[jknappsax]

I work at an FM station that has several call-in shows. We have three
Telos Ones connected to a switchpanel for on-air callers. Over the
last 6 months or so. we've been having problems with all three hybrids
where get a "slinky" sound as the hybrid tries to autonull the line.
We also get into a mode where the host goes into a kind of feedback
intermittently. Each hybrid is being fed a mix-minus, with the common
input including our host. I have checked and listened to the mix-
minuses separately, and they have what they should have, with no
anomalies at the time of test. We had one unit back to Telos to test,
and it came back fine, with better than -52 dB trans-hybrid
bleedthrough. I initially thought that the problem only evidenced
itself when all three hybrids were on air, and figured with the
leakage through 3 hybrids took the level of the host bleeding through
from -52 dB down to about -24 dB as each line added a bit of the host
back into the mix. However, I've been seeing similar issues with each
single hybrid from time to time. We've also swapped known good units
into the one spot where we've had problems, and it had the same
result, leading me to believe that it isn't a hybrid problem at all,
but a line problem, as we've had a very rainy fall and snowy winter,
and I think that some of the lines between the CO and the station are
compromised. The host if going nuts, and we've also tried changing out
a filter chip (a D2912A, U14) that Telos told us sometimes caused this
kind of problem, and it was even worse. I have a firmware upgrade that
supposedly handles bad lines (the "Dallas" software, or version 4X)
that I intend to install to see if it helps. Anyone have any
experience with this type of problem, or have any suggestions for a
direction to look for this. Thanks.

John Knapp
WHYY Ch 12/ 90.9 FM

[Scott Dorsey]

In article ,
jknappsax wrote:
I work at an FM station that has several call-in shows. We have three
Telos Ones connected to a switchpanel for on-air callers. Over the
last 6 months or so. we've been having problems with all three hybrids
where get a "slinky" sound as the hybrid tries to autonull the line.

Can you disable the nulling and just make it set-and forget? In the old
world of mechanical exchanges, the line balancing would change with every
call, because with every call you'd be going through a different sequence
of cables. In the modern era, you're taking the same path to the card in
the exchange with every call. So the nulling won't change from call to call.

The null WILL change as the weather changes and the cabling condition
changes, but that's something you can set periodically by hand.

back into the mix. However, I've been seeing similar issues with each
single hybrid from time to time. We've also swapped known good units
into the one spot where we've had problems, and it had the same
result, leading me to believe that it isn't a hybrid problem at all,
but a line problem, as we've had a very rainy fall and snowy winter,
and I think that some of the lines between the CO and the station are
compromised.

This happens all the time in former GTEland, and there's not much you can do to
convince the telco to fix things. I think, though, that you are experiencing
a problem with the autonulling algorithm going berserk when it is trying to
deal with your bad line. If you can't fix the line, you might be able to
fix the nulling.

The host if going nuts, and we've also tried changing out
a filter chip (a D2912A, U14) that Telos told us sometimes caused this
kind of problem, and it was even worse. I have a firmware upgrade that
supposedly handles bad lines (the "Dallas" software, or version 4X)
that I intend to install to see if it helps. Anyone have any
experience with this type of problem, or have any suggestions for a
direction to look for this. Thanks.

See if there is some way you can disable autonulling completely, except
when you specifically ask it to set the null. You shouldn't have to be
riding the null all day long like you used to have to do in the old days.
Or you can always dig up that old Gately out of the closet....
--scott
--
"C'est un Nagra. C'est suisse, et tres, tres precis."

[jknappsax]

On Mar 11, 11:13*am, (Scott Dorsey) wrote:
In article ,

jknappsax wrote:
I work at an FM station that has several call-in shows. We have three
Telos Ones connected to a switchpanel for on-air callers. *Over the
last 6 months or so. we've been having problems with all three hybrids
where get a "slinky" sound as the hybrid tries to autonull the line.

Can you disable the nulling and just make it set-and forget? *In the old
world of mechanical exchanges, the line balancing would change with every
call, because with every call you'd be going through a different sequence
of cables. *In the modern era, you're taking the same path to the card in
the exchange with every call. *So the nulling won't change from call to call.

The null WILL change as the weather changes and the cabling condition
changes, but that's something you can set periodically by hand.

back into the mix. However, I've been seeing similar issues with each
single hybrid from time to time. We've also swapped known good units
into the one spot where we've had problems, and it had the same
result, leading me to believe that it isn't a hybrid problem at all,
but a line problem, as we've had a very rainy fall and snowy winter,
and I think that some of the lines between the CO and the station are
compromised.

This happens all the time in former GTEland, and there's not much you can do to
convince the telco to fix things. *I think, though, that you are experiencing
a problem with the autonulling algorithm going berserk when it is trying to
deal with your bad line. *If you can't fix the line, you might be able to
fix the nulling.

The host if going nuts, and we've also tried changing out
a filter chip (a D2912A, U14) that Telos told us sometimes caused this
kind of problem, and it was even worse. I have a firmware upgrade that
supposedly handles bad lines (the "Dallas" software, or version 4X)
that I intend to install to see if it helps. Anyone have any
experience with this type of problem, or have any suggestions for a
direction to look for this. Thanks.

See if there is some way you can disable autonulling completely, except
when you specifically ask it to set the null. *You shouldn't have to be
riding the null all day long like you used to have to do in the old days.
Or you can always dig up that old Gately out of the closet....
--scott
--
"C'est un Nagra. *C'est suisse, et tres, tres precis."

There's no way to disable the autonulling; the process sends a short
burst of white noise down the line, and the micro develops an profile
to create a signal that is an inverse of the leakage, and then applies
it to the line. The firmware update I did is one for bad lines. By
tomorrow we should know if this "more aggressive"software helps. The
one thing I think the operators could do (besides pushing the line in
use up in the mix and ducking the other lines about 6dB) is to use the
override function. It's not variable like an analog hybrid; it's
either off or on. there are essentially no tweaks in the box, as it's
all handled by the micro and DSP. We're having Verizon in today to
analyze the lines, but I think that, as we both suspect, it's circuit
by circuit. In my TV control room, I have an old analog Symetrix 108,
and it never gives me the kind of problems I see here.Since I went
through a coarse and fine null about 10 years ago, it's always been
pretty consistent with any of the POTS lines that come up to it. I
always crossride the hybrid and the host, just like I always crossride
multiple lavaliers to minimize phase cancellation. The radio engineers
usually just bring faders up, and never touch them (quiet room with
good mics forgive some of this lack of active mixing). Scott, I
appreciate your response, and am thinking that you and I agree that
some really crappy lines might be a big part of the problem.

[Mark]

On Mar 11, 11:53*am, jknappsax wrote:
On Mar 11, 11:13*am, (Scott Dorsey) wrote:





In article ,

jknappsax wrote:
I work at an FM station that has several call-in shows. We have three
Telos Ones connected to a switchpanel for on-air callers. *Over the
last 6 months or so. we've been having problems with all three hybrids
where get a "slinky" sound as the hybrid tries to autonull the line.

Can you disable the nulling and just make it set-and forget? *In the old
world of mechanical exchanges, the line balancing would change with every
call, because with every call you'd be going through a different sequence
of cables. *In the modern era, you're taking the same path to the card in
the exchange with every call. *So the nulling won't change from call to call.

The null WILL change as the weather changes and the cabling condition
changes, but that's something you can set periodically by hand.

back into the mix. However, I've been seeing similar issues with each
single hybrid from time to time. We've also swapped known good units
into the one spot where we've had problems, and it had the same
result, leading me to believe that it isn't a hybrid problem at all,
but a line problem, as we've had a very rainy fall and snowy winter,
and I think that some of the lines between the CO and the station are
compromised.

This happens all the time in former GTEland, and there's not much you can do to
convince the telco to fix things. *I think, though, that you are experiencing
a problem with the autonulling algorithm going berserk when it is trying to
deal with your bad line. *If you can't fix the line, you might be able to
fix the nulling.

The host if going nuts, and we've also tried changing out
a filter chip (a D2912A, U14) that Telos told us sometimes caused this
kind of problem, and it was even worse. I have a firmware upgrade that
supposedly handles bad lines (the "Dallas" software, or version 4X)
that I intend to install to see if it helps. Anyone have any
experience with this type of problem, or have any suggestions for a
direction to look for this. Thanks.

See if there is some way you can disable autonulling completely, except
when you specifically ask it to set the null. *You shouldn't have to be
riding the null all day long like you used to have to do in the old days.
Or you can always dig up that old Gately out of the closet....
--scott
--
"C'est un Nagra. *C'est suisse, et tres, tres precis."

There's no way to disable the autonulling; the process sends a short
burst of white noise down the line, and the micro develops an profile
to create a signal that is an inverse of the leakage, and then applies
it to the line. The firmware update I did is one for bad lines. By
tomorrow we should know if this "more aggressive"software helps. The
one thing I think the operators could do (besides pushing the line in
use up in the mix and ducking the other lines about 6dB) is to use the
override function. It's not variable like an analog hybrid; it's
either off or on. there are essentially no tweaks in the box, as it's
all handled by the micro and DSP. We're having Verizon in today to
analyze the lines, but I think that, as we both suspect, it's circuit
by circuit. In my TV control room, I have an old analog Symetrix 108,
and it never gives me the kind of problems I see here.Since I went
through a coarse and fine null about 10 years ago, it's always been
pretty consistent with any of the POTS lines that come up to it. I
always crossride the hybrid and the host, just like I always crossride
multiple lavaliers to minimize phase cancellation. The radio engineers
usually just bring faders up, and never touch them (quiet room with
good mics forgive some of this lack of active mixing). Scott, I
appreciate your response, and am thinking that you and I agree that
some really crappy lines might be a big part of the problem.- Hide quoted text -

- Show quoted text -

when you say you are using 3 hybrids at the same time... do you
mean you have 3 phone guests on 3 different phone lines talking to
each other and to the local host?

and each hybrid gets a local mix minus of all the other audio except
it's own output? i.e each hybrid gets fed the other two hybrids and
the host?

is that when you get the problem?

do you get the problem also with only 1 phone line/guest and the host?

are there any dsl filters or dsl modems on these lines?

does it matter if the other end is a cell phone or land line?

I'm thinking as a test I would add some "build out resistance" i.e.
100 Ohms or so in series with each line to stabilize the Z and see
what happens....

or otherwise lower the "send" gain a little bit

and lastley, are you sure that all the remote guests have their radios
turned down :-)

Mark
(in the Philly area by the way and I listen to WHYY)