Black Cube produces excessive RF noise

Hi! recently I came across a severe issue with cube black autopilot. I configured the autopilot and installed it inside the fixed wing frame, and were about to give it a flight when I discovered that my long range RC system (Dragonlink) was failing to pass a range test check and reporting significant noise levels on the receiver side. At first, I thought that either Dragonlink system was faulty or there was a unit inside my plane that radiated too much noise. I spent several hours in attempts to investigate the issue and found that Dragonlink performs normally on a different frame (with different FC installed) without any issues producing excellent range check results in identical environment (with only receiver being inside another frame with a different FC).
Later, I removed the Cube from the frame, powered Dragonlink receiver by USB and ran a spectrum analyzer to assess the noise level coming from the Cube. When I power the Cube (either by USB or by using Cube’s stock power module) the noise level jumps from -110 dbi to -85dbi across entire 433 mhz band! With such high noise levels I have to ground my equipment or find a way how to mitigate interference from the Cube. Any advice on whether such extreme noise is common with Cubes or I did I purchase a faulty one? Thank you.

No it’s not common. I remember only one post where RF noise was traced to the cube. The RF noise was at 122MHZ in that post not 433.

I agree, this is strange. I didn’t expect such kinds of problems with controllers of this kind. This is really frustrating. I will try to send it back to the reseller. If not possible, will have to throw it away…

The cube has been through rigorous testing in many commercial aircraft and been ce and FCC certified in these airframes.
This is not normal at all.
Unfortunately without knowing the full design, layout, and environment of your aircraft, it’s not really possible to give you much advice.
If replacing the flight controller doesn’t fix the issue, then possibly giving more information about your whole system here may allow us to give better advice.

The more information you give, the more we can help.

philip, thank you for your fast reply. In fact, the setup is simple enough. I removed the Cube from the frame to have as a few RF emitting stuff as possible. The only things I have connected are:

  1. Cube Black
  2. CubePilot power module which comes with the controller
  3. Dragonlink receiver connected by USB to my laptop computer. The receiver itself is placed next to Cube with about 5 cm distance from the Cube and receiver antenna is about 50 cm from the cube.
  4. 3s LIPO pack.

The way I conduct the test is the following. I connect DL receiver to my laptop and run spectrum analyzer to inspect the noise level across 433mhz band (the range is from 422-436mhz) before the Cube is powered up. The measured noise level at this point is around -110 dbi and is uniform across the band.
Next, I connect Power Module to 3S LIPO and observe no change in noise floor (Cube is still unpowered) - the noise level remains unchanged.
However, once I power Cube via Power1 or 2 I observe immediate uniform increase in noise level to about -85 dbi.

Thank you and looking forward to receiving any help.

I’m actually not to concerned about that level of interference that close to a receiver, you are installing the dragon link in a fixed wing aircraft, so surely you are installing the dragon link as far away from everything else as possible.

With all RF equipment, distance is your friend when installing antennas. Keep the GPS away, keep payloads away.

Ignoring all of that for a minute though, you now have a baseline.

How are you powering the dragon link? You must never power a high power radio off the Autopilot. The dragonlink must be on its own power system.

If just checking these things cannot fix this, then do replace it, and repeat your test.

I have micro receiver which does not draw much current. Anyway, in the test setup dragonlink receiver is powered from my laptop. nothing is connected to the Cube except power from LIPO.

Try the new cube. Then report back.

OK. Just one newbie question: how do I return my Cube? Can I return it directly to manufacturer?

Contact your reseller, and ask them.

thank you, philip. much appreciated.

1 Like

The Dragon link is extremely susceptible to noise from other systems on the aircraft. Distance is just one factor involved. It takes quite a bit of effort to get the wiring, grounding, bonding optimized as well as layout of the vehicle right to minimize the interference from other systems. You’ll need an oscilloscope, some RF test equipment, and some patience to get it right.

thank you for your input. Today I discovered that actually much of the noise is coming from the power cable connecting FC with power module. I lengthened it by about 10cm as per my setup requirements. Could the lengthening of the cable contribute to noise as well?

Yes. 99% of what you are dealing with is http://ardupilot.org/copter/docs/common-magnetic-interference.html#maxwell-s-equations-the-magnetic-field-on-the-axis-of-a-current-loop

  1. The magnetic field increases as a function of the enclosed loop area Pi R^2
  • Big loop = big magnetic field
  • Reduce the magnetic field by twisting the loop to close it. aka twist the wires together and keep your return path close to the source
  • In the case where you have a PDB and 4 ESCs you have 4 loops that start at the Deans connector on the PDB, go out to the ESC and return back to the Deans connector. In the case of the 4 in 1 ESC the current flow is much more concentrated in one area and you don’t have the big loops so consequently you decrease the magnetic field.
  1. The magnetic field increases as a function of current
  • Where there is the option, you can deliver the same power and decrease the magnetic field by increasing the voltage and decreasing the current.
  1. The magnetic field decreases as a function of the cube of the distance
  • It is more complicated when you are close to the loop but when z >> R the denominator goes to z^3
1 Like

Thank you, I will look into that more closely. However, I do not think that such large noise numbers are a result of my lengthening the power cable. It is also interesting to note that I tried to isolate as many interference sources as I could, by having only a small battery, a power module and a flight controller itself, to estimate the RF noise on the dragonlink receiver. I also noticed that if I remove current and voltage signal wires I get slightly better picture on DL oscilloscope and any movement of power cable relative to DL receiver has a noticeable effect on the noise reported by it.