Phantom 3 Channel <-> Frequency Table

[Shammyh]

Hello all!

https://docs.google.com/spreadsheets/d/1WuKQD3FvbtFnJ6wkFu9DLxhdxHCSsnZPi90CwTGj3ks/edit?usp=sharing

I made a table of observed rf readings via a RF Explorer 2.4 GHz meter and put them in a Google Sheet.

Publicly editable should anyone wish to correct a mistake or add further useful information.

Spoiler: click for BIG rendered image of the table

Few observations:

1) Video is transmitted via ~10 Mhz wide channels from 2280 MHz to 2600 MHz. (600-280)/10 = 32 channels.

2) Control surface (analog sticks, buttons, etc) is ~1-2 MHz wide, and hops in the 2400 MHz 2490 MHz range only. The user-editable "controller channel" does not affect this signal.

3) Phantom channels 13-20 (as well as the control signal hopping) overlap WiFi 802.11b/g/n 2.4 GHz channels 1-14.

I have some screenshots from the spectrum analyzer, but for another post.

Would also like to take a closer look at the control signal in the future. Is it similar to DEEST or FASST? Don't know.

Enjoy!

- Sam

Edit 1: Updated with correct center frequencies and ISM start point.

 

[2nd2non]

Great info. Now we know why distance isn't guaranteed to increase with the 32 channel hack- control signal remains within 2400-2490 range

 

[Dome]

The radio system uses two different modulation for video and controls. Controls seems to use FHSS (Bluetooth like) and video is DSSS-like (similar to old wifi type B). When I made some measurement on an Agilent spectrum analyzer the video channel were no more than 4mhz bw.
To use full spread band for controls you need a config file on aircraft.

 

[Recce2]

Thanks for the great info.

 

[N017RW]

U.S. 2.4GHz ISM is 2400-2500MHz
Your graph shows it starting at 2390MHz

According to the FCC filing, DJI P3 LB down-link center of channel for channels 1 - 8 begin at 2406.5MHz and increase 10MHz with each channel ending at 2476.5MHz.

 

[envisionabove]

Just Because
https://en.wikipedia.org/wiki/ISM_band

 

[ianwood]

Great job. No wonder they had to reduce it to 8 channels! They're way outside the ISM band.

As for the control signal, I'd bet it is FASST (aka DEST).

I would add that the data downlink (e.g. telemetry) is embedded in the video stream most likely in the audio channels.

 

[Shammyh]

U.S. 2.4GHz ISM is 2400-2500MHz
Your graph shows it starting at 2390MHz

According to the FCC filing, DJI P3 LB down-link center of channel for channels 1 - 8 begin at 2406.5MHz and increase 10MHz with each channel ending at 2476.5MHz.

Good catch on the ISM. Just a typo in the table. Will update it.

As to the FCC filing, I took a close look at that before making the table. They (independent lab in China) only tested the Phantom 3 from channel 13 to 20. However, if you look at their graphs, you'll see the channel bandwidth does indeed match my observations.

There's no question that with the full range of channels enabled, that the Phantom is going well outside the ISM range. My RF Explorer can also generate "noise" at specific frequencies, and doing so at say 2355 MHz or 2545 MHz corresponds with a spike in the corresponding channel of the graph in the DJI app. So unless my spectrum analyzer is off by several hundred MHz, the observed portion of my table is indeed accurate. Can't say for sure on the very high and very low channels though, as they're outside my observable range (2350-2550).

That and if you look at some of the specs for Wimax, which I suspect the video signal is similar too, a 10 mhz channel jives well with expected data bandwidth.

In some future testing I might see how the "bandwidth slider" aka video quality slider impacts things as well. However, I assume it's like Wi-Fi where "throughput" at a given channel is determined not by channel width, but by modulation.

 

[Shammyh]

Great job. No wonder they had to reduce it to 8 channels! They're way outside the ISM band.

As for the control signal, I'd bet it is FASST (aka DEST).

I would add that the data downlink (e.g. telemetry) is embedded in the video stream most likely in the audio channels.

My thinking as well. Need to see if I can find some images/data on FASST/DEST to confirm.

On the plus side, if that is the case, a little bit of boosting of the 2.4 GHz signal plus moving the video channels to just above or below the ISM bands (especially in the nebulous 2360-2400 MHz range) might be the sweet spot for maximum control and video range.

It looked to me like the "hopping" control signal was transmit-only from the controller, no return signal, hence a good candidate for boosting.

I wouldn't boost with the very high or very low video channels though, as someone (or someones) will no doubt get annoyed.

 

[N017RW]

As to the FCC filing, I took a close look at that before making the table. They (independent lab in China) only tested the Phantom 3 from channel 13 to 20. However, if you look at their graphs, you'll see the channel bandwidth does indeed match my observations.

The center freqs. are still wrong.

 

[43k]

Or this one:

From the thread here:
http://www.phantompilots.com/threads/more-channels-hack.50034/page-23#post-479356

 

[N017RW]

Not arguing but simply repeating the FCC filing...
DJI #s 13-20 also have a .5 Mhz. i.e. #13 = 2406.5 MHz

 

[Dome]

I don't bet on FASST (it is not well accepted over S-FHSS, by chip maker due to restriction on affective power dbm usable). You need a waterfall spectrum analyzer to see the eventual difference. Anyway doesn't change much between the two. For the control system.

Which fcc doc are you referring about aircraft ?

 

[Shammyh]

Not arguing but simply repeating the FCC filing...
DJI #s 13-20 also have a .5 Mhz. i.e. #13 = 2406.5 MHz

Correct. Looks like I'm off by about 1.5 MHz. Not exactly what I observed, but with the limited resolution of my spectrum analyzer, I'm not totally surprised.

I can only measure to +/- 1 Mhz, but again, based on other people's data, and the FCC filing, I'll update my table accordingly.

I did not see any spacing between channels, but again, could be limit of the the resolution of my observations.

Also didn't see that BudWalker had already done some tests with a Bullet M2. Glad to see that by and large (minus the 1.5 MHz offset) our results agree. Nothing like independent discovery for the win!

Going to do a waterfall and some more tests on the "control signal" when I have the time.

Edit: Table and image updated.

Thanks for all the feedback as well from everyone!

I'm quite proficient in some aspects of technology, but I'm not terribly well educated on the RF world (and this is my first round with a frequency analyzer) so all the feedback is invaluable to get the most accurate picture possible.

 

[N017RW]

Correct. Looks like I'm off by about 1.5 MHz. Not exactly what I observed, but with the limited resolution of my spectrum analyzer, I'm not totally surprised.

I can only measure to +/- 1 Mhz, but again, based on other people's data, and the FCC filing, I'll update my table accordingly.

I was going to inquire about your test set-up but now you've cleared that up.

Thanks for the effort.

 

[ianwood]

It looked to me like the "hopping" control signal was transmit-only from the controller, no return signal, hence a good candidate for boosting.

That would be consistent with FASST.

 

[Shammyh]

I was going to inquire about your test set-up but now you've cleared that up.

Thanks for the effort.

No problem! First post on the forums, so wanted to contribute something.

RF Explorer 2.4 GHz ISM with the stock 2.4 GHz pigtail antenna connected to my PC.
Done in a fairly clean RF environment or at least as quiet as one can get in the Boston metro suburbs (aka, my office/basement).

I used their older free application for testing, but might try some waterfalls and other data logging with a slightly more advanced app.

 

[Shammyh]

Also, of potential note, reviewing the legacy DJI Lightbridge FCC filing (https://fccid.io/pdf.php?id=2219514) it looks like the "Phantom/Inspire-brand Lightbridge" is indeed different.

Looking at the above FCC doc, you can clearly see that both the channel width is different (~2MHz vs ~10MHz) and the channel-to-frequency assignment is different.

Also might explain why P3/I1-Lightbridge is not compatible with the standalone Lightbridge.

 

[N017RW]

Correct.

Do not use the stand-lone LB filing.

Use the P3 Pro or Adv.

 

[dedstik]

Hello all!

https://docs.google.com/spreadsheets/d/1WuKQD3FvbtFnJ6wkFu9DLxhdxHCSsnZPi90CwTGj3ks/edit?usp=sharing

I made a table of observed rf readings via a RF Explorer 2.4 GHz meter and put them in a Google Sheet.

Publicly editable should anyone wish to correct a mistake or add further useful information.

Spoiler: click for BIG rendered image of the table

Few observations:

1) Video is transmitted via ~10 Mhz wide channels from 2280 MHz to 2600 MHz. (600-280)/10 = 32 channels.

2) Control surface (analog sticks, buttons, etc) is ~1-2 MHz wide, and hops in the 2400 MHz 2490 MHz range only. The user-editable "controller channel" does not affect this signal.

3) Phantom channels 13-20 (as well as the control signal hopping) overlap WiFi 802.11b/g/n 2.4 GHz channels 1-14.

I have some screenshots from the spectrum analyzer, but for another post.

Would also like to take a closer look at the control signal in the future. Is it similar to DEEST or FASST? Don't know.

Enjoy!

- Sam

Edit 1: Updated with correct center frequencies and ISM start point.

Great job Sam...a nice clear explanation.

Thanks.