made some RF discoveries today

[james_hunt]

Thanks for your hard work James - very informative. Can I ask a question. I have a P4 with a boosted MaxxRange antenna, as well as the Evo Lite. MaxxUAV recommends leaving the setup in auto. I'm wondering why that is and if that logic stands up to your findings? You mentioned in your video the sweet spot was quite narrow? Is it worth unlocking the extra channels if there is so little wiggle room for adjustments?

This is my setup as well. I think, as mentioned, MaxxUAV can't legally recommend using the hack. But that doesn't mean it's not going to work better. Based on the VSWR of that antenna, you'll get the most out of it when using ch. 20-24. But using other channels will still be better on that antenna than on the stock. Here's why:

The MaxxRange antenna has a forward gain of 10.5dB. The stock antennas are 3.5dB. The VSWR of the MaxxRange would need to reduce gain by 7dB before it has no benefit over stock. A VSWR of 6 equates to about 3dB reduction, while VSWR of 15 = 6dB of reduction. Look at the chart I plotted, which shows the entire 32 ch range is less than VSWR 15. So down at the lower channels you're getting gains of about 4dB over stock, while in the sweet spot of ch.20-24 you're getting the full 7dB over stock. Up at ch.32 you're getting near the stock levels of 3.5dB.

 

[The Flying Swan]

Here's why:

That was exactly the explanation I was looking for - THANK YOU. So in actual fact, interference aside, the antenna physically can't operate to it's full potential without the channel hack?

Do you find in the 20-24 sweet spot you can pretty much set and forget it, or do you need to monitor and tweak within that narrow range of channels?

 

[alexdds]

Anyone know if Autopilot supports 32 channels?

 

[The Flying Swan]

The MaxxRange antenna has a forward gain of 10.5dB. The stock antennas are 3.5dB.

Also, James, do you know if the sweet spot the same for both the MaxxRange and the ITE-DBS02 (the smaller one ~9dbi)? I have both but no spectrum analyzer (or knowledge)

 

[james_hunt]

Also, James, do you know if the sweet spot the same for both the MaxxRange and the ITE-DBS02 (the smaller one ~9dbi)? I have both but no spectrum analyzer (or knowledge)

I don't have the 9dBi to test, but it is lower gain, so that means usually a wider sweet spot. If you had to operate on a lower channel (<13) I would probably start with the 9dBi. Wider pattern and about the same gain at the lower freqs.

I always look for a good channel before I fly, using the Dji app. Once I decide on one, I'll choose my antenna accordingly, and only make adjustments to the channel if I experience some dropouts. I find that the spectrum doesn't change much during flight, so if I've chosen well before takeoff, there's no reason to tweak it.

 

[The Flying Swan]

Wow, so you actually choose your antenna in the field based on channel noise?

Shame I'm not in the US or I'd send you the 9dBi to add to the list. it would be great to see the data given it's one of the more popular antennas on these forums. Maybe someone else would be willing to sacarfice their antenna for a week of testing or add to your data?

What kind of hardware would I need to beg/borrow/steel to figure this out for myself?

Great thread.

 

[jephoto]

Can thE OP repost a link for that video about the channel interference?

 

[billy_LB2]

Glad to help!

- 10 MHz channel width for the LB video feed, though it only seems to be using maybe 8 MHz or so at a time.
- Just stock for the tests, though I aligned the antenna polarity throughout the tests, so I expected pretty high levels.

Interesting thoughts about the "rolling" aspect of the LB video signal, wouldn't surprise me at all that DJI is doing whatever they can to boost the legal signal levels.

Hope you get your hands on a P4P so we can see what's going on in the 5.8 GHz space!

On your P4 Controller when you did the video capture (of the 2.4Ghz and 5.8Ghz), what was the linerate set to for the Video on the APP?
P3A - P4P have selectabled Video Linerates of 4Mb, 6Mb, 8Mb and 10Mb.
Its would worthwhile checking with your cool and groovy Spec An the spectrum produced at each of those settings. I think here you may find your answer.

 

[james_hunt]

On my P4, the occupied bandwidth doesn't change when you select lower bitrates. That would be cool though. Lots of wasted bandwidth for a 4 Mb bitrate.

 

[Shammyh]

On your P4 Controller when you did the video capture (of the 2.4Ghz and 5.8Ghz), what was the linerate set to for the Video on the APP?
P3A - P4P have selectabled Video Linerates of 4Mb, 6Mb, 8Mb and 10Mb.
Its would worthwhile checking with your cool and groovy Spec An the spectrum produced at each of those settings. I think here you may find your answer.

I did actually try changing transmission bandwidth/bitrate/whatever slider in Go 4, and did not notice any difference in the video channel width nor in the amplitude. I didn't test it exhaustively/closely at the time, but I'll take another look at it.

I need to break out the RF Explorer again anyway, as I'm mildly curious about some of these SDR boost mode things that are available for other dji platforms in firmware.

 

[billy_LB2]

I did actually try changing transmission bandwidth/bitrate/whatever slider in Go 4, and did not notice any difference in the video channel width nor in the amplitude. I didn't test it exhaustively/closely at the time, but I'll take another look at it.

I need to break out the RF Explorer again anyway, as I'm mildly curious about some of these SDR boost mode things that are available for other dji platforms in firmware.

The Downlink or Video Bandwidth does not change with the selection of the different Line-Rates modes.
If you can take a capture of the entire UP and Dn Links transmitting simultaneously, you will see that the Downlink transmission bursts out and that the Uplink Hops will transmit between the Downlink bursts.

Someone in this thread or another has such captures already.

 

[Shammyh]

The Downlink or Video Bandwidth does not change with the selection of the different Line-Rates modes.
If you can take a capture of the entire UP and Dn Links transmitting simultaneously, you will see that the Downlink transmission bursts out and that the Uplink Hops will transmit between the Downlink bursts.

Someone in this thread or another has such captures already.

If you watch the video I posted a while back screen capturing my session with the RF Explorer, you'll see it's trivial to capture both the controller uplink signal and the video transmission signal at the same time. In fact, by using the channel hack, you can move the video feed to outside the uplink feed frequency space, therefore making then non overlapping.

The video feed does not impact the uplink FHSS at all. They don't take "turns", they literally just go over each other. In fact, you'll even occasionally see a FHSS spike within the video feed signal if you look closely. They are almost entirely independent of each other.

I've been scoping out LB for a while now... So about as familiar with the spec as one can be without decompiling the protocol directly on the SDR.

Maybe there's a language barrier here... But your question was if the video signal changed depending on the bitrate "switch" in DJI GO 4, and the answer to that is "not in any way I can observe". I assume that lowering the bitrate just increases the ratio of "guard" (likely some form(s) of error correction) to "signal". Aka, it's always this 8mhz wide signal, but at higher bitrates, more of the carrier is used for video, less for guard. And at lower bitrates, it's reversed. This part I'm mostly just speculating on though.

 

[billy_LB2]

So here's that new video I promised about meaDsuring the various antennas. The VSWR is a complicated measurement for sure, but a few basics are:

1) Most manufacturers pass or fail their products compared to a baseline VSWR measurement of >2. Meaning if it is higher than 2 across the frequencies it is intended to work in, it will fail and be rejected.

2) A VSWR of 1 is perfect and never achieved in the real world.

3) VSWR is logarithmic. A VSWR of 2 represents a loss of 10% of the transmit power. A VSWR of 3 represents a loss of 25% of the transmit power. Example: If you have a 100mW transmitter (like the P3A/P4), using an antenna with a VSWR of 3 for the channel you're operating on, can only output 75mW.

Dude this is great but I think these results are shady, I think the results are skewed by the cat, did you take into account the attenuation caused by the cat???

 

[billy_LB2]

If you watch the video I posted a while back screen capturing my session with the RF Explorer, you'll see it's trivial to capture both the controller uplink signal and the video transmission signal at the same time. In fact, by using the channel hack, you can move the video feed to outside the uplink feed frequency space, therefore making then non overlapping.

The video feed does not impact the uplink FHSS at all. They don't take "turns", they literally just go over each other. In fact, you'll even occasionally see a FHSS spike within the video feed signal if you look closely. They are almost entirely independent of each other.

I've been scoping out LB for a while now... So about as familiar with the spec as one can be without decompiling the protocol directly on the SDR.

Maybe there's a language barrier here... But your question was if the video signal changed depending on the bitrate "switch" in DJI GO 4, and the answer to that is "not in any way I can observe". I assume that lowering the bitrate just increases the ratio of "guard" (likely some form(s) of error correction) to "signal". Aka, it's always this 8mhz wide signal, but at higher bitrates, more of the carrier is used for video, less for guard. And at lower bitrates, it's reversed. This part I'm mostly just speculating on though.

Hey - so they dont just go "over each other". I dont know what level of FFT HackRF can support, but if you apply an FFT of say 32Khz with some overlap say 80%, I suspect you will find the answer is that UP and Downlink dont interfere with each other, this makes sense.
There's a reason why the Quad beeps like crap to show it has sync, which I assume to mean Timing Sync of both UP and Down links.

There has to be some serious timing synchronization taking place for the DN link to exist in the same RF as the UpLink.

Does that make sense.

 

[wa5tef]

Great video and info! As a ham I truly appreciate your methods of results. Thanks for posting.
Jim -WA5TEF

 

[handyman]

Hi !

I have a DJI Phantom 4 Pro + and I'm doing some frequency analysis.
As you said, I noticed the 2.4GHz stream band and the FHSS channels for commands.
Due to the frequency used, i supposed that the drone communicates over Wi-fi ? However i did'nt find any SSID (hidden or not) when i tried to analyze it, do you know if it's really using wifi model or maybe another proprietary protocol ?

Thanks for your help !

 

[wa5tef]

Handyman, I admire your interest in figuring out the Chinese and their secrecy on this matter! I am not familiar with the P4P as all I have is a year old P4. But I’d guess that they would not use simply WiFi for control. I would think they may be using a signal superimposed on another. Can you see any of evidence of that? It might even be shifted out if phase. Would your equipment show that? Please keep us informed!! This is very revealing information into dji’ technology And very interesting!!
Thanks
Jim
WA5TEF

 

[handyman]

Hi !
Thanks you for answering, I will informed you if I made some discoveries !
Do you know if someone tried to analyze the hardware of the drone ? I'm curious about the embedded chipset used in the drone to communicate

Regards

 

[wa5tef]

Handyman, no I have not heard any thing further. If I do, I'll pass it on to you. Thanks again for your research.
Jim
WA5TEF

 

[james_hunt]

Hi !

I have a DJI Phantom 4 Pro + and I'm doing some frequency analysis.
As you said, I noticed the 2.4GHz stream band and the FHSS channels for commands.
Due to the frequency used, i supposed that the drone communicates over Wi-fi ? However i did'nt find any SSID (hidden or not) when i tried to analyze it, do you know if it's really using wifi model or maybe another proprietary protocol ?

Thanks for your help !

It is definitely not wifi, just "wifi band".