On the morning of August 12, 2017 I and my co-conspirators flew a High Altitude Balloon mission out of the Wyandot County Airport, near Upper Sandusky, Ohio. One of the goals of this mission was to show we could operate a Raspberry Pi throughout the flight. I had hoped to fly an SHT31A Humidity sensor, but when I could not get the Explorer pHAT ADC to work, I replaced it with a Sense HAT carrying an HTS221 humidity Sensor. This flight showed that, as I had suspected, inside a payload box is not the optimal place to position environmental sensors.
Data that we can compare to is available from the National Weather Service, though I use a University of Wyoming site. Above are charts showing the 8AM balloon sounding data from Detroit and Cincinnati. They show the tropopause as being between about 250 mb, where the dew point drops drastically from the atmospheric temperature.
As an aside, Images of Lake Erie are available from the MODIS Satellite for about the time of the flight. This can be used to help understand some of the flight photographs.
Above are the pressure readings throughout the flight. At a first glance they appear reasonable. The pressure measurement did get down to nearly 5 mb. I did not expect the Sense HAT’s LPS25H to be able to read that low. However, at 105 kft, the Standard Atmosphere gives a pressure of 8.5 mb, instead of the measured 5.3 mb, a 3 mb or a nearly 60% error.
Relative Humidity and temperature measurements were also taken.
I don’t expect the humidity or temperature measurements to be very good during the ascent. The sensors are in the box, and during ascent the box is venting gas, the ambient air is not circulating. During descent the box is sucking outside air into the box, so I expect the measurements to be better. Temperatures of -20C during ascent and nearly -40C during descent. Temperature sensors are built into the pressure and humidity sensors, I presume to allow temperature compensation. The Pressure sensor’s temperature measurement appears to bottom out near -25C. The previous analysis of Sense HAT specifications suggested the humidity sensor’s temperature sensor is better.
The Humidity sensor spent a lot of time near 20%. I’m not sure I trust it at high altitudes.
The Standard Atmosphere agrees fairly well with the morning balloon soundings from Detroit and Cincinnati. It’s estimate is a little low, but not by much. Without other data, it looks like a good guess at the pressure altitude relationship. To convert times to altitudes I did a linear interpolation by time between the launch and burst and again between burst and landing. The rise rate was around 1100ft/min, give or take maybe 20%.
It is the descent portion of the flight where the sensor is reading high. The ascent is pretty close to the other curves. Though I admit I don’t understand why the readings are high. Incorrect temperature compensation? Pressure wave from the fall?
The pressure sensor did quite well until it reached about 12 mb much better than the specified 260 mb. At low pressures it may not be meeting its specifications, but appears to work. It may be possible to build a corrected data set to modify low pressure readings.
SOhio 2017-1 Relative Humidity and Temperature vs Pressure
There is some humidity and temperature hysteresis, differences between ascent and descent. This may be due to venting during ascent and ‘inhaling’ during descent. This is more pronounced with the temperatures.
The temperatures do not reach the expected -60C during descent. In the past I’ve thought this was due to the thermal mass of the computer boards. However, I am beginning to wonder if this might be due to compressive warming of the air inside the box.
The humidity sensor spends a lot of time reading near 20%. I don’t trust this reading. It could be due to being inside the box. Its also doing this at the coldest temperatures, and appears to be trying to function at the highest, slightly warmer altitudes. It is supposed to function down to -40C, though perhaps the pressure environment is inappropriate. I’d like to get the humidity sensor outside the box. I’d like to see it reading more like 10% or even 5%.
I have not done the dew point calculations because I don’t think the data is good enough to warrant the effort. For -20C and 20% humidity would give a -40C dew point which is too high for Stratospheric conditions.
The above chart is from my humidity sensors post.
At landing, the humidity reaches 100%. This is expected as the still cold payload lands in a higher dew point environment.
The pressure sensor did well on the ascent, but read high during descent. I’m not sure why this is.
Measurements inside the payload box to not represent the outside environment. It might be interesting to place the Sense HAT outside the box. But I doubt I will do this.
Next time I hope to use the SHT31A, outside the payload box. Hopefully this will perform better.