openrcforums

Here is a link to my website with some documentation of my antenna tracker project which uses FrSky telemtry to get the aircraft position data.
http://www.zoomworks.org/fpv/antenna_tracker/index.html
There is also various source code associated with the project, which I hope to eventually combine into one repository.
https://github.com/kwikius?tab=repositories

regards
Andy

Sent from my GT-I9195 using Tapatalk

I am working with Andy... 'skyscraper' on these forums.

It is building on these ideas...

http://www.zoomworks.org/fpv/antenna_tracker/index.html (Andy's project)

Just introducing new ways of getting the signal to the tracker.

With reference to
http://www.zoomworks.org/fpv/antenna_tracker/index.html
It does seem a bit slow in response (unlike your signal strength version), the plane actually flew by only once that I could see it in the video and the tracking lag was probably about 2 seconds.
It's a pity that there was not an overlay of the received video from the plane.
This might be a better method for long distance cross country tracking but not for close around a field flying.

If you're so close that the tracker lag is significant, you don't need a tracker in the first place

Correct.

All GPS trackers lag close by.

The resolution of the GPS is just not that fine!

Once you get 50 or so meters away things start to track with more accuracy as you have some difference in the GPS locations to bother moving the antenna.

Sent from my GT-I9300 using Tapatalk

The resolution of the GPS is just not that fine!

That's why I asked if the ground station was going to use a GPS as well, the resolution of the GPS can be improved by a factor of 10 or more by a method of error reduction assuming the stations a fixed and thus known co-ordinate the error it gives in relation to its actual position is compared subtracted or added to the reading from the plane giving a much more accurate positioning.

the AAT system is using a magnetometer in the set up to set 'home' location and orietation. I am not sure I know exactly how that works. But it can be affect by magnetism of the electronics in the control box and environment.

Here is my thought: if the software can store a GPS location as reference, by putting the plane/GPS right on the tracker and push a button during initial setup, the control box/software can memorize the GPS location as home location (including altitude). then once the plane start moving, all the calculation can use that 'home location' as reference point to track. During the initial setup (power up) the initial state of the tracker elevation should be 90 degree straight up (panel horizontal) and the azimuth reference should be parallel to the elevation tracking arc. This way it does not need a 2nd GPS on the tracker to set up home location and no need to know the 'compass' orientation of the set up.

That's pretty much how GPS trackers work

Sent from my GT-I9300 using Tapatalk

Bill wrote:With reference to
http://www.zoomworks.org/fpv/antenna_tracker/index.html
It does seem a bit slow in response (unlike your signal strength version), the plane actually flew by only once that I could see it in the video and the tracking lag was probably about 2 seconds.
It's a pity that there was not an overlay of the received video from the plane.
This might be a better method for long distance cross country tracking but not for close around a field flying.

Hi Bill,

There is a low pass filter on the gps position input to stop it jiggering around.

quantracker/ground/tracker/v1/telemetry.cpp ( lines 26 to 44)

There is a tracker cam vid on youtube
http://www.youtube.com/watch?v=JzvukQ9rDgk
regards
Andy

kaos wrote:the AAT system is using a magnetometer in the set up to set 'home' location and orietation. I am not sure I know exactly how that works. But it can be affect by magnetism of the electronics in the control box and environment.

Here is my thought: if the software can store a GPS location as reference, by putting the plane/GPS right on the tracker and push a button during initial setup, the control box/software can memorize the GPS location as home location (including altitude). then once the plane start moving, all the calculation can use that 'home location' as reference point to track. During the initial setup (power up) the initial state of the tracker elevation should be 90 degree straight up (panel horizontal) and the azimuth reference should be parallel to the elevation tracking arc. This way it does not need a 2nd GPS on the tracker to set up home location and no need to know the 'compass' orientation of the set up.

The above is how my V1 tracker works. It uses a quadrature encoder on the pan so, once indexed, it always knows where its pointing. You align it using a handheld compass so its pointing north and the plane (sending its telemetry data) is close by when you press the "go" button, so it sets its index and home position from that.

A magnetometer and a gps is nice and I'm planning that for the next one..

regards
Andy

why a compass or a magnetometer is even needed? If on power up, the elevation is indexed (that is a word, I was trying to say ) some how straight up and the pan is indexed at any given point. There is no need to know it is north, south... It is all relative to the GPS/altitude.
May be I am thinking using a stepper motor, using a servo may need that info?

No. Not needed.

But nice to have.



Sent from my GT-I9300 using Tapatalk

Not needed, but helpful.
When you put your tripod down it may know where it is (home position), but it doesn't know where it's facing, so when the pan servo is centered it can be pointing anywhere, it has no reference.
So either you need to put it down facing a predetermined way (in which case YOU need a compass to make sure you put it right), or it can have its own compass to find it out by itself and act accordingly automatically.

ah, you are all thinking servo centering. I am thinking a stepper motor. There is no centering for pan, once it is indexed on start up. it just changes by the GPS difference for 'relative' Azimuth and elevation in relation to 90 degree/straight up.

Rob: good move of the topic. I guess we are side tracked from the beginning of RSSI to GPS as main track.

kaos wrote:ah, you are all thinking servo centering. I am thinking a stepper motor. There is no centering for pan, once it is indexed on start up. it just changes by the GPS difference for 'relative' Azimuth and elevation in relation to 90 degree/straight up.

Rob: good move of the topic. I guess we are side tracked from the beginning of RSSI to GPS as main track.

A stepper motor would be a good alternative to a motor with an encoder certainly. The only slight advantage of the quadrature encoder is that it is more reliable. For example:
If, while the tracker is running you physically stop the pan moving on a stepper motor version ( e.g stall it), when you release it, it wont know its in the wrong place, wherease a quadrature encoder will and the electronics will now point it back to where it should be.

However if you use a magnetometer for direction this wouldnt be a problem. Seeing as how magnetometers are very cheap to buy ( e.g look for HMC5883 on ebay) that would seem the best option to find your direction in pan Then you can use the most economical option to drive it, either a motor( no quadrature encoder required) or a stepper.

For tilt you only need 90 or 180 degrees rotation. For that it I would say that RC servos are a simpler cheaper option than a stepper.

regards
Andy

skyscraper wrote: If, while the tracker is running you physically stop the pan moving on a stepper motor version ( e.g stall it), when you release it, it wont know its in the wrong place, wherease a quadrature encoder will and the electronics will now point it back to where it should be.

ha, never thought about 'what if' the motor is stopped during operation. I wonder how those 3D printing control board handle situation like this?

As for 180 degree tilt, i am thinking if I fly on top of a hill while flying the plane on the bottom of the hill or flying over a cliff while the tracker is at the edge of the cliff, then the tilt needs to go further than 180. I would like to see the tilt can point downward as far as physically can. But a servo can also achieve 270 degree fairly easy.
like flying here over the air and 3 ft near the sea level.

kaos wrote: [...]
As for 180 degree tilt, i am thinking if I fly on top of a hill while flying the plane on the bottom of the hill or flying over a cliff while the tracker is at the edge of the cliff, then the tilt needs to go further than 180. I would like to see the tilt can point downward as far as physically can. But a servo can also achieve 270 degree fairly easy.
like flying here over the air and 3 ft near the sea level.

That is true. I hadnt thought about that. Theoretically your tilt should be from straight up to straight down. you could be tracking from a hot air balloon or an aeroplane! Certainly the ability to point downwards is a desirable feature, especially for slope soaring etcetera.

regards
Andy

My skylark can point downward.

Actually... Their approach to the design of the pan tilt is superb. So simple and effective.



Sent from my GT-I9300 using Tapatalk

what is that skylark? Isn't that a plane? sounds interesting.

http://www.skylarkfpv.com/store/index.p ... #container

Sent from my GT-I9300 using Tapatalk

wow, it almost identical with the foxtechfpv AAT system. It does not seem to have the slip ring which is better.
Hmm, it has the BT, the V3 AAT from foxtechfpv removed the BT and using usb.
But both uses servos and 90 degree tilt.

Based on this thread title, are you thinking using Frskky GPS info for tracking? Could you elaborate on what you are thinking as for how you get the FRysky info to the tracker control or you plan to do this all in the Tx?
At the Zoomworks site it mentioned several options as how to get GPS data, ie, frysky, FPV audio ch......
I think the most independent/ versatile one would be using audio ch like AAT system.
but the easiest one, as most people here use Frysky system, would be using the frysky telemetry with a single GPS attached to a arduno and let the Tx do the calculation and control the tracker servo/motor?? Is that what you are thinking?
for what I can tell Skyscraper's current system is using frysky telemetry but using a separate board to control the tracker.

Skyscraper:
in regard to mavlink to frysky conversion using Arduno. Does that only convert Ardupilot mavlink data? or multiwii will work as well?
also, does that only convert GPS data or all the telemetry data from the FCB? I can't figure out from the linked source site.
IF that works for multiwii board and converts all the telemetry data (GPS,volt,amp, accel, temp ....) I may just go ahead and make one just to get my multiwii FCB telemetry data to my Frysky telemetry and try.

Skyscrapers current system relies on a link from the serial port on the djt into the tracker system.


Essentially filters out the GPS signal and uses this to move the tracker.

Downside of this is the cable.


One potential solution is a relay system. Frisky back module and relay piggy backed to each other using cppm...this could be installed on the tracker. Result is you bind your tracker to the plane, and your radio to the tracker.

Our hope is to do something a bit different here. Essentially some 'magic' to make it allot simpler and allot more resilient.

Very early days still - so not going to elaborate more as it may potentially not even work!

Rob

Sent from my GT-I9300 using Tapatalk

Rob Thomson wrote: One potential solution is a relay system. Frisky back module and relay piggy backed to each other using cppm...this could be installed on the tracker. Result is you bind your tracker to the plane, and your radio to the tracker.

I thought that is the simplest way to do it. Now I am really looking forward to your 'magic' to make it simpler than that.

Just one question. Using the audio ch to transmit GPS back to tracker you can use a higher power vTx and higher gain ant to boost range. But I remember Frysky telemetry send back power is limited, beside using a higher gain ant you can't boost the power. It seems to me using a Vtx would potentially have longer range. (I can't remember what is the power for Frysky telemetry return signal.) Correct??

No idea what the return power is.

Regardless... Andy's system can send back on both frisky and audio

Sent from my GT-I9300 using Tapatalk

Range is the same for control and telemetry with FrSky.

If we add a, say 13 dB, antenna on Tx, will the transmission and receiving range still the same?
if we use a wifi booster on Tx, that seems definitely will result in tx range > rx range?

Do this....



Sent from my GT-I9300 using Tapatalk

that is something similar of what I have in mind, only with helicals.

Have a Skylark, making 2 Rssi, now a frysky tracker
Rob, you are crazier than I thought.

Did you add a diversity on that Skylark? I also see a fat shark clover/planar skiew on it.