[Asterisk-Users] Hardware Compatibility and zaptel driver

[Steve Underwood]

Nir Simionovich wrote:

>Hi Steve,
>
>  Hmmm.... lets see, NOT !
>
>  In Israel the connection of the E1 at the wall socket is not an RJ45, but
>a British Connector.
>
>  I'm fully aware of the fact that the connection of E1 on the Router/Server
>side are 1,2 and 4,5 as pair for TX and RX. Steve, my question may sound a
>little stupid, but
>here is my current setup.
>
>1. The cable connecting from the wall to the server is a flat 4 wire cable,
>used by our Telco.
>2. The cable also has a noise filter on it, the same one used for Data
>Connection by our Telco.
>3. The cable is around 60cm long, which means that it's just the right
>length.
>4. The length of connections from my wall socket to the E1 Grouping device
>is around 30 meters,
>    well within the specs.
>
Maybe that 30m cable is paired incorrectly.

>  However, what you said poses a really interesting question. Although I'm a
>CCNA, I couldn't find an answer to it. As far as I know, E1 is supposed to
>be an un-balanced
>medium, which suggests that there is no significance to the order of wires
>within the pair. Question is, is this true? If not, this might be the cause
>of problems. However, the line works fine
>from my side, and I due to the fact that I can't see any errors on my side
>(actually, I don't have an
>ability to see errors), I'm a bit at a loss.
>
Er, unbalanced means there _is_ significance in which way round you wire 
things. Balanced means there is not. In the case (or cases) of E1 you 
can get either. If you have is supplied on TP is should be balanced. On 
co-ax it will be unbalanced. You have TPs, so your is balanced. However 
this is entirely irrelevant. E1s and T1s can both tolerate reversal of 
the data.

>  Now, here are the symptoms according to our testing (The Telco and
>Myself):
>
>1. We've managed to bring down the number of errors to around 60 per hour.
>    This was achieved by changing the default ToneZone in zaptel.conf to FR,
>while before it
>    was set to UK. That's why I believe that the problem resides zomewhere
>in the zonedata.c file.
>
I don't think so. zonedata.c only affects the data content of the E1 
stream. Your CRC4 errors mean you have bit errors in the stream itself. 
Changing the tone zone has probably just changed the data on the line to 
something more benign that is giving less trouble.

>2. When errors go up, I can create outgoing calls, but incoming calls
>receive busy signals on all         PRI channels.
>3. The errors seem to go away without any specific reason. The only common
>denominator
>    for reseting the errors is, performing several resets to the zaptel
>driver and performing a
>    reset to the Telco System 12 connection card. And only after a variable
>number of resets,
>    everything works fine again.
>4. It can work fine for a couple of days, and then start making hell.
>
Problems tend to come and go when things are marginal. It sounds like 
crosstalk, weak signal or some such PITA. With T1s you can get quirky 
problems like this with AMI vs B8ZS mismatches. E1 is always HDB3, and I 
can't think of any other optional things which give almost but not quite 
right results.

Regards,
Steve