. 
154 POWER DISTRIBUTION FOR ELECTRIC RAILROADS. 
for the larger sizes. Slower speeds and higher voltages 
than those mentioned are very likely to reduce the effi- 
ciency by omne or two per cent, even more for small 
sizes running at unusually low speeds. If the prime 
mover is of low speed, such, for example, as a Corliss en- 
gine, it is quite easy to lose more in efficiency by using 
small direct coupled generators of, say, 100 k., w. or less, 
than would be lost by belt driving. 
Inasmuch as practically all railway work is at present 
done by continuous current, the energy received at any 
substation, transmitted by alternating current, simple or 
polyphase, must be changed into continuous current for 
use on the working circuit. There are various ways of 
effectiig this transmutation, all of them, unfortunately, 
quite inefficient compared with the results obtained from 
static transformers, and what is worse, all requiring atten- 
tion which, however slight, cannot be dispensed with. 
‘The most obvious plan is to employ a motor driven from 
the alternating circuit by belting or coupling it toa contin- 
uous current dynamo. Such is the simplest and often the 
cheapest method when existing stations are to be converted 
into substations operated from a transmission plant. The 
engine can be removed or merely disconnected, and a 
synchronous motor installed to take its place in driving the 
dynamos. This is the arrangement which has been used 
for several years past at Hartford, Conn., and Taftville, 
Conn., in both of which places the already existing gen- 
erators were driven from polyphase synchronous motors, 
The same practice is followed in the Folsom-Sacramento 
transmission. At the latter place generators for the elec- 
tric railway and for other purposes are driven from a coun- 
tershaft which receives its power from three phase syn- 
chronous motors. The generator room of the Sacramento 
substation, which is a typical example of the practice 
under consideration, is shown in Fig. 88. 
Obtaining continuous current in this way is often 
very convenient, but is most reprehensible from the stand- 
point of efficiency. It may answer well enough for the 
  
LB