Home        Store        Learn        Blog

Dual T-200 jittery motors only when both are running simultaneously

Hi All - I have the build posted here

It is 5s20p with arduino mega sending signal to two Blue Robotic ESC. Both motors seem smooth at 100% but anything short of that feels like it is toggling the power off and on extremely fast (based on the sound). If only one motor is running it sounds extremely smooth. I do not believe it has anything to do with lack of power as this situation occurs at 5% speed as well. Any ideas? Is it possible noise is being injected into ground somehow? My potentiometers signals seem to be extra jumpy while motors are under power as well.

I’m away from my motors for sometime before i can test any changes but maybe it is something i can do with my box in the meantime.

Hi @cody00,

I’d note that unshielded jumper wires and the like aren’t exactly renowned for their signal integrity, especially when they’re close to high power, high frequency sources of noise like motor wires.

General Noise Reduction

Some common approaches to noise reduction include:

  • maximising physical distance between noise sources and signal wires/receiving circuitry
  • minimising how far each signal has to travel, to reduce the region(s) where noise can be picked up
  • shielding noise sources to keep noise contained, and/or shielding signal wires + circuitry to keep noise out (that can be anything from wrapping things in foil to enclosing them in metal boxes and/or using cables with built in shielding)
  • using twisted pairs and equal-length wires for any differential signals, to keep noise on each wire as equal as possible so that it cancels itself out when measured (‘impedance matching’ is the more general concept here)
  • using high contact (e.g. soldered / screw terminals on bare wire or crimped ferules / seated-connector) connections, to avoid changing resistance/connectivity from minor movements
  • ensuring components in the same system share a common, low-resistance ground, without forming loops (e.g. try to have ground connections branch out from a central location)

Noise Detection/Measurement

Of course it’s easiest to isolate noise sources from signals if you’re able to detect where and when noise is occurring. If you can access one, an oscilloscope can be really useful for probing and measuring noise levels. More generally, systematic testing of features in isolation can help determine where the largest problems lie.

Your Scenario

In this case you’ve found that the pot signals are jumpy when the motors are running, so a good starting point may be to run a program that tests both motors without taking input from the pots. If that’s smooth then it may help to try isolating/protecting the pot signals from the motor power and output cables.

If it’s not smooth then there’s either an issue with the signal getting to the motors (consider isolating/protecting), or there are problems with the power supply (either insufficient capacity, which you’ve already determined as unlikely, or noisy from the thruster draw, which you can protect against/smooth with decoupling capacitors between power and ground, and/or inductors/ferrite beads in series with/on the power line, preferably as close as possible to the affected device(s)).

Hope that’s helpful! :slight_smile:

1 Like

Thanks for the helpful information Eliot. After reviewing your information with my setup I believe the #1 suspicion lies with the 21V → 10V Buck Converter. Apparently it can feed some massive noise back into the input. Ill look at doing decouple capacitor on it or figuring out a better way to power my components.

1 Like

I actually think the easiest test is to power the arduino with a USB battery bank. Disconnect the buck converter and see how everything works. At that point it would just be the ESC tied into the main battery.

1 Like

I ended up replacing my cheaper/noisier buck converter with a buck converter that I believe will work better. It also has terminal on the both in and out for playing around with different capacitors.

I will not be testing it until spring at this point but will report back.

1 Like