Installing ezUHF, fighting noise – part 2
After quite a long wait for my ezUHF JR Module, i finally received it last week.
After a quick setup (I’m not going trough the binding and failsafe setup here, at this point) i got out and did a range test.
Results of the build in spectrum analyzer shown in my last post made clear there is a lot of noise on the F550. The noise floor turned out to be around 87 dBm. Noise is a big problem on multirotor platforms combined with UHF. A lot of equipment is crammed together in very little space.
So, for the first results:
EzUHF JR on LOW: ~450m
EzUHF JR on HIGH: ~650m
So…. time to get back in the ring with background noise for round two!
In part one, I didn’t want to open up the F550 and disconnect the UBECs. If the range test got me to ~ 1500m I wouldn’t take the hassle to continue the fight.
But, it didn’t brought the F550 to ~ 1500m, it brought me to ~450m, so I opened up the hexa, and disconnected the UBECs, my prime suspects. I also disconnected the cabling from the RX to the Naza flightcontroller.
Putting everything back to getter, I found out there are two troublemakers on my platform: The 12v 4,5A UBEC, and the Naza flightcontroller.
For now, I disconnected the 12v UBEC (that was powering the FPV cam and 800mW 1,3Ghz video transmitter) and hooked the cam and VTX on and separate 3S battery.
The DJI Naza flightcontroller emmits noise over the cabling to the RX. Using ferrite core’s on the servo wires improved the noise level from 89dBm to 95dBm! That is a factor 2 less noise! In total, noise is now reduced from 86 dBm to 95dBm, 9 dBm is a factor 3 less noise. It didn’t ever crossed my mind the naza was the main cause of the problem!
As soon as the weather and spare time allows, I’ll be out and testing again! If the range is acceptable now, I will need to find a better way to power the VTX and FPV cam. I might try connecting them to the balance leads of the 4S battery, I will have to test if it doesn’t discharge the lipo cells to unbalanced.
Below is the final graph. I still need to figure out where the 0dBm spike is coming from…
I will report back here with results soon!