One element of interest during the previous balloon flight was the clarity of the Moon pictures. The small Nikon S3200 has a too small lens to provide a consistent zoom. Even worse, the zoom cannot be controlled through GPhoto2lib on this camera model.
For the next flight, the plan is to use different camera , with interchangeable lenses and a telephoto objective. From our shelf, we decided to put at good use the Z-CAM E1 + m34 to Canon adaptor + Canon EF 55-200mm f/4.5-5.6 USM.
At such longer zooms the payload can no longer be left free and we must use a method to stabilize it and control its position precisely towards the moon.
This way we started experimenting with reaction wheel control.
Fig 1: Z-CAM E1 and Canon tele lens
For control I decided to use NI myRio system since I am familiar with LabVIEW and NI provides a wide range of signal processing and control functions
As reaction wheel a small BLDC motor was used driven by a common toy car ESC. the ESC is controller through pulse duration modulation between 1ms and 2ms , period between pulses it’s not critical, usually 20ms.
Fig 2: Prototype payload
To keep the orientation of the camera toward the moon we need a reference. there are a couple of options: electronic compass, image recognition of the moon, but we choose something I considered more interesting : a Sun Tracker.
For this a webcam attached to the myRio was used together with a wide angle lens 2.1mm covered in solar film. The only image the CMOS sensor will acquire will be the one from the Sun if this will be in the range of the camera.
Fig 3: Sun Tracker
Different types of controls were investigated to achieve accurate tracking using the reaction wheel: PID, Fuzzy Controller, and in the end most promising results were achieved using LQR controller… big thank you to NI for the self explanatory examples.
One downside of the reaction wheel is the saturation . When this happen the payload can start spin uncontrollable and the reaction wheel is no longer able to correct this. This is why a desaturation system is required. for this project a couple of ideas will be investigated: passive desaturation, or CO2 cold jet system.
On this second flight we will retry to send SSTV images as before. For this the “2m RF Shield” was redesigned to make sure RF influence over digital buses is minimized.
Fig4: “2m RF Shield” REV2 on left, REV1 on right
One important task was to evaluate the RF power vs DC power – Efficiency.
Fig5: Measure DC power and RF power during SSTV transmission
The 19.29dBm indicated by the spectrum analyzer is measured though a 20dB attenuator so the RF power is 39.29dBm ( 8.5W ), the occupied bandwidth is ~7Khz . DC power is 23.5W, so the efficiency is ~36% not great but not bad from an old hybrid power amplifier + raspberry pi 3 doing other things.
I knew from the start that the MyRIO it’s not only overkill from the performance point of view but in case the payload is lost the financial burden might be a bit too painful. All these details were accepted before I found a wonderful tool to replace it: OpenMV H7 . This is a small embedded module designed to perform image processing using Python.
Another aspect that had to be added into equation is the vertical position of the Moon, which varies around 10 deg/h, so a servo is required to tilt the camera accordingly.
The setup was changed from the Farnell box to a bunch of wood sticks on which all item were attached, most importantly the servo holding the Z-Cam E1 and the lens. Also the lens were downgraded from the 200mm Canon to the 100mm Panasonic Lumix to keep mass on servo low.
Fig6: Lab Test Setup
After many test we set the launching date to 18 Jan 2020. The angle between Moon and Sun is ~80 Deg, not ideal but good enough
Fig7: left – Moon elevation, right – angle between Moon and the Sun. Variation over 180min starting 18 Jan 2020 10:00 am Bucharest time
To be sure we recover our valuable payload we had 3 APRS trackers:
- one is generated by the “Pi RF shield” and has SSID -11
- second is a improvised PCB we call DirtAPRS to test PIC32MK, has the SSID -10
- thirst is a PicoAPRS Light, this would turn off at height above 2000m, SSID – 12
Fig8: felt – APRS tracking from YO3HCM-11, right – live recovery https://www.youtube.com/watch?v=ynO9_IMbrg0
Fig9: preflight pictures. Main camera trying to point towards the moon but it’s cloudy
The nacelle got to an maximum altitude of 28.5Km. the highers altitude from where we got a clear picture of the moon is 18.1Km. images from higher altitude were gathered but unfortunately these were out of focus
Fig10: Moon from 18.1Km altitude though 103mm FL on micro four thirds sensor
Andrei Hapenciuc 