IP hardphones, analog trunks and a buzzing sound

[mforrence]

I have a new IP406V2 3.0(40) install with a total of ten 5610 & 5620 hardphones and analog trunks (via daughterboards in control unit). When connected to an outside party via an analog trunk, there is a buzzing sound that matches the inside user's speech pattern. Not an echo (well, there's that too, but we can live with it.) but a buzz. The outside caller does not hear the sound. The 5400 series TDM phones don't have the trouble. System is on a "closed" ethernet network (only IP phones & VMPro on the network).

Here's what we've tried:

1. Changed transmit dB on trunks to -4dB. Slight change for the better, but not enough. Changed receive dB on trunks to -4dB. No change.

2. Trunks measured appx 31ma - slightly out of spec. Installed current limiters to bring them to 25ma. Slight change for the better, but not enough.

3. Plugged IP phone directly to 406 lan port (with mid-span power brick). No change.

4. Tier 3 passed this to me: "They had squeal or a loud buzz when they did a trunk to trunk transfer. We were able to help by changing the gain on the trunks to -4db however it did not completely eliminate the buzz. This was a common issue on Legends and Magixs. The RFE added some 185a network terminals to the lines and this got rid of the buzz. If the tech can not find these in his garage a similar part can be ordered from Chicago Capacitor Company. Part# RCP 4405154JN400." I can't find a "Chicago Capacitor Company" or the particular part number...

5. I've read thru past posts on this forum and found several references to the echo - none about a buzz. None of the "echo" posts seemed to have a definitive answer.

Suggestions are most welcome. Thanks
Mike

 

[aeverson]

mforrence,

I had the same exact problem and it drove me crazy. The only difference was that I was using an Analog Trunk Module instead of the 4 port analog daughter card. Avaya has a notice about this and pretty much what I found is if you have a line that is out of voltage specs you will get this buzz. Your lines seem way out of specs what is the voltage on the line? The standard is 48V my lines were about 51V but Avaya wants like 45V (very strange of them) I think. The analog trunk card is much more forgiving and I actually cured my problem by using an analog trunk card. It sounds like you need to get with your local phone company and try to get them to put those lines closer to specs.

If you want to boggle them take a line and run it from one of your analog extensions back into a trunk port. Showm that the trouble does not exist when you do this. Then plug back in the line and show them it sounds bad. Its an Avaya design flaw that shows up on out of spec phone lines but dont let them know that..

Also check and make sure you are using a good ground. I went as far as running a ground from the same point that the phone lines were being grounded from to trouble shoot the issue. Actually I found ungrounding the system actually made it bearable.

Hope this helps

 

[Ragnorak]

There is a lot more to it than just voltage. A phone line is an electrical circuit and Avaya designs its equipment expecting input that falls within certain parameters. If these parameters are not met, there is a problem with the line from your local phone company. Good luck getting them to test it much less fix it. Avaya could have designed their equipment with greater tolerances, but it is still your local loop providers issue. Here are the paramaters-- they differ slightly from state to state and they are usually specified in the state's PUC or PSC tariffs.

Loop Current mA
Acceptable > 23
Marginal > 20 < 23
Unacceptable < 20

Circuit Loss dBrn
Acceptable < -8.5
Marginal > -8.5 < 10
Unacceptable > 10

Metallic Noise dBrnc
Acceptable < 20
Marginal > 20 < 30
Unacceptable > 30

Power Influence dBrnc
Acceptable < 80
Marginal > 80 < 90
Unacceptable > 90

Balance dB
Acceptable > 60
Marginal < 60 > 50
Unacceptable < 50

LOOP CURRENT
Current on the line coming from the Central Office

CIRCUIT LOSS
Measurement of the power loss of a 1004hz signal measured in dBrn.

NOISE METALLIC
Noise measured across the tip and ring of a circuit that the customer actually hears on the line.

POWER INFLUENCE
Noise measured between ground and the tip and ring conductors shorted together. The subscriber does not hear it, but it affects the amount of noise metallic the subscriber does hear.

BALANCE
The amount of current cancellation along the tip and ring conductors. The subscriber hears non-cancelled current as noise metallic. BALANCE (dB)=POWER INFLUENCE minus CIRCUIT NOISE

Typical Problems

Loop Current and Circuit Loss work together. As the loop current approaches 23 mA the circuit loss should be about -8.0dBm.
Loop Current low-Loss OK
This is generally a defective loop aid, such as the Office Equipment card in the CO or the card in a pair gain device.
Loop Current OK-Too much loss
Indicates incorrect loading or a bridged tap.
Both are unacceptable
This indicates a problem with the resistance levels in the loop make-up. It probably exceeds the maximum 1300 Ohms.

Bad Noise measurements indicate a defective cable pair. This is probably a loose connection or a broken or corroded wire.

Unacceptable Power Influence indicates an open cable shield or missing shield ground.

A Balance problem indicates a one-sided open in either a bridged tap or in the cable pair past the point of service.

 

[aeverson]

True and if you do get them to come out all they are going to do is plug their butt set in (which is very tolerant of noisy lines) and tell you it works and its now his lunch break. I dealt with two different companies in two different states about this issue and everytime they blamed it on Avaya and Avaya blamed it on them. Leaving me in the middle scratching my head and looking confused to my customer. Unfortunately to resolve the issue my customers ended up just getting a T1. Very expensive way to solve a problem.

 

[mforrence]

After extensive testing, Tier 3 passed to Tier 4 - who replied that they have had other sites with similar trouble.

"It is thought at this time that the issue appears to be an interaction between the 5600 series and the analog module"

I'll update with the final outcome...
MIKE

 

[JeSTeROCK]

Would a resistor in series with the line help ? or is it all analogue lines?

 

[mforrence]

I've tested with as much as 800 ohms in series (on tip and ring) on the CO lines - made some difference, but not enough. It also took the line current below the recommended threshold of 20ma. Changing the Transmit level on the line (lines section of manager) to -4db also made a noticable difference... but not enough.
MIKE