PRIVATE LINE DATA TRANSMISSION 1459 



accurate oscillator or "clock" at the receiver into the proper phase 

 relation at the beginning of each word. 



The phase of the oscillator is readjusted slightly as necessary at the 

 beginning of each word Init this adjustment is purposely made sluggish 

 so that an occasional noise burst will not throw" timing out badly. Thus 

 the timing information is, in effect, by its repetitive characteristic, highly 

 redundant. The recei\-ing device is capable of getting in step without 

 any manual adjustment but it may require as many as 10 words (or 

 that number of synchronizing pulses) before it gets into exact phase, on 

 an initial connection or after losing synchronization. 



2.1.3 Transmission Considerations 



Such a signal leads to the following transmission considerations: 



1. The signal is rather similar in its spectrum to a telephotograph 

 signal, and as a first order approximation requires about the same type 

 of facility for its transmission. 



2. In particular, the signal is expected to require something like the 

 same over-all delay equalization as the telephotograph signal. A brief 

 discussion of this point was given in a paper by one of the authors.' The 

 conclusions there are that the telephotograph etjualization limits of 

 ±0.4 times a signal element duration in envelope delay are generally 

 reasonable, although these are probably overly severe with respect to 

 very fine structure irregularities in the residual envelope delay charac- 

 teristic. The formal limits which have been set on the envelope delay 

 distortion for the 1,600 bit per second signal are ±250 microseconds. 



8. These limits constitute a rather less severe problem for the bulk 

 of the expected circuits (200-300 miles) than they present for telephoto- 

 graph circuits which are equalized for 3,000- to 5,000-mile lengths. For 

 the small expected number of very long circuits the problem would of 

 course be the same as for the telephotograph ser\'ice. 



4. The delay ecjualization problem recjuires special consideration in 

 view of the nature of practices which have developed in the telephoto- 

 graph art. The number of circuits which haA'e been equipped for tele- 

 photography has in the past been rather small. The adjustment of the 

 delay equalization has involved arithmetical calculation in the process 

 of fitting various manufactured delay equalizer sections to the measured 

 ilelay distortion. These methods are not generally suitable for large 

 scale operations. It is believed that a process of equalization by prescrip- 

 tion will l)e useable for the 200-300 mile circuits. This is discussed in 

 more detail in Section 3.3. 



5. So far, it has not been found expedient to transmit telephotograph 

 signals generally over unmodified XI carrier or other compandored 



