Saturday, October 27, 2018

Ground Station Setup For Converged IoT Platform

Antenna Cluster Planning

The mission scenario sketch example is in the following picture. The small matchbox at the lower left corner is the base station. The modeling is made in Blender. And, based on the mission scenario, the transmitter/FPV antennas are arranged without internal interference or obscuring/blocking signal line of sight.

Transmitter Setup

FrSky R9M 900mhz Long Range module is chosen as a substitute of TBS micro module to save modification in the QX7 transmitter itself. 10-pin, 1-meter ribbon cable connects TX pins to module. Breadboard jumpers tethers the module.

The Frsky R9M module can be set to output 500mW using Taranis transmitter's regular power pins without external power source. The flight video shows long distance dive without any transmitter RF glitches, which would manifest itself in fail safe mode level free-fall aborting diving. 

This flight sequence rehearses the mission plan, diving from 333 feet east, 444 feet north, ground distance 555 feet, air distance 1000 feet, relative to ground station. 3 flips above the landing patch area. Level traversal between 333 feet east, 444 feet north, and 333 feet straight east.

Video Receiver Setup

The Hyper Tough cigarette power plug is specifically chosen because it has very tight fitting to fix a suspected cause of crash of pilot's knee knocking the plug during flight.

Retractable USB-micro cable powers to the FPV receiver. Its host plug breaks out on the chassis column and is accessible for quick disconnect after craft landing to reduce heating of the receiver.
The FPV receiver explicitly based on the RX 5808 module is chosen because experiments on the FXT Triversity receiver had video "blueouts". The real fault point, including all previous blackouts, is the FXT Triversity receiver, not the FPV goggles.
The video output pin 6 is soldered to the AV yellow cable for video on the PCB. Pin 3 is ground. The tail end of the AV cable and the row of pins, including the pin base are encased together by the clear mounting tape. The male plug of the AV cable goes to FPV goggles. On the True-D receiver, the casing groove is sanded down to level the surface with the clips. Then 2 Velcro strips sandwich the clips and the flattened surface together.

The alternative RJX receiver without casing uses the 3/4 inch velcro strip to wrap around the lower corner LED as an anchor, then it goes over the bottom side of the receiver and up. There are 3 mistakes in the following 2 pictures though. First, the black Velcro traps heat. White Velcro should be used. Second, the antenna plugs make a natural tilt angle for the velcro strip to slide down on the RF cavity cover, but during a maintance, pulling back the Velcro also pull open the RF cavity. The wrapping should stop at the edge of the RF cavity. Third, I left the receiver on the hot roof top of my car with full power on it between flights in a summer noon heat, and the blackouts reappeared when signal is weak and there should be white snowy noises. I should have the habit of disconnecting USB cable from the chassis column right after landing the craft.

The retractable USB coiling mechanism needs to be removed as the picture-in-picture below. The right angle SMA adapter for the mushroom antenna needs to reverse the polarity with the inner core pin stem on both sides because both the mushroom antenna and the receiver side has the inner core pin hole. The patch antenna with 9.4dBi is the directional antenna projecting the craft's diving trajectory. The 45 degree SMA adapter is essential in making the near vertical radiation pattern.

The receiver antennas allow quick release folding. Just loosen up the lower antenna's 90 angle connector with 5-6 counter-clockwise twists, then fold the upper antenna backward 90 degrees and fold the the lower antenna backward 180 degrees.

Mod FPV Googles For AV Input

The FatShark Velcro sticker does not need to be removed. Just use an art class knife to cut a cross on the 4 outer bore holes, and the Hypertough #1 driver can unscrew the screws. For the inner 2 shallow holes, one wedge of the crosscut needs to be sliced off to give room for the driver. Before mod grinding begins, tape over the optical cavity and the MicroSD opening to block saw dust.
The small diameter rotary disk will grind the glass fiber of the RF module. Hollow out the fiber glass between the top and bottom header leads first, then polish away the solder until the art class knife can slide in between the RF module and the main board from the left edge to the center.
To prepare the AV cable, stick a piece of transparent mounting tape on the main board to temporary fasten the cable and adjust the cable leads. The ground lead goes to pin 3; Signal wire goes to pin 6, same as diversity receiver's pin-out order. Stick a piece of paper underneath pin 1 and pin 2 of the RF module to block power supply to the module.

Use the 1mm milling bit to notch the SMA antenna hole for the AV cable and thread the cable through the hole before soldering as above picture-in-picture. Tape the cable one on top of the temporary mount and another next to the SMA antenna base after soldering. 
After reassembly, to adjust the fitting, use cuticle scissors to slash the cushion tube and the foam laterally. Real human faces are not completely symmetrical and I need to remove the left side cushion to align my pupils to the center. Also removing the side cushion sections flattens the curve.

The original Makerfire off-brand EV800D has good DVR video quality and is a alternative substitute. The original version has the circuit board seen in the picture on the right. The USB charging port is fragile and can easily break when USB plug dangles on the side. The picture on the right has the repairing Picoblade plug. We can modify a USB sync cable to terminate with any plug the picture-in-picture on the right.

The worst case noise interfering with FPV feed is with 12-16V direct battery power supply to the Caddx Turtle cameras (v1, v2, and Baby) in the following video if you watch the original feed video file (download it then watch it on a video player instead of watching it online with online player because online video servers process out the noises to save their internet bandwidth).

The FPV feed is clear of interference lines when using VTX03's 5V aux power, which filters out low frequency ripples. But, the voltage sensor of the camera only operates between 10.0 and 12.0 volts. Any voltage outside the range has the idle display of "12V".

The Wolfwhoop power supply, with 10uH inductor, has the video ripple level between direct battery power and VTX03's aux power. The direct battery power has the same level of ripple noise as the Bluesky 5V power supply, which has a 4.7uH inductor. This is because that the Caddx and Runcam split camera systems have a internal switching converter to generate 5V when power supply isn't 5V, and that buck converter has the same oscillator 4R7 inductor circuitry as the Bluesky 5V power supply.

The ripple noise can not be eliminated with a 10uH inductor LC filter because the ripple frequency is similar or lower than the 10uH inductor working frequency. The ripple noise can not be eliminated with a electrolyte audio A/C capacitor because the ripple frequency, around 1MHz is much higher than the electrolyte capacitor's audio working frequency. The runcam system consumes 600mA, and can not be powered by VTX03's aux power supply, which is limited to 360mA. For this reason, the run cam needs dual Wolfwhoop power supply in parallel.

In-Car Battery Charging

Fight batteries should be warmed/cooled with car vent using a gardening wire to hook up the charging plug before charging in cold/hot temperatures. 

When the car engine is offline, the car battery voltage has a slight drop, so the flight battery charger needs to be set to output 1.2A instead of the usual 2A, to prevent low voltage cut-off during charging, which can happen in the middle of a mission. The ToolkitRC charger has a continuous USB power output suitable for FPV goggles

Tuesday, October 16, 2018

Blade 180 - Oxy 210 Rotor Characteristics

Product source is ebay seller amain. Item "OXY2-081 OXY Heli Carbon Plastic Main Blade 210mm (Black) (2)"

The airfoil profile is (NASA predecessor)NACA0015. The symmetrical airfoil is generally known to have the center of pressure CP at quarter blade width from the leading edge during cruise flight when attack angle is smaller than 2-3 degrees in research paper. As shown in the picture below, the CP point is at 17.5mm/4 = 4.375mm from the leading edge but the bolting is at 4.3mm, so, when the attack angle is very small, it has perpetual pitch suppression. The research measurement also shows that the airfoil has perpetual up-pitching torque at cruise attack angle of 6 degrees because of "center of pressure moving forward of the quarter chord". 

When the mechanical margin of error is added between actuator servos and actual blade position, the blade angle shifts as the following time line.

This means that large slop gap can result in an unstable craft. And in the above example depiction, the slop should be improved from 3.2 degrees to 3 degrees to escape the perpetual pitch suppression slump. And to avoid over-correction, the slop needs to be smaller than 2.1 degrees.