106 SURFACE CONDENSERS. 
however, are valueless for comparative purposes because the conditions under which 
the experiments were made are not stated. 
The heat transfer of 2,750 British thermal units given in the first paragraph on 
page to1, if it were accomplished under high vacuum conditions, is a very remarkable 
result, but if this was effected under heater conditions it is no more than what is accom- 
plished in every-day results. 
Most of the condensers on which high heat transfers have been obtained heretofore 
have been equipped with tubes of 34 to 1 inch diameter, and it is difficult to determine 
whether the high heat transfer obtained in the tests on the condenser on the Cantigny 
are due to the general design of the condenser or to the fact that smaller tubes were used 
than is practical in land installations where the item of foreign matter in the circulating 
water is of large importance. 
There have been no direct experiments published in which an attempt has been 
made to determine the relative efficiency of the different diameter tubes. Theoretically 
the economic length of tubes should be shorter and the heat transfer higher, say about 
15 per cent for each 14-inch smaller diameter of tube. The matter is complicated by at 
least three unknown factors, and it is hoped some careful experimenter will conduct a 
series of experiments with air-free steam having this particular point in view. 
The general design of the Lovekin condenser for the battleship Tennessee, etc., 
embodies no fundamentally new features. The matter of providing lanes between the 
tubes near the steam inlet to the condenser is one that has been in use for a great many 
years, the writers having used such lanes in condensers some fifteen or twenty years 
ago and demonstrated the fact of higher heat transfer when a liberal amount of space 
is made for the steam to get at the tubes. The first use of “‘lanes’’ in the tube bank 
dates from the late seventies. 
While the lanes into the tube bank are aids to an efficient condenser, they do not 
cut down the steam velocities materially, as the residual velocity from the turbine 
buckets plus the space velocity is still of the order of 600 to 800 feet per second and the 
velocity of the tenuous water vapors approaching the tube surface must be of the order 
of 1,500 to 1,900 feet per second unless seriously slowed down by air aggregations in 
the immediate neighborhood of the tube surface. 
It is unfortunate that the author did not measure his air leakage before and after 
coating with bitumastic varnish. Good practice limits the air leakage into a 20,000 
square-foot condenser to 3 cubic feet of free air per minute while anything above 8 cubic 
feet should mean an overhaul and stopping the leaks. Of this 3 cubic feet at least one 
is due to air brought into the system with the feed water, the remainder being true 
leakage air. It is probable that the leakage into the 8,000 square-foot shell exceeded 
25 cubic feet of free air per minute if the vacuum was reduced so large an amount. 
In the matter of baffling so as to insure the condensate draining to the hot well 
without passing over any more tubes than is necessary, this is a modification of the 
Atwood Wheeler dry-tube condenser, a number of installations of which we have had 
in service for about twelve years. Practically all condenser manufacturers have used 
modifications of this type. 
The general arrangement of tube surface is very similar to that supplied by the 
Westinghouse Company in their radial flow type condensers, dry air suction being 
taken from practically the same points. The radial flow type condenser, as manufac- 
tured by the Westinghouse Company, provides a larger open space between the bank 
