Telemetry / Antenna Tracking Board with Ethernet Capability

[Arthur Benemann]

I'm developing a Telemetry board, to be used in conjunction  with Mission Planner or my android GCS app (DroidPlanner, but i think Andropilot will run without a problem).

 It's just a HopeRF's HM-TRP radio (the same module used in 3D robotics telemetry modules), an uC, an Ethernet controller and servo outputs (for an antenna tracker). The Ethernet connectivity option is great since multiple devices can be used to control the drone, like an android device and a PC. The whole system can communicate trough WiFi to multiple devices if you attach a cheap WiFi router to this board. Here is a 3D render of the board:

I think it would make a great addiction to the USB and air modules already avaliabe. I think a BOM cost of $10 is possible (excluding the RF module and SMA connector, which i don't know the price). The features I'm thinking to add are:

• Serial To TCP bridge
• Serial To UDP bridge
• Antenna Tracking - Via two servo outputs
• Battery Voltage Sensing - Since this will probably be connected to a Lead Acid battery for field operations.

I have attached a PDF with a sketch of the schematic. The board files and source code can be downloaded from this git repository. I have a working prototype working, capable of handling multiple TCP connections. I could post a video if there is interest.

Any comments are greatly appreciated.

PS:  

1. DEV call is tonight right? Could I join, if so where can I get the required password?

2. What is your policy about image and file attachments on this list?

[Craig Elder]

Thanks Arthur.  This looks very interesting.

I'm just answering your other email so I will give you the details on the call in it.

--
You received this message because you are subscribed to the Google Groups "drones-discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to drones-discus...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.
 
 

[Kevin Hester]

Looks awesome Arthur!  Are you going to do a small build, if so could I send you some money to get one?  I'll happily add andropilot support.

On Mon, Mar 4, 2013 at 10:55 AM, Arthur Benemann <arthur....@gmail.com> wrote:

[Arthur Benemann]

Looks awesome Arthur!  Are you going to do a small build, if so could I send you some money to get one?

I'm hopping to have 3DR support to build some prototypes, since to have the size I showed on the 3D model the board is must be very miniaturized (QFN, and 0402 components).

 

 I'll happily add andropilot support.

I think it will be supported out of the box, since the board is just a TCP server. I have tried to configure Andropilot to a TCP connection but have failed.

[Michael Oborne]

Hi Arthur

 

Just a couple of questions

how does the user set an ip on it?

does is support dhcp?

i believe many users are not advanced enough to go changing these settings, both hardware and pc side.

 

Michael

[Arthur Benemann]

Just a couple of questions

Befor I answer: I have not develop the TCP/IP stack, as that would give me a lot of trouble. I'm using Microchip TCP/IP stack, and that is really handy. The only problem is that the stack can't be distributed with the code (It has to be downloaded from microchip site), but that doesn't mean it can't be used. 

how does the user set an ip on it?

It has a DHCP client, so it ask your gateway for an IP. The MAC is currently set on code, but in a production run a pre-programed I2C eeprom will hold the MAC. If it is connected directly to a PC it starts a DHCP server, and give an ip to the PC.

It respond to ping queries, and has support for NETBIOS names so connecting to the board is really easy.

The stack also has support for HTTP ( as I demonstrated with this board), but I leave that one off to shrink the code to the maximum so a cheaper IC could be used. I thinks is better to leave the UI to the client side.

does is support dhcp?

i believe many users are not advanced enough to go changing these settings, both hardware and pc side.

Sure, things must be as simple as possible. The DHCP question is answered above. 

On the command line it's just "telnet ethernetgcs 9761", I think it's easier than messing with COM port numbers and baud-rates. 

[Arthur Benemann]

The only interaction to the MAVlink stream the board will be handling is for the antenna tracker.

I will just intercept GPS packets and use their coordinates to move the antenna, and also receive packets from the GCS to set the antenna position and orientation. Most of it is written on the ardustation code.

[Arthur Benemann]

Here is a picture of the prototype (no xbee or RF module, just a wired connection of the UARTs):

And a video of it working with MissionPlanner and DroidPlanner at the same time.

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

[Arthur Benemann]

I have finished the layout of the Ethernet GCS board. The files are on GitHub.

One nice addition I made is a port to connect a GPS receiver. If Tridge's endeavor with Differential GPS works, this can now be used as the base station. Having Ethernet means long cables (or Wifi) for connection, and the ability to easy setup multiple receivers.

Here is a 3D model of the current version:

[Robert Lefebvre]

Wow, pretty cool.  The potential for a completely wireless antenna tracker gets closer.  Just go stick it out in the field somewhere, sit at a comfortable table in the shade with your laptop...

[Robert Lefebvre]

Pete, is that program working, or you would just like it to be working?

On 21 March 2013 00:31, Pete Birley <peteb...@gmail.com> wrote:

Arthur, this is great - is there any chance you could upload a pdf of the layout - I dont have eagle at the moment.

I'll take this opportunity to make a shout for ardu-imu (or similar) support for compass / attitude compensation to provide a real 'plug and go' solution again :)

[Arthur Benemann]

 Pete Birley wrote:

Arthur, this is great - is there any chance you could upload a pdf of the layout - I dont have eagle at the moment.

Attached the schematic and top layer of the board (in 1:1 scale), for the other layers it's best to have the right tool (Free version at this link).

I'll take this opportunity to make a shout for ardu-imu (or similar) support for compass / attitude compensation to provide a real 'plug and go' solution again :)

I tough about that (including a HMC5883 magnetometer) to provide orientation control, but that would result in a bigger board and more cost to the BOM (about $2 at digikey).

 

[Robert Lefebvre]

I think I've missed the plan here a little bit.  But if at the end of the day, I could end up with a completely wireless antenna tracker, that would be awesome.  I have the tracker mostly built.  It has 3 antenna, 900MHz patch for telemetry, 2.4GHz patch for control, and 5.8GHz CP helical for video.  I was hoping to be able to set it up as a totally wireless relay station. I could put the antenna out in a good spot, while I sit back at a table in the shade.  Or in a vehicle, etc.  I know I can set up a 5.8GHz video relay.  And I think I can set up a 2.4GHz relay system.  So I just need the telemetry, and servo drive to get figured out.

On 21 March 2013 17:15, Pete Birley <peteb...@gmail.com> wrote:

The basic update of the Ardutracker code is working fully - I've put a cleaned up version on dropbox that is good to go (57600 baud, Pan and Tilt on RC 8 & 7 respectively).

I'm looking get the other version with full (dynamic?) tracking, finished, tested and documented within the next 3 weeks. I've got a hellish workload on at the moment, and I currently have to do this in my spare time, though after then will have full day a week on this stuff :)

[Arthur Benemann]

robert.lefebvre wrote:

I think I've missed the plan here a little bit.  But if at the end of the day, I could end up with a completely wireless antenna tracker, that would be awesome.

That's what I'm aiming for with the EthernetGCS. Anyone has tried the Ubiquiti radio modules, to keep a WiFi link to the drone? That would allow for video streaming and telemetry with only one link.

And since you mentioned:

 I could put the antenna out in a good spot ....  or in a vehicle, etc

If the antenna tracker has a GPS and a mag. it possibly could follow the target (the drone) while the vehicle is in movement.

[Robert Lefebvre]

That's sort of along the lines of what I'm thinking.  Put some kind of IMU, a GPS and mag on the tracker, and then just walk it out and set it down with no calibration required.  Is that what you're working towards?

I'm not sure the servo system could handle the wind load of being vehicle mounted. That would be pretty strong.

--

[Arthur Benemann]

Pete Birley wrote:

I'm installing eagle now, in the past I've always used NI Circuit Design Suite but it's horrifically expensive and overkill for most apps.

Open source hardware projects are mostly using eagle since it has that free version. (I home ardupilot stays that way and don't switch to Altium)

 

Though it would be great to have orientation control onboard for cost and simplicity; I think it should ideally be a separate module - this would allow much more flexibility during installs where magnetic interference and weird tracker designs may cause issues (e.g. mounted on the roof of a car)

That's a great idea, all I need to do is put an external connector on the I2C bus and the magnetometer board stays optional (like on APM1). 3DR even has the magnetometer board in stock.

 

Would it be possible to get my hands on one of these boards when you get a run made?

I'm working with Craig from 3DR to manufacture some prototypes, since the board is quite difficult to make (four layers,QFN and 0402 parts). I have a prototype working (picture in a post above).

[Pete Birley]

Cheers Arthur

That sounds like a great addition: forgive me, but does the magnetometer also have the ability to find the attitude of the tracker body? This is the solution I'm really interested in having: it would allow the tracker to be used in all sorts of additional environments: eg. boats, off-road vehicles, backpack mounted systems, lazy setups etc.

[Arthur Benemann]

That's sort of along the lines of what I'm thinking.  Put some kind of IMU, a GPS and mag on the tracker, and then just walk it out and set it down with no calibration required.  Is that what you're working towards?

Yes, I have some steps to achieve this goal:

1. Code the TCP-Serial Bridge - Done
2. Test the bridge with a 3DR radio on the link - Just waiting for my order arrive
3. Build the hardware for an antenna tracker.
4. Implement a poor man tracker (no GPS, no Mag) - Will work with Mission Planner (or the GCS app I'm developing (DroidPlanner)). The tracker coordinates is set in MP, or with the help of an Android device (which has GPS, and sometimes a mag).
5. Improve the tracker by using a GPS and a magnetometer - This would require no user intervention, except on the first setup to calibrate the range and trim the pan/tilt axis.

I'm not sure the servo system could handle the wind load of being vehicle mounted. That would be pretty strong.

Since I'm a E.E. I normally forget about the hard part, the hardware. But that is just a question of getting a big enough servo.

 

[Pete Birley]

I'm not sure the servo system could handle the wind load of being vehicle mounted. That would be pretty strong.

Since I'm a E.E. I normally forget about the hard part, the hardware. But that is just a question of getting a big enough servo.

 

The servos I am using are pretty big: Pan (with a 5:1 gearbox) and Tilt, wind loading at high speeds would be massive, but easily mitigated by adding a radome if needed.

[Arthur Benemann]

Pete Birley wrote:

That sounds like a great addition: forgive me, but does the magnetometer also have the ability to find the attitude of the tracker body? This is the solution I'm really interested in having: it would allow the tracker to be used in all sorts of additional environments: eg. boats, off-road vehicles, backpack mounted systems, lazy setups etc.

I thought only in a antenna mounted with pitch = 0º. To measure rest of the attitude information a acceleration would be needed (no gyros if the antenna is static). Interfacing to ArduIMU or a MPU9150 would be as hard as to the magnetometer, it just adds cost since those parts are more expensive.

[Pete Birley]

Going with the ArduIMU would be easiest initially I think, though it is a bit more expensive, as it would require only a few lines to get a full ARHS from it (including gps). I can't see the picture you've posted of the breadboard setup you have for some reason, but I'll try and get a replica knocked up in a week or so to play around with. 

[Robert Lefebvre]

I've got the SPG785 with 5:1 for pan:

http://www.servocity.com/html/spg785a_top_mount.html

And this on tilt:

http://www.hobbyking.com/hobbyking/store/__27213__TGY_1270HV_Metal_gear_Digital_Servo_w_Heat_Sink_40kg_18sec_170g.html

That servo is a monster.  

Still, don't think it would work on in a wind.  My pan tilt shaft is 24" long, and the two patch antenna are mounted on the ends.  So the pan moment is huge.

I think your setup would work for a single patch, however.  At least at reasonable speeds.  But for something like mine, it would have to be much more robust.

--

[Pete Birley]

True - my setup consists of a 5.8Ghz Crosshair and 433Mhz Moxon, they're tiny in comparison to your rig!

[Arthur Benemann]

Going with the ArduIMU would be easiest initially I think, though it is a bit more expensive, as it would require only a few lines to get a full ARHS from it (including gps). I can't see the picture you've posted of the breadboard setup you have for some reason,  

Here are some direct links to the pictures:

Prototype with wired link to APM.

Prototype youtube video

EthernetGCS 3D model

but I'll try and get a replica knocked up in a week or so to play around with.

Both boards I used are custom designed by me, I don't see how you can make a replica? If you want I can send you the board files (one of the boards can be found here).

[Arthur Benemann]

Pete and I talked off the group to stop  cluttering this list. The conclusion that we came to is that any of the external boards would work (HMC5883,MPU9150,ArduIMU,GPS) since there is power,ground, and two generic I/O pins on the GPS port of the board. So no modification is needed, is just a matter of writing the code.