1n high al ti·tudes# since it freezes below -BOOC and will flow
readily at low temperatures. Also# this fluid will absorb only
a very slight amount of water which might freeze aloft.
Basically three different principles have been used in the control
of ballast flow. The first of these is calculated constant flo·•;
the second is displacement-switch control; and the third is rate-
of-ascent switch control.
(1) Constant Flow
In the simplest of the control systems, liquid ballast is allowed
to flow continuously through an orifice (Figure 8) at a pre-
determined rate. This rate is set to slightly exceed the ex-
T~~
i
1
.SECTION .A-A \ i ~1tA8CMJC
•O$TAI4 B
A
Figure 8. Orifice for fixed-rate ballast flow.
pected loss of lift of the balloon due to leakage and diffu-
sion. If this method is successfully used, the balloon stays
full because the gas remaining in it has less load to support.
Therefore, the balloon will rise slowly as ballast is dropped,
maintaining equilibrium between the buoyancy and the load. In
the General Mills 20-foot balloon, for example# diffusion losses
are about 200 grams per hour at altitudes near 40,000 feet. The
balloon at its ceiling of 40,000 feet with a 26-kilogram pay-
load rises about 700 feet with each kilogre.m of ballast dropped.
This means that such a balloon using this constant-flow type
control will float at a "~ing" which rises at the rate of
about 140 feet per ·hour. Constant flow was first obtained by
use of the manual ballast valve shown in Figure 9. Due to
excessive clogging of this valve, caused by itsannular ring
opening, gate -type valves were tested, and finally the use of
-17-