GoatBoy
04-02-2017, 11:27 PM
In case you guys were wondering, when I’m not off annoying paintballers with my pedantry, I’m sometimes annoying RC modelers with my pedantry.
The analog 5.8ghz video systems in common use are really… regressive technology. However, they’re plentiful, cheap, and easier to fit onto micro models.
My recent test platform has been a brushed micro. Here it is in its (printed) carrying case.
https://lh3.googleusercontent.com/hdpQIhx_BE4AtImuvrk2nejy-ORy7Z2U44pNoWF0KSwT9Gxpu65t8P5WNml2HwlpO49UzutrsVS AkQ=w640
That’s running a 25mW 5.8GHz video transmitter (about the smallest/lightest/cheapest/least powerful you can find). Since space and power are limited on that platform, I decided it would be more interesting to focus on the ground station to increase range.
This is what I came up with:
https://lh3.googleusercontent.com/j-Hm57Gj56RFqQ5Aw27wmCUNLVjFpTK1RwEj1HOPDKrxeWDjtQsA SgqHlXRYVmR0TeiNzFpfSqknrw=w640
https://lh3.googleusercontent.com/IVqvQeDm0tbWntFkefRnx-TuHHk2kg6xrxtW3qEVGYvDFVhB2WCoXsJ4F8lqwZuYq2_RQUm0 SPS2yw=w400
Functionally, there are two directional antennas feeding a diversity receiver. The name of the game is signal-to-noise. The FR632 picks the stronger antenna for the video feed, however I also tap the receive signal strength indicator from each receiver and feed that to an embedded computer. The computer tries to keep the signal strength equal between the two antennas by rotating the servo in the direction of the weaker antenna.
Here’s the mount that the case/servo just rests on top of:
https://lh3.googleusercontent.com/L3fnV0Uh1K7KdFATwdufaZaE9yxE_gU_toQtbKsAinEJ4zfBmA CQTzoznrOXti9oQVqQAhKD3gHwKw=w640
I could have used a typical tArduino or whatever, but I decided for the features I wanted this would be easier with a $9 C.H.I.P. So as is, the feature list looks like:
Linux based
>200 degree horizontal sweep (360 possible)
Video RX agnostic (any diversity receiver with RSSI accessible should work)
2s to 3s operation
Extra VBAT pins
Extra 5V out
Extra video signal pins (in/out)
Wifi based controls (I control it with my phone)
Videos can be immediately downloaded after recording
Video files have actual, correct date/time
USB host port (used by Easycap)
Second OTG port which can either be host or device
Configurable 1s battery charger/backup/monitoring
About $150, including tripod, excluding battery.
The C.H.I.P. has a configurable 1s backup/battery charger so you can do stuff like this really easily:
https://lh3.googleusercontent.com/-C89drHVkxEM/WF3fsrpVCTI/AAAAAAAAJuA/eRG_izKvI9M8kpNV7HgN_pQ0RuHCBSCzQCJoC/chip_graph.png
In fact the entire modification is a pile of scripts; the main web interface and servo control are both written in Python.
The weather has really not been cooperative, but I finally got in a decent range test today. You’ll have to pardon the flying; the PIDs are still stock… from a different quadcopter. And I was getting eaten alive by mosquitos while I was trying to fly.
http://www.youtube.com/watch?v=UEYZwz0Owns
Took a few seconds to lock in, but once it did, it stayed locked in past 300M:
https://lh3.googleusercontent.com/V1ov9o2u7nBkXjvS9rIbKv2Ta1Kudt0RtEYt83noSL6xVrja5p zcgXOso0q5n81QEWA1wRWv3J8KLA
That’s not exactly recordbreaking, but I was able to put a fairly wide sweep on the area while maintaining a solid video link, and I might be nearing the limits of the micro RX’s range. (And that is a looooong walk of shame if I crash.) And this was still an urban area.
The notable thing is since this is the ground station, it’s fully transferrable to a different model. I’m looking at building a larger (faster) brushless quad now that’s still under 250g and carries a Mobius.
The analog 5.8ghz video systems in common use are really… regressive technology. However, they’re plentiful, cheap, and easier to fit onto micro models.
My recent test platform has been a brushed micro. Here it is in its (printed) carrying case.
https://lh3.googleusercontent.com/hdpQIhx_BE4AtImuvrk2nejy-ORy7Z2U44pNoWF0KSwT9Gxpu65t8P5WNml2HwlpO49UzutrsVS AkQ=w640
That’s running a 25mW 5.8GHz video transmitter (about the smallest/lightest/cheapest/least powerful you can find). Since space and power are limited on that platform, I decided it would be more interesting to focus on the ground station to increase range.
This is what I came up with:
https://lh3.googleusercontent.com/j-Hm57Gj56RFqQ5Aw27wmCUNLVjFpTK1RwEj1HOPDKrxeWDjtQsA SgqHlXRYVmR0TeiNzFpfSqknrw=w640
https://lh3.googleusercontent.com/IVqvQeDm0tbWntFkefRnx-TuHHk2kg6xrxtW3qEVGYvDFVhB2WCoXsJ4F8lqwZuYq2_RQUm0 SPS2yw=w400
Functionally, there are two directional antennas feeding a diversity receiver. The name of the game is signal-to-noise. The FR632 picks the stronger antenna for the video feed, however I also tap the receive signal strength indicator from each receiver and feed that to an embedded computer. The computer tries to keep the signal strength equal between the two antennas by rotating the servo in the direction of the weaker antenna.
Here’s the mount that the case/servo just rests on top of:
https://lh3.googleusercontent.com/L3fnV0Uh1K7KdFATwdufaZaE9yxE_gU_toQtbKsAinEJ4zfBmA CQTzoznrOXti9oQVqQAhKD3gHwKw=w640
I could have used a typical tArduino or whatever, but I decided for the features I wanted this would be easier with a $9 C.H.I.P. So as is, the feature list looks like:
Linux based
>200 degree horizontal sweep (360 possible)
Video RX agnostic (any diversity receiver with RSSI accessible should work)
2s to 3s operation
Extra VBAT pins
Extra 5V out
Extra video signal pins (in/out)
Wifi based controls (I control it with my phone)
Videos can be immediately downloaded after recording
Video files have actual, correct date/time
USB host port (used by Easycap)
Second OTG port which can either be host or device
Configurable 1s battery charger/backup/monitoring
About $150, including tripod, excluding battery.
The C.H.I.P. has a configurable 1s backup/battery charger so you can do stuff like this really easily:
https://lh3.googleusercontent.com/-C89drHVkxEM/WF3fsrpVCTI/AAAAAAAAJuA/eRG_izKvI9M8kpNV7HgN_pQ0RuHCBSCzQCJoC/chip_graph.png
In fact the entire modification is a pile of scripts; the main web interface and servo control are both written in Python.
The weather has really not been cooperative, but I finally got in a decent range test today. You’ll have to pardon the flying; the PIDs are still stock… from a different quadcopter. And I was getting eaten alive by mosquitos while I was trying to fly.
http://www.youtube.com/watch?v=UEYZwz0Owns
Took a few seconds to lock in, but once it did, it stayed locked in past 300M:
https://lh3.googleusercontent.com/V1ov9o2u7nBkXjvS9rIbKv2Ta1Kudt0RtEYt83noSL6xVrja5p zcgXOso0q5n81QEWA1wRWv3J8KLA
That’s not exactly recordbreaking, but I was able to put a fairly wide sweep on the area while maintaining a solid video link, and I might be nearing the limits of the micro RX’s range. (And that is a looooong walk of shame if I crash.) And this was still an urban area.
The notable thing is since this is the ground station, it’s fully transferrable to a different model. I’m looking at building a larger (faster) brushless quad now that’s still under 250g and carries a Mobius.