Open RSSI Antenna Tracker

[Rob Thomson]

I always use a standalone balance lead tester

But would be useful

Sent from my GT-I9300 using Tapatalk

[Rob Thomson]

Test version running...

http://www.youtube.com/watch?v=TrSRZiL6sDk

[hageha]

Rob,
you have the RSSI outputs of the receiver connected directly to the Arduino?
Or with resistors and possibly with capacitors blocked?

[Rob Thomson]

I simply go direct.

I have not found a need for a buffer on my receivers.

Apparently this works well too: (the buffer circuit)

http://www.michael-heck.net/index.php/f ... nentracker

Rob

Sent from my GT-I9300 using Tapatalk

[kaos]

I think in theory the RSSi tracker concept surely works.
At 1:09 when the tracer snap around it seems the center helical is not pointing to where you were? or may be you are so close, the cloverleaf was taking over?
I guess if you use stepper motor that would eliminate the 360 degree turn around?

[jhsa]

Cables wouldn't.. that is the main reason why you can't have it rotating 360º

Just my 2c

João

[kaos]

When the battery and all set up is on the pan tray, there won't be cable issue. the tilt will only go 180 to 270 degree (some say 120 degree) anyway so they won't tangle.

[jhsa]

And how do you connect your goggles or recorder to the pan tray? Wirelessly or do you sit on it and rotate along?

Sent from my GT-I9195 using Tapatalk

[kaos]

of course wireless, as Rob is doing now. there is a relay. I won't even think about wires from Tx/goggle to a tracker. If there is wire between me and the tracker that won't be my tracker. Throw the tracker out on the ground and start flying.
I am trying to learn/study from the one Rob is making and add radio directional antenna on it as well.

[jhsa]

So you're saying that you will have the tracker receiving from your aircraft with all those antennas, and then transmitting to your video display/recorder. That sounds a bit like too many tx/to all together.

Sent from my GT-I9195 using Tapatalk

[kaos]

No one says it is going to be easy.
I am also looking at how this is going: http://openrcforums.com/forum/viewtopic ... ian#p54315
this will make the tracker with less hard ware. How good it is? yet to see how the project is moving along.
I think the GPS system would be affected more by weather while the RSSI is a direct radio signal.
but with more accurate GPS, something like this, may be it does not matter:
http://dx.com/p/u-blox-lea-6h-high-accu ... fpv-261328

[Rob Thomson]

The snap back is because the servos in this unit do not do more than 360 degrees.

So you have to snap!

Now... Try remember something.

The beam width gets wider and wider. If you are 200 meter out, you already have a beam about 1/2 a mile wide. So even if the snap is not spot on, it will still be OK.

It snaps, then relocks. All within 2 seconds.

Part of the reason for the snap taking a while to occur - you will note the overshoot in then video is to allow me to be sure that you ate actually continuing past the snap point. Essentially, you cannot snap if you are sitting on the 360 of 0 mark, as they are the same place! You would induce a constant snap-snap-snap etc. So you leave a large margin, and don't snap back to the exact opposite

I tuned up the code yesterday. So no more random funnies, and all seems to be smooth. Test flights today.

Now... What I do to get a video feed to the goggles is to use a separate tripod located 2m from the tracker. I simply rebroadcast on a different frequency. I intentionally separate it as I found lots of issues when tunning the transmitter in the tracker. The power supply is just too dirty, and the antennas interfere with one another. Separated them... And works perfectly.

Sent from my GT-I9300 using Tapatalk

[kaos]

The snap back is because the servos in this unit do not do more than 360 degrees.
So you have to snap!

Part of the reason for the snap taking a while to occur - you will note the overshoot in then video is to allow me to be sure that you ate actually continuing past the snap point. Essentially, you cannot snap if you are sitting on the 360 of 0 mark, as they are the same place! You would induce a constant snap-snap-snap etc. So you leave a large margin, and don't snap back to the exact opposite

I guess this is where a stepper motor may shine.

Test flights today.

video, video

Now... What I do to get a video feed to the goggles is to use a separate tripod located 2m from the tracker. I simply rebroadcast on a different frequency. I intentionally separate it as I found lots of issues when tunning the transmitter in the tracker. The power supply is just too dirty, and the antennas interfere with one another. Separated them... And works perfectly.

This is where a GPS system will shine.

great work!

[Rob Thomson]

I guess this is where a stepper motor may shine.

Nope. My previous version had a 360 continuous servo + slip ring.

This version uses a snap back because the chassis I used could not accommodate a slip ring.

Regardless.. no actually issue. Snap back is a work-around, used by pretty much all commercial trackers, and causes no issue bar a 1 second video break. In practice, you position your antenna / tripod so the 'snap' position is not in the direction you intend to fly. 99% you never actually hit it!

This is where a GPS system will shine.

Not really. It will make little difference. A GPS unit still has the same functional mechanics and will introduce the same dirty power.

This can be fixed with LC filters, but it is cleaner IMHO to use a separate power supply for the relay.

The 'relay' in my case only needs separation because I am relaying out on the same band that I receive on. If I did video on 900, and relay on 5.8, I would not have this issue. But in the UK, 5.8 is the only frequency that is 'legal' and 'safe' to use for fpv. So best option is for me to separate them

[Flaps 30]

The snap back is because the servos in this unit do not do more than 360 degrees.

I don't understand why you should need servos that have to do more than 180 Degrees. The last time I looked, it was possible to do pan and tilt with two 180 degree servos to cover the whole sky with large satellite aerial arrays. No snap back on them.

[jhsa]

Last time I saw Michael heck's device at the field I think it wouldn't rotate 360 degrees but the tilt mechanism would go over to the other side when he flew over to the back. All you need is two 180 degree servos I would say.

João

Sent from my GT-I9195 using Tapatalk

[Rob Thomson]

Totally correct.

My servos do 180 degrees. But there is a 'gear' which ups it to 360 on the tracker unit.

[Rob Thomson]

First flight test...

http://www.youtube.com/watch?v=xcrXeGhW9eg

With a present at the end!

Tracker followed fine. Certainly within tolerance.

I need to tune the gain settings a bit - as it is still wobbling a little to much. But nothing major.

[tilmanb]

Nice finale

[jhsa]

Rob, you missed the tracker. Shame on you. must aim better next time

[Rob Thomson]

Stupid tree had a branch I did not see!



Done some tuning since the video.

It is reacting faster, and wobbling less now.



Sent from my GT-I9300 using Tapatalk

[Bill]

Stupid tree had a branch I did not see!

Be honest you didn't even see the tree....
As I predicted the prediction algorithm is over predicting

[kaos]

At 0:19 what's the cause of the tracker turn exactly the opposite to the plane? signal reflection from the back woods?

When you trying to fly and look at the tracking that's what gives you the finale. your finger moves with tracker instead of your eye.
I think you need that little girl to fly and you look at the tracker.

[Rob Thomson]

Just too close to the tracker.

I have not added code yet to detect that you are so close, no point moving!



Sent from my GT-I9300 using Tapatalk

[Flaps 30]

My servos do 180 degrees. But there is a 'gear' which ups it to 360 on the tracker unit.

It is the 360 Degrees that makes no sense, as it isn't required. The tilt mechanism deals with reciprocals.

Done some tuning since the video. It is reacting faster, and wobbling less now.

Yes I saw the wobbling on the heading going on. It reminded me so much of the Automatic Scan Tracking (AST) on the Ampex VPR2's and 3's that sometimes does that when the gain is set too high on the head drive piezo assembly causing the head to overshoot and wobble around. Showmaster should know what I'm on about if anyone else is lost on this bit.

[kaos]

It is the 360 Degrees that makes no sense, as it isn't required. The tilt mechanism deals with reciprocals.

I lost you on that. if you fly in circles you need 360 degree?? Isn't tilt only track the signal for vertical angle while the pan tracks the horizontal signal strength?

[Rob Thomson]

You can technically tilt right over to do the other side.

But prefer to do it as I have

Sent from my GT-I9300 using Tapatalk

[Kilrah]

But tilt can get past vertical towards the back, so you can do 360° azimuth coverage with only 180° pan and 180° tilt. Snaps would be quicker too.

[jhsa]

That's what I was trying to say earlier.. But I'm not into FPV so my opinion mtght not be the best.

[Flaps 30]

It's okay JHSA. Your are a thumb flier so we can let you off this time.

I lost you on that. if you fly in circles you need 360 degree?? Isn't tilt only track the signal for vertical angle while the pan tracks the horizontal signal strength?

Draw a circle with the aerial (antenna) in the centre. The up down (Y axis) is dealt with by the tilt mechanism. The left right (x axis) is looked after by the pan mechanism. From that you can work out that all movements, regardless of the size of the circle can be covered by using no more than 180 degrees on pan or tilt, regardless of where that circle is with respect to the aerial.