I'm thrilled to report that my radio kit arrived and it works well.  Andreas is likewise happy with his kit.

Writeup here:

http://eastbay-rc.blogspot.com/2012/04/review-3dr-900mhz-radio.html

Fixing the air unit network id:

http://eastbay-rc.blogspot.com/2012/04/setting-network-address-on-3...

Thanks 3DR, Tridge, and everybody else that worked in producing this unit!

Views: 3804

Tags: 3dr-radio

Comment by Jack Crossfire on April 21, 2012 at 6:44pm

It's a closed source Hope electronics module soldered to a microcontroller which presumably translates MAVLink.  You'd think 3d robotics could just design the entire radio.

Comment by Greg Fletcher on April 21, 2012 at 8:07pm

Error in your review, wt. is 17.5g not 175g.


Moderator
Comment by Mark Harrison on April 21, 2012 at 9:46pm

Thanks Greg, fixed!

Jack, all in good time, I'm sure this won't be the last radio system offered.  My perspective on this one: worked painlessly, half the price. Onward and upward!


Developer
Comment by Andrew Tridgell on April 21, 2012 at 9:48pm

Hi Jack,

It's not closed source - the full schematics of the design are available from Si (it's just the Si1000 reference design), plus all the source code is open. This is the most open telemetry radio I'm aware of.

Cheers, Tridge


Developer
Comment by Andrew Tridgell on April 21, 2012 at 9:51pm

Thanks for the review Mark! I was really looking forward to the first reviews coming out so it's nice to get a positive one first up!

Cheers, Tridge

Comment by Ayush Gupta on April 21, 2012 at 10:04pm

Hi Mark, Could you please do some Range Test and Let the community know what kind of Ground range and In air range one can expect out of this Beauty.


Developer
Comment by Andreas M. Antonopoulos on April 21, 2012 at 10:16pm

I also flew my 3DR radios a few days ago. Very happy with the purchase, much smaller and lighter than Xbee, easy to use. 

Tip: Even though you *can* plug the groundside radio directly into your laptop's USB port, DON'T. Use the provided USB extended cable and maybe tape the radio to the back of the screen. Within minutes of plugging directly into the laptop, I managed to bend the USB jack and almost snapped it off. It's not designed to have any resistance to bending, there's no backstop or bottom plate. So the USB jack will simply pivot on the solder points, if you're lucky, snap them off if you're not. 

Very nice radio solution: half the price, less weight, less soldering (one header). Can someone from the team comment on the antennas? I noticed the ones provided for the 3DR radios are shorter than the tapered antenna that comes with the XBee groundside radio. How do they compare from a signal strength perspective? I don't have enough space to fly far enough to test.  I will be doing a drive test this week (that's where I put the quad in the back window of the car and drive around the lake with the GCS at home logging). 


Developer
Comment by Andrew Tridgell on April 21, 2012 at 10:59pm

Hi Andreas,

The antennas are reasonable, but may benefit from having a ground plane. It would be interesting to do an experiment watching the reported RSSI value with and without an aluminium foil ground plane.

Regarding range testing, you can extrapolate the range quite well from the RADIO packets in a tlog. What you need to look for is the difference between the RADIO.rssi and RADIO.noise levels at both ends of the link. Convert that difference to dBm and use the rule that each 6dBm doubles the range. The formula for RSSI to dBm is on the 3DR wiki page.

For example, here is a graph from a flight I did today:

The 'RADIO.rssi' and 'RADIO.noise' graphs show that at a distance of 500m the GCS was getting a RSSI of 66 (which is -92dBm) and a noise level of around 27 (-112dBm). That means we had 20dB margin. If we allow for 6dB of fade margin then that gives us 12dB spare. That corresponds with 4x the range, so it would have gone to 2 km very easily (the 6dB fade margin is generous).

The radio in the plane is a different story for this flight. It was getting a similar signal level (about -92dBm), but the noise level was very high, sitting around an RSSI level of 60, which is -95dBm. That is only 3dB below the signal. It did keep a solid link, but it was quite marginal. If we had flown any further it would have lost the link.

So in this case what we need to do is work out why this plane (a 50cc petrol plane) had such a lot of noise in the plane. The same radios when I flew them yesterday in my SkyFun gave much lower noise levels, and extrapolating that result implies a range of about 5km or so for those flights.

So the range you get very much depends on the noise levels you have. One of the things we aimed for in the design of the radios was to make it easy to log the signal and noise levels so you can work out why a link is poor. With most radios its just guesswork, but with logs like the one graphed above its easy to see what needs fixing.

Cheers, Tridge


Developer
Comment by Michael Oborne on April 22, 2012 at 12:33am

just a fyi,

in the planner under "Status" and "tunning" you can see all the values tridge included in his graph live.

Comment by Andrew Dunlop on April 22, 2012 at 2:34am

Hi Tridge,

Great job on the radios.

There are a number of reasons why the receive performance on a small mobile platfrom can never be as good as on the ground, but it is worth exploring them to try and minimise the difference.  You have noted an increase in the noise floor, so this is worth considering first.

Number one, there is the increased proximity of other RF emitters, which can be physically separated on the ground.  Although out of band, if strong enough or close enough out of band emitters can penetrate the input filters of the receiver in question and degrade its sensitivity.  Image frequencies are particularly troublesome.  Separation is difficult to achieve in an aircraft, but every little helps.


What sort of ignition does your petrol engine have?  Electronic ignitions are notoriously noisy electrically.  Is this being radiated or conducted into the radio?

I recall that you said adding an LNA inproved the receive sensitivity of your base radio.  This suggests that the front end of the radio is not quite optimal (not intended as a slur on the design - there may be nothing you can do about it with such a broadband front-end).  Again there are a number of reasons this could be the case.  Impedance match between antenna and input stage is one possibility.  If the LNA does a better job of matching to the antenna at the frequency of interest than the radio does, then this will improve the receive perfromance.  Noise figure itself is another possibility.  The LNA is likely to have a better noise figure than the radio, and together with a bit of gain this will improve the receiver sensitivity, dB for dB (almost).  An LNA on the aircraft might improve the receiver performance, but it might not be practical.

On the subject of LNAs, they should be located as close as possible to the antenna.  That is why they often accept power up the output coax - to simplify installation if they end up being mounted up a pole.  Every dB lost before the first amplifier worsens the receiver sensitivity, dB for dB.

There are no doubt other factors that I have missed.


Cheers,
Andrew.

Comment

You need to be a member of DIY Drones to add comments!

Join DIY Drones

© 2014   Created by Chris Anderson.   Powered by

Badges  |  Report an Issue  |  Terms of Service