On this page you will find some information about the equipment I will use for FPV flying.
Video Transmitter and Receiver
The transmitter and receiver was taken from a video link set like this:
Such set includes a big transmitter, which was rated on the box for 300mW. A measurement discovered >500mW of output power using a tuned antenna for good VSWR.
Another video link set exists with a small transmitter rated for 25mW. Measuring the power showed really 25mW output power.
To compare the transmitter size:
I disassembled the receiver and put the tuner and PCB in my DIY antenne tracker, see below.
Since the video downlink is working in the 1.2GHz range I ordered a patch antenna RE1208P-SM from L-com which covers the range and the size and weight is not to much for use with a DIY antenna tracker.
The plots show that the antenna is only usable in a maximum range from 1110-1190 MHz where the VSWR is <2.00:1. Best VSWR is 1.047:1 @ 1149 MHz which is close to channel 3 (1160 MHZ) of my video downlink which I will use then. So the antenna parameters, at least the VSWR, provided by the manufacturer seems to be dressed up a bit.
Since I replaced the coax cable with a more flexible one, due to use of the antenna with my tracker, some pictures are discovering the inside of the RE1208P-SM antenna. By this chance I took also measurements of the antenna patch construction.
after certain time with the patch antenna the following Yagi was reviewed by members from RC-groups and rated for very good results. So I could not resist to order and try myself.
The new Yagi came from China for around 15€ and free shipping. This antenna performs best for the price. A range test was stopped at 13km due to lost line of sight. But it’s sure 15km+ can be reached as long as you have line of sight to the transmitter.
Although the Yagi is rated from 1120-1280MHz it worked best for me at 1200MHz. Therefore I changed my transmitter frequency to a new channel. Side effect of the frequency change is also that I got rid of the LPF. At 1200MHz GPS is no longer affected by the VTX.
You can order this antenna at eBay: http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&item=280402380910
On the transmitter side it makes only sense to use an omnidirectional antenna since the transmitter will move around during flight.
A usable solution will be a lambda/4 or lambda/2 dipole, where the halfwave dipole has the advantage of a small or even no ground plane required.
I read through all types of possible antennas and there advantages and disadvantages, also about the construction details and impedance matching.
My first try will be a coaxial antenna made from a small brass tube and RG174.
Such antenna is easy to construct and matching is also easy since you just need to cut the upper part of the dipole for best VSWR.
For all tests the brass tube remains the same length of 61mm (Reduction factor: 0.95*lambda/4) and inner/outer diameter of 5mm/6mm.
The first attempt for impedance matching was to measure the current drawn by the transmitter.
With 74mm length of the upper part the lowest current around 190mA @ 12V was drawn. Later on a VSWR measurement showed best resonance at 1019MHz which is far of the desired frequency 1160MHz.
While connected to the VNA I cuted the antenna until I got the best VSWR for my desired frequency at 1160MHz:
ending up with a length of 56mm for the upper half of the dipole and a current of 188mA @ 12V drawn by the transmitter on power-up. The length of 56mm corresponds to a reduction factor of 0.86*lambda/4 which is far of the 0.97 I found in the books.
Here are the resulting details of my coaxial antenna for the 1.2GHz video transmitter:
One important thing to keep in mind for such type of antenna is the fact that the lower part, the brass tube, must be isolated to ground because the tube is a radiating part as well!
While the receiver antenna has been changed the one on the transmitter was changed as well.
Now I use a DIY IBCrazy inverted-V antenna, measured and tuned at 1200MHz.
As shown on the picture the inverted-V antenna is now integrated in the leading edge of the left wing. With both antenna’s, the eBay Yagi and the inverted-V, I archived the range of 13km LOS while video was crystal clear.
See this thread on RC-groups about all details of the inverted-V antenna:
Also check IBCrazy’s blog on RC-groups for more DIY tutorials regarding FPV:
Using a a medium or high power video transmitter together with a 2.4GHz remote control will cause heavy interferences due to second or third order harmonics. A drastic loss of performance and decreased range of the remote control will be the result.
Also when a GPS receiver might be influenced by a 900MHz or 1.2GHz video transmitter which could lead to decreased performance or even to the complete loss of the GPS fix.
In all cases a low pass filter between the video transmitter and the antenna will help to reduce interferences to a minimum.
The plot shows the performance of the LPF-900 low pass filter in the frequency range 800MHz – 2.5GHz.
Attenuation for the GPS L1 frequency at 1575.42MHz is >37dB while at 2.4GHz the attenuation is even higher with 52dB.
Pass through attenuation is <3dB at 900MHz and 1.2GHz.
Pass through attenuation is <2dB at 900MHz and 1.2GHz.
Attenuation for GPS L1 is 32.5dB and 39dB for 2.4GHz.
All in all such filters are suiteable if you have interference or performance loss with your GPS or 2.4GHz remote control.
This is my DIY antenna tracker.
Based on a Hammond metal box it includes everything necessary for video and telemetrie downlink reception.
The arm beside the patch antenna will hold a 5 elements Yagi for 868MHz later on.
The lower part of the box holds a 5V/2A power supply, a 12V/1A regulator and a part of the video receiver.
The upper part of the box holds the pan servo, video tuner module, the tracker controller board and the RSSI indicator.
The tracker controller is based on ATmega162 and FT232R. The board accepts control inputs from a PC by USB and uses the Pololu protocol for servo commands.
One serial input of the ATmega162 will connect to a Wi.232 moduleto receive telemetrie data and of course the aircrafts position. A second serial input can receive optional NMEA data from an external GPS module (not used for now).
By a jumper block the serial signals can be routed from and to the PC and/or the controller board to the Wi.232 mainly for use with a ground station software (MK-Tool, Mission cockpit) on the PC.
Actually the controller board can drive 4 servos but only two are in use. With 4 jumpers the servos can be inverted independently from software.
The connector front provides USB, video and audio, power input and a DIP switch to select the video channels.
In the following picture you see the top of the box without pan&tilt mechanic, only the pan servo output shaft is visible.
The whole mechanical part is fixed by a M2.5 screw to the pan servo shaft. The round base plate takes the wheight of the tilt mechanic and prevents it from shaking. For easy moving I put some grease on the bottom of the base plate.
The pictures are showing also the cable routing for -90°-0°-+90° pan angle.
On one side of the tilt mechanic is a single ball bearing mounted, on the other side the servo output shaft has two ball bearings.
Also the counter weight, 220g of Lead, is visible.
A word about the servos: I’m using standard RC servos, size 41x41x20, with full metal gear and two ball bearings on the output shaft. The output shaft is from metal too.
Since a GPS data downlink is anyway necessary to operate the antenna tracker I thought showing some of the data would be nice.
So I upgrade the tracker with an LCD. Because the box is not that big and space is limited an LCD from a Nokia mobile phone was used. I had some displays from the 3310/6150 series laying around. With 84×48 pixel it gives a good 14×6 character display. That’s enough to show various data from the GPS.
After power ON the tracker is not calibrated and waits for GPS data from the aircraft. As long as calibration is not initiated the GPS data is just shown like above.
“S 09” means the number of sattelites used.
“M A3” tells the mode of GPS operation, in this case Automatic with 3D solution.
The tracker is now calibrated and some more informations are visible.
The altidude above ground – AGL – is the difference between the GPS altitude stored during calibration and the actual aircraft altidude. “SR” is the slant range (or line of sight distance) to the aircraft.
In case of a crash out of sight the tracker will always show the last know (received) aircraft position. This will be very helpfull to recover the lost aircraft.
The following files will include the schematic, board layout and firmware source code for my AVR based FPV antenna tracker:
Download the Target3001 Viewer to open the board layout: DOWNLOAD
Schematic of AVR Controller for my FPV Antenna Tracker
Board layout of AVR Controller for my FPV Antenna Tracker
Firmware source code for my FPV Antenna Tracker
|License:||GNU General Public License|
A documentation will be available soon.
The full operational antenna tracker:
The Yagi antenna is used for the 868MHz data downlink established using a pair Wi.232 modules.
Receiver Modifications and Improvement
Obviously the most important modification is the improvement of the tuner module inside the receiver. Inside the tuner module is a SAW filter in the IF section with a bandwidth of usually 27MHz which is good for satellite LNB but this is way to wide for terrestrial use.
With this deviation, the better solution is a 17 MHz wide SAW filter which means sensitivity is increased due to less noise floor presented to the IF and demodulator. This modification have dramatically improved receiving performance.
See the following discussion in rcgroups.com for more details about the modification: http://www.rcgroups.com/forums/showthread.php?t=1028300
Now watch the following video it will explain all:
Here are some pictures of the tuner modification for my video receiver.
The original 27MHz SAW filter in my tuner was an SBU4872U.
The original filter used 4 pins( 2 GND, In, Out) while the replacement ECS-D480A has only 3 (GND, In, Out).
Pinout for both filters is the same so easy to replace.
The new 17MHz SAW filter ECS-D480A from Digikey.
I was lucky to get hands on a broken EVG920 for some €, almost complete but without IR remote control.
As revealed on the first test the battery pack worked fine but there was no video on the displays.
After disassembling the video goggles and cleaning the PCBs with alcohol they worked fine for some time. But frequently the displays turned black. I also noticed when touching the display wiring the video was mirrored horizontally.
Due to the ROHs compatibility the soldering of the PCBs looked awful. So I re-soldered all components by hand except the SSD5102 BGA package.
Since then the first problem, black displays, was gone and it was obviously related to bad soldering.
The second problem with mirrored video was solved with a pull-up resistor on a signal line going to the displays. There are two signals on the displays to mirror the video horizontally and/or vertically. Both signals are controlled by the SSD5102 display driver. It seems there is a bad soldering on the BGA package as well since the signal for horizontal mirror was floating. With the pull-up both displays are working fine now.
The video decoder TVP5150 and VGA display driver SSD5102 are controlled with a PIC12F683 by I2C. On power up the PIC initializes some registers on both video ICs. During operation the PIC polls frequently the video format from the TVP5150 and changes the configuration of the SSD5102 if the video format was changed.
(I tried to read the PIC12F683 but obviously code protection is enabled 🙁 )
Also IR remote controlling of the EVG920 is handled by the PIC. But since I have no IR transmitter I can not do any adjustment.
For the TVP5150 and SSD5102, datasheets are available where all control registers are explained. So building a DIY controller, maybe with AVR, is easy. An advantage would be the full control over the video goggles.
The picture above describes the pin-out for the PIC12F683 controller of the EVG920. The I2C bus is self explaining, the reset to the SSD1502 is essential on power up for clean initialization of the displays. Pin2 seems not connected somewhere while pin 3 is used to blank the right display when forced HIGH, must be LOW for normal operation. I think this pin is used as some sort of shutter when the goggles are operated in stereo or 3D mode.
Here are the datasheets for the ICs used on the EVG920 board:
SSD1502 Display Driver for Micro VGA Color Display
TVP5150AM1 Ultralow-Power NTSC/PAL/SECAM Video Decoder
Thanks to Sebastian from Globe Flight for the IR remote controller he sent me. For now there is no urgent need to continue with the goggles controller based on AVR.
Increasing the eyes distance
When everything worked fine I just cut the plastic frame which holds the video displays together into two half’s. 😀
Why? Simply my eyes distance is so width that I could not see correctly with the standard display distance. A piece of metal sheet with four slots, four screws and the video goggles fit for me.
Of course the plastic housing will not fit anymore but since I want to use the video displays anyway with ski goggles it doesn’t matter.
The modification of my video goggles are done.
Everything fits now in nice ski goggles:
Since the video optic doesn’t fit anymore into the original housing I made a new housing from 1mm plywood to cover the optic and electronic.
Only the original IR window was re-used.