开云体育

I'd tend to mostly agree with Max.

With DC and low frequency AC, and sometimes at RF, grounding with individual
wires or traces from each circuit to a single point often works best, as KF6VB says.
Otherwise, if two circuits share a ground path, the E=I*R losses of ground currents
from one circuit can be injected into the local ground for an unrelated circuit.

In the case of RF, there is a conflicting requirement of preventing unwanted radio signals
from getting into those ground wires.? These unwanted signals might be from an external
source (such as an automotive ignition system), this is best addressed by grounding each
outside connector where it enters the metal box, the metal box forming?
a Faraday cage around all the electronics.

Sometimes we want to avoid having our rig generate unwanted radio signals that might?
float across the room and get into other gear.? Again, this is addressed with a Faraday cage
around the entire rig, connectors grounded where they enter the cage.

Sometimes we want to avoid having signals radiated from one circuit in the rig getting
received by some other part of the rig.? This is why ham gear is often built as a bunch?
of little metal boxes inside the big external chassis.
Think of this as a chassis full of little transmitters and receivers that must be kept
from talking to each other.

Between single point grounding vs Faraday cage,? it is not always obvious which is best.
Grounding can be a black art, where you just try stuff and see what works
under the constraints of cost and time.

In the large digital designs I have done, it was generally best to just use a multilayer board
with a solid ground plane across the entire design.? Only in special cases was it?
found necessary to cut small islands into that ground plane for especially sensitive analog parts.
With a solid sheet of copper, the E=I*R losses are very low, and unwanted ground return voltages
are quite small.

This is what has been done with the uBitx main board, it's cheap and generally works just fine.
In some cases a good Faraday cage (metal chassis) around the uBitx will help, but most
of us don't have strong RF sources in the vicinity that would require such construction.
The uBitx does have some trouble with signals from one part getting into the other (perhaps the
reason for some residual transmitted carrier, and the power amp might inject some crap into
the IF chain),?but it seems to work well enough.

We might see less noise in the receiver if the ground return from the Raduino
was routed back to the main power connector coming into the uBitx chassis.? However, now the three
clocks out of the Raduino might have power supply noise on their nearby ground returns, so it's a tradeoff.
I'm guessing that for the Raduino, it's best to just make sure all those grounds (mostly the power ground and
clock signal grounds, the digital control signals into the uBitx main board might also be of concern)
are well connected to the main uBitx ground plane.? This is how it is currently implemented.

Jerry, KE7ER