Basically signal ground connected to chassis, then lifter between chassis and earth.
You have the EM susceptibility in the loop and you have the common impedance problem. The 'head on' way to deal with the common impedance issuer is to keep the 0V return impedance between the two channels to a minimum so that the voltage drop seen at the input signal grounds before the HBR is minimized.
The HBR will 'split' the loop current so that most of it is confined to the 0V returns, and only a very little flows through the input signal grounds which are usually joined and the connectors located next to each other in a good conventional layout.
So, the first conclusion is that we have to maximize the ratio between the 0V umbilicals and the HBR signal ground loop current - I threw out a figure of 5 mOhms for the umbilical - can we get it lower? It might be possible to get it to 2 or 3 mOhms or less, and the HBR is at 10 Ohms so the loop error current is now at 1/5000.
Of course you can take a hammer to the problem and build it with dual mono supplies.
Another option is to cancel the signal ground loop current using an opamp where you invert and then sum the signal ground loop current - purists may not like it though.
A balanced input of course solves all of this . . . and is very elegant
I don't agree with that approach.
I would always keep the input connector barrel isolated from the chassis. Its only connection to the 0V is via the HBR on the amplifier modules.
The ground lifter is a separate subject and is there to break the AC ground loop caused by the loop created by the interconnect cable and the earth (ground) connections to the mains supply. we have mainly been talking here about cross channel ground loops and not the AC power ground loop - two separate things.
(BTW I see in both of those diagrams on the links you posted, the HBR is taken all the way back to the star ground which in a system where you have two amplifier modules and a centralized PSU would create very large loop areas. I think a better way to do it is to take is back to the amplifier module PCB star ground - ilimzn's posts (see post #1 on this thread - 3rd link) shows this clearly and I concur with his view.)
Last edited:
There's no difference in his view and mine - I've put his posts up in the first post on this thread (3rd link) in PDF format. He's captured the whole essence of the unbalanced input noise problems superbly.
That's why I asked what keantoken was referring to and he said the loop through safety earth
Post 52? The second diagram in your link. Also from post 36 "Also, do not shield your components connecting the case to the earth wire! You are not building a stove or washing machine. If you do that, it will capacitively interact with everything in the box, via the voltage difference between the neutral mains wire and earth - pretty much any electrical disturbance or load in the building and possibly more.
The case should be connected to the component's ground and then that should go to earth via an earth loop breaker."
The case should be connected to the component's ground and then that should go to earth via an earth loop breaker."
Last edited:
As a newbie I've tried to read some of the stuff about grounding and what for me seems to be
most important principles about grounding (besides safety reasons!) is to reduce induced noise by
1) Reducing loop area and think loop area with a system approach: all components that
shares common ground or interconnect (pre-amp, amp, digital line level etc).
2) twisted pair cable can shield EMI if and only if the second conductor is the sole return path of the
first conductor since any induced current will then cancel out. And with coax the same if sole return current is carried on the inside of the shield.
I am not yet convinced that hum breaking resistors are really necessary if a system approach
is followed.
This guy below was a little bit funny: He recommended bonding all chassis inside a ham shack
with a short and thick copper braid (in parallell with interconnect cable) and then only couple
the last to ground.
http://k9yc.com/GroundingAndAudio.pdf
http://www.audiosystemsgroup.com/KillingRXNoiseVisalia.pdf
most important principles about grounding (besides safety reasons!) is to reduce induced noise by
1) Reducing loop area and think loop area with a system approach: all components that
shares common ground or interconnect (pre-amp, amp, digital line level etc).
2) twisted pair cable can shield EMI if and only if the second conductor is the sole return path of the
first conductor since any induced current will then cancel out. And with coax the same if sole return current is carried on the inside of the shield.
I am not yet convinced that hum breaking resistors are really necessary if a system approach
is followed.
This guy below was a little bit funny: He recommended bonding all chassis inside a ham shack
with a short and thick copper braid (in parallell with interconnect cable) and then only couple
the last to ground.
http://k9yc.com/GroundingAndAudio.pdf
http://www.audiosystemsgroup.com/KillingRXNoiseVisalia.pdf
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.