Posts: 1,504
Threads: 83
Likes Received: 944 in 654 posts
Likes Given: 2,142
Joined: Sep 2016
Reputation:
24 (24-Aug-2016, 05:28 PM)trojanGoat Wrote: I don't ground any of my ESCs next build I guess...
Hey TG, I know its almost a year ago now, but did you notice any significant difference in quad behavior after connecting grounds?
Has connecting them just become SOP now due to the growing popularity of DShot?
Windless fields and smokeless builds
• Posts: 18
Threads: 3
Likes Received: 4 in 4 posts
Likes Given: 0
Joined: Dec 2017
Reputation:
1 I'm just building a first quad from scratch, and have been wondering about ESC signal grounds. It is my first attempt at this, but I do have an electronics background - and in noisy environments.
It seems to me it is impossible to say what is right without seeing the FC and ESC schematics and PCB artworks - and I've seen neither. Noise, ground loops etc is a tricky business, so is understanding what is happening.
I can see the arguments for connecting them, but also there is the problem that (a) one almost certainly creates ground loops and (b) it is inviting large earth currents to flow places where they probably shouldn't. Both of those are bad ideas.
If the signal was balanced (RS485 style) with good common mode rejection, or if the ESC was opto-isolated, there wouldn't be a problem. But the former is not the case and the latter seem uncommon, it seems to me that non-BEC ESCs are often called opto, but actually they aren't. If they were, there would probably be no connection between between signal and power grounds, mine for example read 0R.
There are other ways of reducing noise a bit without ground loop / high current issues. One is to connect one end of the ground only as a shield (ideally use coax). The trouble with that is one might just create an aerial. The other is to connect the open end with a small cap. That way high frequency currents flow, low don't. On the other hand, that is just a high frequency ground loop.
So, what to do? Difficult. These ESCs are going to run DShot300. That's a purely digital protocol, and from memory, I think it has a 4 bit CRC. The motors are only 1306s so current is not massive. There's no commutator and therefore no sparking, so frequencies won't be that high.
So based on all that, and a desire not to cook the FC if things to wrong (as they do, seeing that this is my first go). I'll be starting with signal wire only, and see what happens. Can always add the ground later.
Posts: 1,504
Threads: 83
Likes Received: 944 in 654 posts
Likes Given: 2,142
Joined: Sep 2016
Reputation:
24 This is a great vid from Josh Bardwell about why you SHOULD use signal ground. Basically, (though from your post I doubt basic is necessary!) it allows a low a resistance path between ESC and FC to simultaneously recognize the voltage potential of the noise, so that the voltage potential of signals are still recognized accurately. Using common ground adds delay to the ESC - FC signal connection, due to all the other components that are also connected. Considering this, the baseline of the ESC may be different to the FC due to the delay, thus the voltage potential of the received signal is different to the intended value.
Windless fields and smokeless builds
•