QUESTION:
Is it a fair statement that, to properly operate, a digital PBX with digital interfaces to more than one carrier (local or long distance) with DS1 service into digital switches at the carriers' POPs, requires that timing be taken from one of them (should not matter which), and all timing has to be Stratum One traceable, and the PBX clock must be 'slaved' to one of the DS1s (or be fed from a Stratum One clock on site). That is, all digitally interconnected DS1 devices and lines must operate at a common rate; and since the carriers are Stratum One, everything else must be, too.
ANSWER:
The short answer is... YES. However, that does not fully address what happens if the PBX timing reference is not up to snuff. Generally speaking, most PBXs have an internal clock that is Stratum 4 quality (32 ppm) so this sets a lower threshold on how bad things can get.
One way to appreciate the need for good timing is to consider what happens to the analog signal from a telephone connected to the PBX. The PBX "line circuit" contains the codec where the analog-to-digital conversion takes place. This involves sampling the analog signal at 8 kHz (nominal) and encoding each sample with 8 bits for a net bit-rate of 64 kbps. The codec clock is slaved to the PBX time-base (either internal or locked to an external reference). These samples will, eventually, be converted back to analog in the codec at the distant end.
If the conversion clock at the A/D and the D/A are not equal then some mechanism to "equalize" the rates must be employed. This is the basis of a "controlled" slip whereby one sample is either deleted or repeated. This slip operation could take place anywhere along the (circuit-switched) path between the codecs, most probably at the Digital Switch to which the PBX is connected to.
The effect of a slip, for human-to-human (voice) communications is not catastrophic, especially since we have been conditioned in recent times to the often abyssmal quality on mobile telephones. The effect of a slip could be a pop or crackle or, if it happened during a silent portion of the speech, nothing observable at all. However, the effect on machine-to-machine communications can be quite disasterous. Fax machines and modems do not like such slips, do not have periods of silence, and do not have the ability of the human auditory system to mask imperfections.
Considering that the distant codec could be anywhere in the world, the logical way to address the frequency difference issue is to make all codec clocks Stratum-1 traceable.
|
