Summary of field testing of the glidersonde and powered sounding system in South Africa
(January 13-24th, 2002)

February 27 test flights
March 08 test flight
New!March 13 test flights New!
Screen shots of powersonde flight Figures Summary of costs Photos Photos (2)

The field testing of the powersonde and glidersonde, developed at the National Severe Storms Laboratory (NSSL) in Norman, Oklahoma, was carried out just south of Pretoria, Republic of South Africa (RSA), between January 13 and January 24th, 2002. Davis Egle, retired OU Aerospace Engineering professor, and Mike Douglas (meteorologist/NSSL) participated in the tests, along with Lucian Banitz from the South African Weather Service (SAWS). The field tests were sponsored in large part by the International Activities Office of the US National Weather Service.

The field tests were carried out at the Irene Weather Office, about 20 km south of Pretoria, and at a nearby model aircraft flying field.

The original plan was to carry out high altitude ascents of the glidersonde, and then additional testing of the powered aircraft radiosonde system ("powersonde"). During the first three days after our arrival, the glidersonde system was tested, but the Irene facility proved too small for landing the glider easily. Gas was easily available from the radiosonde facility at Irene, but one glider was damaged during a descent from the balloon, and rather than risk losing the flight computer, we decided to proceed with powersonde testing.

In addition to glidersonde problems during the first week of testing, several aircraft had to be built up during the first week, delaying the powersonde tests. In addition, Mike Douglas gave 3 lectures to the SAWS staff during the stay; one day was effectively lost to testing due to our participation in the visit to the main weather service office in Pretoria (well away from the Irene weather office). On other talk days, both Lucian Banitz and Davis Egle continue with testing and only Mike Douglas went to the main office.

On the positive side, the aircraft successfully navigated during both ascent and descent portions of the flight. During descent, when 500 ft above ground level was reached the engine throttle was opened by the flight computer and the aircraft climbed again. When 6000 ft Above Ground Level (AGL) was reached the engine throttle was closed by the flight computer and the aircraft descended. This basic flight mode was successfully demonstrated not only on the last flight (the day before we left) but on a number of previous flights as well.

During the second week of testing the powersonde was flown several times on successive days, with increasing height coverage. On the last day of the tests the powersonde carried out a particularly impressive flight test. Using 32 oz. (~one liter) of gasoline (.25 gallon) the plane flew (with radiosonde under one wing transmitting to a ground station)for about 40 minutes, and made 4 successive ascents to 6000 ft. above ground level without landing. During this and several other autonomous flights, the operator needed only to pilot the aircraft during take-off up to an altitude of 200 feet and during landing from less than 500 feet. All this time the plane was navigating around the home point, never exceeding 1 km from home and most of the time staying within 300 m of home (Figures 01, 03). The engine performance was impressive considering that the altitude of the airfield was 4850 ft. above sea level, so that the aircraft was climbing to about 10,800 ft above sea level. Although the aircraft was still climbing at this altitude (Fig. 02) and could have reached higher altitudes; we chose to limit the altitude to 6000 ft AGL for safety reasons.

No apparent hysteresis problems were noticed with the RS-80 radiosonde during our tests; however a later test flight after we left from South Africa showed a problem (see figure 08) that is being investigated. However, a data recording glitch on the laptop used to record the data prevented any radiosonde data from being recorded. Although this was disappointing (the problem with recording the radiosonde data has since been solved), the key point was that there was no interference between the radiosonde transmissions and the flight computer or any other component of the powersonde.

Several technical problems were identified during the flight tests.

(1) A change between the pressure sensor determined altitude and GPS-determined altitude, which took place at 3000 ft AGL, created some spurious height data (since corrected).

(2) On some of the descents, the descent rate was less than expected due to the throttle changing from a low to intermediate value when the aircraft was further away from home than a preset value (1000ft/321 meters). Because of the rather stiff winds on the last flight day, the aircraft would fly farther than 321 m from home moving downwind. This caused the flight computer to open the throttle to a "cruise" setting at which the aircraft would not descend. The computer set the throttle back to the descent setting when the aircraft returned to within 321 m of the home point. This problem could not be corrected in the field but is easily corrected by reprogramming the flight computer.

(3) Finally, the code to determine when to cut-back the engine used the criterion that if the aircraft descended for more than 3 seconds the aircraft was assumed to be unable to climb and the descent mode (with powered back engine) was triggered. Because the aircraft did sometimes, in turns, descend for more than 3 seconds, the descent mode was triggered and this prevented the powersonde from climbing as high as it could have. This was part of the reason we set 10,800 ft ASL as the upper limit for the tests — we could not change this part of the code of the flight computer in the South Africa. This setting will be changed to something like 10 seconds, to better judge when the powersonde cannot climb any higher.

The Irene facility, consisting of a meteorological radar, radiosonde facilities, surface station, and associated instrument maintenance shops and storage facilities, is very good for hardware development and testing of the glidersonde and powersonde. However, the office lacks a suitable runway for testing of the vehicles. Our field testing at a model flying airfield, 10 km away from the Irene weather office, took about 3 hours each day away from our test activities. This is because the aircraft had to be loaded into a trailer, the trailer attached to a car, the car driven to the airfield, the equipment set up, and the opposite procedure repeated during the return. If a model aircraft flying field had been located at the Irene weather office, this time would have been available each day for testing. The Irene Office is situated on a large Government-owned agricultural research station (farm) and on the weather service property there is space less than 100 m away from the main building for the construction of a dirt runway. This possibility has been communicated to the SAWS; such a runway would greatly facilitate future testing and training of personnel who might fly and maintain the powersonde.

Additional positive aspects of the testing were that the personnel of the SAWS were supportive of the test, helping wherever needed, and that the daily expenses in South Africa were very reasonable (due to deflation of the value of the Rand compared with the dollar). A newer version of the glidersonde is being readied for testing; the parachute deployment (for automated landing of the glider) has also been demonstrated and works well.

Back

Approximate hardware costs for the South Africa field trials

2 PowerSonde Aircraft @ $3,285 each $ 6,570

Spares for PowerSonde @ $2,785 each $ 2,785

2 GliderSonde Aircraft @ $2,195 each $ 4,390

Spares for GliderSonde @ $1,975 each $ 1,975

 

Total Hardware Costs $15,720 (excluding ground station)

Travel related expenses of Davis Egle and Mike Douglas: ~$6,000

Notes:

• Costs do not include labor for assembling airframes nor integrating components.

Spares consists of complete airborne electronics and powerplant, all but the airframe.

• Much of the hardware continues to be used in South Africa and in Norman for flight testing.

Costs Details

PowerSonde

GliderSonde

Navigation Computer

$ 500

$ 500

Servo-Computer Harness

$ 50

$ 50

RC/MC Switch w TLM Control

$ 150

$ 150

MIM Board

$ 100

$ 100

9.6 v 900 mah Battery

$ 20

$ 20

Laptop Extension Cable

$ 15

$ 15

GPS Extension Cable

$ 15

$ 15

Pitot Tube

$ 50

$ 50

MC/Yaesu HT Audio Cable

$ 25

$ 25

GPS receiver

$ 200

$ 200

Wood

$ 100

$ 50

Composites

$ 200

$ 100

Servos

$ 500

$ 300

Receiver

$ 200

$ 200

Switches

$ 60

$ 45

Landing Gear

$ 50

Cables

$ 100

$ 50

Batteries

$ 100

$ 50

Engine & Accessories

$ 550

TLM TXR

$ 200

$ 200

Misc

$100

$75

$ 3,285.00

$ 2,195.00

Ground Station

Laptop Cable

$ 25.00

Data Capture Program

$ 250.00

Config Program

$ 250.00

Receiver

$500

Terminal Node Controller

$350

Laptop Computer

$2,000

Antenna

$200

RCTransmitter

$1,000

Other cables

$100

$ 4,675.00