LoRa Causes Unstable Power and Induces Servo Jitter

[CCCanyon]

I have a tilt bicopter that has two servos, power schematics as the pic.
One Arduino Nano, an AMS1117 3.3v regulator, two servos provided with 5V by a mini 360 regulator
The Lora module runs on the 3.3V from Nano and AMS1117 3.3v regulator.

The 5V power line is unstable, causing the reading of the pot inside the servos unstable, ultimately causing servo jitter. The jitter happens at the exact timing when my flight controller receives and transmit packets. The jitter stops when I touch the end of the antenna or the 3-wire servo cable strip with metal or hand.
I've tried adding a capacitor to the power lines of the LoRa module and servos, it didn't work.
Coiling the servo cables to DIY EMI suppressing rings made of paper clips works, but I think there must be a better way. I'd like to know what else can I do.

[jremington]

It is not a good idea to have the Arduino and motors share the same power supply, for the reasons you posted, as excellent power supply decoupling is absolutely required.

The safest and simplest approach is to use separate regulators, one for the servos and another for the MCU and peripherals.

Make SURE that the servo regulator can handle more than twice the servo stall current.

[Ballscrewbob]

BUCK converters can sometimes be underpowered so make sure yours is capable of supplying all the amps required for ALL the devices at once.

As it appears to be a NOISE issue that could still be the convertor OR some bad shielding on the part of the layout. Maybe a good picture of your current setup would be worth looking at.

[Paul_KD7HB]

See if the regulator input voltage ever gets below the minimum for the regulator. 4.75V ≤ VIN ≤ 12V from the data sheet.
Paul

[Idahowalker]

Going back to the capacitor. I have a ESP32 CAM taking frequent pics for the last year. I had trouble with getting decent power to the ESP32CAM. Along with taking pics I used WiFi to FTP the images from the cam. WiFi use would, from time to time cause various power issues.

I used 2, 2.7V 1F capacitors wired in series across the 5V line. Solved my issue.

[CCCanyon]

Thanks!
@jremington : I'll try ditching the AMS1117 3.3v and power the Arduino directly from the 2s li-ion.

@Ballscrewbob : here's the cramped setup

@Paul_KD7HB : no, it's 2s li-ion 7.4v to 8.4v

@Idahowalker : my helicopter is tiny, it won't fit

[Ballscrewbob]

I foresee some signal issues in the near future

[CCCanyon]

I detach interrupts from Lora DIO pins 1ms early. It solves the problem of servo pwm signal being chopped 500us occasionally.
Power jitter still not subside after adding bypass capacitors.