Channel banks and modem connection speeds.
Moe & Melissa
I have a customer with a T-1 connected to a Telco DCB-24 channel bank.
The T-1 is set up ESF/B8ZS. This T-1 carries 11 inbound 800 voice lines,
512K internet. Recently we added a 888 number to ring into a channel for
the customer to access his internet connection. This channel rings into
a Hayes Accura 33.6 external modem. So far we have only been able to
connect at speeds ranging from 19.2k to 28.8k. Most connections are
26.4k and less. We have swapped modems of differing brands and
capabilities but the connection rates have remained on the low side.
In order to clear the 888 routing part of the network, I had the 888
number ported to a POTS line. We were able to achieve 31.2k on a regular
basis and sometimes 33.6k.
The channel bank is set up correctly as far as timing, framing and
coding. The performance on the 512K internet has been great.
The T-1 and this DS-0 run error free during BERT test.
I have tried adjusting the channel card rcv/xmt levels up and down to no
avail. Telco had no other suggestions beyond adjusting the levels.
Is it usually hard to get decent connect speeds through a channel bank?
What other tests could be suggested to try to improve the connect speeds
we are seeing?
TIA,
Moe Kunkle
Satch
The best test would be to get IEEE-743-1995 compliant Transmission
Impairment Measuring Sets (TIMS) such as the Sage 930A with the 23-tone
test option installed, and use that to scope out your digital channel to
see what's going on. I suspect that the line cards in your channel banks
roll off too soon, and that coupled with any low-frequency noise would
narrow the available bandwidth significantly so that the higher rates are
impossible.
The TIMS can also sniff out IMD problems.
You also need to worry about crosstalk with channel banks, particularly if
the battery on the thing is small. I once worked with a channel bank that
had no battery at all and the crosstalk between channels was bad enough to
affect the ability to speak over the channels in the thing, let alone modem
transmission.
Satch
--
Stephen Satchell, Satchell Evaluations
(Formerly of Motorola ISG)
Opinions expressed here are my PERSONAL opinions
To reply via e-mail, remove the last domain part.
Hatredonalog
In article <3528CF...@ro.com>
Moe & Melissa <m...@ro.com> wrote:
> I have a customer with a T-1 connected to a Telco DCB-24 channel bank.
> The T-1 is set up ESF/B8ZS. This T-1 carries 11 inbound 800 voice lines,
> 512K internet. Recently we added a 888 number to ring into a channel for
> the customer to access his internet connection. This channel rings into
> a Hayes Accura 33.6 external modem. So far we have only been able to
> connect at speeds ranging from 19.2k to 28.8k. Most connections are
> 26.4k and less. We have swapped modems of differing brands and
> capabilities but the connection rates have remained on the low side.
>
What is the dialup #?
Just a Thought...
-Hatredonalog
-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
..o0 digital phreak p1mps 0o..
http://internet.monks.net/p1mp
Currently down due to 9x's retardation!
use: http://members.tripod.com/~p1mp
digital phreak p1mps e-zine
Where do y0u want to p1mp today?
-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
"scream at me again, if you like.
Throw your hate at me, with all your might.
Hit me cause I'm strange, hit me.
Tell me I'm a pussy and you're harder than me. "
-KoR, "Clown"
-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
Rick Ellis
In article <35292b0...@news.mindspring.com>,
S.C. <roamer1@-removeme-pobox.com> wrote:
>>Is it usually hard to get decent connect speeds through a channel bank?
>If it's a line-side T1, yes.
How about trunk side?
Stanley Cline
It depends. Sometimes trunk-side T1's give good connections through
channel banks, sometimes they don't -- it seems very dependent upon
the channel bank and framing (D4/AMI, ESF/B8ZS, etc.) used on the T1.
-SC
--
Stanley Cline (Roamer1 on IRC) ** GO BRAVES! GO VOLS!
Network Operations Center Engineer, MindSpring, Atlanta, GA.
scline(at)mindspring.net -- http://scline.home.mindspring.com/
SPAMINATE! help outlaw spam - http://www.cauce.org/
Rick Ellis
In article <Pine.LNX.3.96.98040...@mulder.mindspring.com>,
Stanley Cline <scl...@mindspring.net> wrote:
>It depends. Sometimes trunk-side T1's give good connections through
>channel banks, sometimes they don't -- it seems very dependent upon
>the channel bank and framing (D4/AMI, ESF/B8ZS, etc.) used on the T1.
Does AMI make much of a difference?
Floyd Davidson
Stanley Cline <scl...@mindspring.net> wrote:
>Rick Ellis wrote:
>> S.C. <roamer1@-removeme-pobox.com> wrote:
>>
>> >>Is it usually hard to get decent connect speeds through a channel bank?
>>
>> >If it's a line-side T1, yes.
>>
>> How about trunk side?
>
>It depends. Sometimes trunk-side T1's give good connections through
>channel banks, sometimes they don't -- it seems very dependent upon
>the channel bank and framing (D4/AMI, ESF/B8ZS, etc.) used on the T1.
Framing shouldn't have any affect at all. Neither would line coding.
However, the original description included (if memory serves, which
it doesn't always anymore) trunks, an agraggate data circuit, and
this one line all in the same T1. At the distant end there is either
a digital cross connect system or another channel bank... and more
than likely both! (For the one line at least.)
Given the connect speeds described, his results match a channel bank
on both ends with an analog interface to the switching system.
Floyd
--
Floyd L. Davidson fl...@polarnet.com
Ukpeagvik (Barrow, Alaska) or: fl...@barrow.com
Moe & Melissa
Floyd Davidson wrote:
>
> Framing shouldn't have any affect at all. Neither would line coding.
>
> However, the original description included (if memory serves, which
> it doesn't always anymore) trunks, an agraggate data circuit, and
> this one line all in the same T1. At the distant end there is either
> a digital cross connect system or another channel bank... and more
> than likely both! (For the one line at least.)
There is a channel-bank at the customer premise only. The other end for
this 888 line would be where ever the customer happens to be when he
dials it. The T1 terminates in the carriers switch and digital
cross-connect system.
>
> Given the connect speeds described, his results match a channel bank
> on both ends with an analog interface to the switching system.
The only analog interface is the channel-bank at the CPE. Of course, a
lot depends on the CO where the 888 call originates from.
How about a small ISP? What did these guys do before the digital modem
banks that direct connect to the T1 existed? Didn't they have to use a
channel-bank connected to a seperate modems?
Moe Kunkle
Floyd Davidson
[posted and emailed]
Moe & Melissa <m...@ro.com> wrote:
>Floyd Davidson wrote:
>> However, the original description included (if memory serves, which
>> it doesn't always anymore) trunks, an agraggate data circuit, and
>> this one line all in the same T1. At the distant end there is either
>> a digital cross connect system or another channel bank... and more
>> than likely both! (For the one line at least.)
>
>There is a channel-bank at the customer premise only. The other end for
>this 888 line would be where ever the customer happens to be when he
>dials it. The T1 terminates in the carriers switch and digital
>cross-connect system.
>>
>> Given the connect speeds described, his results match a channel bank
>> on both ends with an analog interface to the switching system.
>
>The only analog interface is the channel-bank at the CPE. Of course, a
>lot depends on the CO where the 888 call originates from.
Hmmm...
>How about a small ISP? What did these guys do before the digital modem
>banks that direct connect to the T1 existed? Didn't they have to use a
>channel-bank connected to a seperate modems?
Lots of small ISP's are essentially equipped as you describe,
minus the digital cross connect system. The essence of your
configuration is that there exists only a single
codec/linecard/AD-DA on your loop, and to top that off the cable
between you and that linecard is exceedingly short! That should
result in consistently great connection speeds.
If you have or can get a USR Courier or one of the other modems
that have really detailed diagnostics it might be relatively
interesting to see what the modem says it is seeing. Lacking an
extra conversion to analog and back to digital it would seem
there are only two or three likely problems that I can think of
off hand.
One is if the T1 facility is suffering from clocking problems.
If you can determine definitively from the channel bank it would
be interesting. Otherwise if the telco end would look at their
receive from you and see if it is locked to their transmit. Or
if you can check the channel bank to make sure it is loop timed
to the receive signal. Another way to test that is by setting a
modem up for 2400 bps with no error correction or compression
turned on and the calling a BBS or someplace where you can get a
shell login. Watch for noise bursts, in particular with
repeating patterns like "{{{w3{{{w3" in them.
Another would be if part of this system is less than 64Kbps per DS0,
as in one of these ADPCM channel banks. I doubt that could be in
the digital cross connect, but it could be. Except I doubt you would
get as good of connections as you do if that was the case.
The only other thing I can think of right now is that you just happen
to have a line card in the channel bank that has a relatively narrow
bandwidth for some reason. There are systems with different filtering
on different versions of some line cards. Actually I've seen some
cards that were just plain defective (NEC Channel banks, though right
now the model number escapes me, but it is one of the older versions
and I doubt anyone but a telephone company would have any!)
Hmmm... a diagnostic reading from a modem sure would be interesting...
(I probably will not be able to respond to this thread again for
several days as I'm about to take a tour of the villages on the
Arctic coast west of Barrow for a few days... to diddle with the
bandwidth of satellite modems no less... :-)
Pierre Lepoureau
listen it sounds like the parameters are correct at b8zs esf and the
phone numbers for inbound doesn't make a difference on the bandwidth.
lets look @ this you have 11 64k channels for 800 inbound (Voice) and
512k for internet that equals 8X64k channels that equals total usage of
19X64k timeslots what's up with the rest of the timeslots "5." Look
into this?!!? also check you modem settings and if that does work go to
ISDN (Primary Rate 23 +1D) Later bro!