue with a lot of telecom equipment, I thought you might be interested in a chart for the colors of standard telecom cable:

http://www.shout.net/~wildixon/telecom/cable/color.html

Notice that the chart also has the tip-ring code for the old JK (red-green-black-yellow) cable. These are the standard colors. Adjustments may need to be made for old engineers who created job security by changing the color code at every termination.

 

[RolfTaylor]

Since tip/ring IS an issue with a lot of telecom equipment, I thought you might be interested in a chart for the colors of standard telecom cable:

http://www.shout.net/~wildixon/telecom/cable/color.html

Notice that the chart also has the tip-ring code for the old JK (red-green-black-yellow) cable. These are the standard colors. Adjustments may need to be made for old engineers who created job security by changing the color code at every termination.

Good post, but since the outside plant crews in many of the Telcos no longer pay attention to TIP vs RING, if in doubt check it with a meter and remember "red/ring/negative"

 

[Brian Bowers]

Since I needed to install another 4 Caller ID box setup in another studio, like the one I first mentioned in the thread, I kept a number of these suggestions in mind. What I did discover was that all phone line input polarity could be switched (uniformly) in my Telos 1x6 phone inputs (without the Caller ID boxes installed), and all lines would conference without any problems. Switch the polarity on one line, and it would create a loud buzz when conferenced with the other lines. But it ended up that only combination of phone line polarity into the Caller ID boxes (while externally powered by the one source adjustable power supply), and one combination of phone line polarity into the Telos 1x6 worked 100% correct. With this setup, what is the correct way to isolate the power supply? I am printing this thread and saving it for future reference, as I doubt listener phone lines in radio stations probably won't be going away anytime soon. Thanks.

 

[RolfTaylor]

Brian,

thanks for posting your results back to this thread. An opportunity for all of us to learn.

Here are a couple of things to try (and do keep us posted)

1) Try lifting the ground at the power cord of the power supply.

2) If the power supply also has both + and - outputs feed the caller ID from those outputs without a reference to ground (set one for -2.25 volts and the other for +2.25 and feed the boxes across those outputs) If the problem is as I surmise this should take care of it.

3) Buy some of the newer switching regulated 5 volts "wall warts" and a power strip and keep them isolated.

Good luck, keep us posted.

Since I needed to install another 4 Caller ID box setup in another studio, like the one I first mentioned in the thread, I kept a number of these suggestions in mind. What I did discover was that all phone line input polarity could be switched (uniformly) in my Telos 1x6 phone inputs (without the Caller ID boxes installed), and all lines would conference without any problems. Switch the polarity on one line, and it would create a loud buzz when conferenced with the other lines. But it ended up that only combination of phone line polarity into the Caller ID boxes (while externally powered by the one source adjustable power supply), and one combination of phone line polarity into the Telos 1x6 worked 100% correct. With this setup, what is the correct way to isolate the power supply? I am printing this thread and saving it for future reference, as I doubt listener phone lines in radio stations probably won't be going away anytime soon. Thanks.

 

[Brian Bowers]

Thanks Rolf. I can tell you that both of these systems work perfectly. The power supply is not bipolar, so all I really have is a positive and negative available to me. Since I've already shelled out $80 for each of these power supplies, I doubt I'll now go the route of the individual wall warts for each Caller ID unit. The power supply is fused, so hopefully that will provide some protection if the situation ever requires it. It would be simple for me to lift or isolate the ground pin on the AC power cable for the power supply, but I hesitate to do that. If anything, this has taught me that in unique phone line applications, polarity may be an issue.

 

[ChiefOperator]

Since tip/ring IS an issue with a lot of telecom equipment, I thought you might be interested in a chart for the colors of standard telecom cable:

http://www.shout.net/~wildixon/telecom/cable/color.html

Notice that the chart also has the tip-ring code for the old JK (red-green-black-yellow) cable. These are the standard colors. Adjustments may need to be made for old engineers who created job security by changing the color code at every termination.

Good post! I had the opportunity to work with an old Bell techinician who taught me an easy method to memorize the entire 25-pair sequence. In fact, when I punch down a 25-pair cable, I do so from memory based on a couple Bell telco sayings.

But first, the easy way to remember the Green, Red, Black, Yellow quad cable
"Christmas Tree" (Green, Red), "Bumble Bee" (Black, Yellow).

Now, to the 25-pair sequence. First, note that the sequence is nothing more than 5 "binder" colors combined with 5 "dominant" colors. The binder colors are the "tip" side of the circuit and the "dominant" colors are the "ring" side.

First, memorize the following saying for the binders (tip side of circuit):

Why Run Backwards You'll Vomit
Colors: White, Red, Black, Yellow, Violet

Now, memorize the following saying for the dominant colors (ring side of circuit)

Bell Operators Give Best Service
Colors: Blue, Orange, Green, Brown, Slate

That's all you need to know to punch a 25 pair cable. I'll show the first 10 pairs to illustrate:

We first punch the white binder ("Why") and the 5 dominant colors:

Pair 1: White/Blue (Why/Bell)
Pair 2: White/Orange (Why/Operators)
Pair 3: White/Green (Why/Give)
Pair 4: White/Brown (Why/Best)
Pair 5: White/Slate (Why/Service)

Next, we'll move to the Red Binder ("Run") and the 5 dominant colors:

Pair 6: Red/Blue (Run/Bell)
Pair 7: Red/Orange (Run/Operators)
Pair 8: Red/Green (Run/Give)
Pair 9: Red/Brown (Run/Best)
Pair 10: Red/Slate (Run/Service)

And so on to pair 25.

These sayings have worked well for me.. I hope they do for you as well.

 

[Brian Bowers]

I've just been told by one of our AM (250 watts ERP non-DA) hosts, that during his talk radio show, he was punk'd by Sirius Radio the other day. Sirius Radio actually came up on the Caller ID units. Possibly the Stern show? Having Sirius myself, and being an occasional listener to the Stern show, I know that these guys regularly do this. Even for that one situation, it was worth every penny and minute installing the studio Caller ID system to confirm this actually happened. Our host was so quick to realize what was going on, he didn't even need the caller ID info to confirm it, but he informed me he was once again appreciative of the system. The FM guys were so jealous I favored the AM over the FM, in installing a Caller ID system, they requested a system too. Surprise surprise.