The Roswell Report: Fact vs. Fiction in the New Mexico Desert

The following statistics for a given pressure (1001.8 millibars) were computed:

 N.Y.J. Shop Model L-416	Brailsford Model B-501
on the mean	12.5%	34%
within 0.1% of mean	25%	50%
"	0.2%	"	41.5%	70.5%
"	0.3%	"	62.5%	85.5%
"	0.4%	"	75%	91%
"	0.5%	"	95.6%	100%

Other conclusions arrived at as a result of this test were:

(a) Since changes of speed of the motors did not occur simultaneously in the two instruments, the speed changes probably are not due to slight changes in pressure or temperature.

(b) Sensitivity varied from 0.1 to 0.9 millibars.

(c) Sensitivity increased with increase of rate of pressure change.

It was recommended as a result of these tests that the records of flights when the balloon is floating be read to the nearest two-tenths of a percent of a cycle, or approximately two-tenths of a millibar, for high accuracy. When using the instruments manufactured by Brailsford and Co., satisfactory accuracy will be attained, if necessary, when the record is read to the nearest one-tenth of a percent of a cycle.

In the second group of tests the pressure was reduced slowly to about 20 millibars and increased to sea-level pressure at different temperatures.

The most comprehensive series of calibrations was made with the first instrument made by Brailsford and Co. Two runs were made at room temperature (229C), one at -100C, one at -30 to 370C and one at -56 to -620C. On the last test at the lowest temperature, the unit was found to be completely unreliable. The cause of failure was the erratic motor operation at extremely low temperatures. This had been observed previously during flights when the Olland-Cycle was not thermally insulated.

The other curves were plotted on a single chart in order to study their spread (see Figure 31). The envelope of curves thus obtained showed no regular temperature effect over the range $ + 2 2^{\circ} \mathrm{C} $ to $ - 3 7^{\circ} \mathrm{C} $ . In general, the envelope was less than 10 millibars wide although at some higher pressures it was as much as 12 millibars wide. The curves at low pressures fell closest together and were all within 3 to 4 millibars apart between 50 and 150 millibars and 6 millibars apart between 150 to 200 millibars.