Purpose of individual channel GND in NCxxxMP modules?

[RickS]

Indeed, as @Speedskater says, with such things, there can be many acceptable solutions. Are you saying you run multiple wires from the board GND to each case panel? Again bypassing the RC connection via the heatsink? edit: You probably short wires between the panels?

No, but sometimes for the back panel. It depends on the case. The case supplied by Hypex not only had wave washers on the back panel, but the front had aluminum bars grounding the case top and bottom (assuming conductive screws are used). Hypex supplied conductive screws for the front rather than anodized ones.

My module is a NC500MP rather than NCx500, so not sure how much applies. In any case, my board uses the GND provided by J6.4 (the connector with the 5V standby power supply), and it does not connect it to chassis. J4.3 and J6.4 are shorted together on the NC500MP, I just checked this. With regards to points 2+3, connecting the amp to chassis via pin 1 already introduces a second connection between module GND and chassis, I guess what matters is that there aren't two or more shorts between amp module GND and chassis. I think this may go against what I was understanding earlier (above paragraph) about you using multiple ground wires.

I have NC250MP and wired the same as other modules as this is Hypex guidance. Am not a EE but Hypex refers to better noise immunity so I have followed just in case. When I have built custom amps for others, I want the best noise performance and ground loop protection so I also follow AES48. There are some special considerations if you are using a non-conductive case or XLRs that are non-conductive. In any case, a short wire from pin 1 to earth ground supplied by the power plug ground is simplest.

p.s.

Thought might be useful to point out that am pragmatic about the grounding. If your earth ground is to the bottom of your (conductive) case and your XLR connections are mounted on a conductive back panel, the Hypex diagram suggest that you make multiple connection to the earth ground via the back panel. The diagram suggests one ground point for the pin1, one for the input cable shield and one (potentially) for the XLR case. This assumes the back panel has a good path to the earth ground AND every other ground point can make good contact to the back panel. This is often impractical and can be problematic. Have not had any problems with short wiring to earth ground. For more than one input, may add another ground to the bottom panel and wire grounds to it. So far, have not had any issues (measured or in actual application).

 

[morethinkingcapacityneede]

To answer the OP main (subject) question - separately grounding each channel back to the chassis ground point will minimise crosstalk.

A shared ground connection would allow ground currents from one channel to influence the other, and vice versa.

So, this actually goes against what others have said. Those CH1 and CH2 are most likely shorted, and go to chassis via an RC.

 

[Speedskater]

The Safety Ground/Protective Earth should be attached to the metal chassis near where the AC cord enters the chassis.
All cable shields should be attached to the chassis at their connector. (not near the AC cord)
Audio circuit common should be attached to the chassis near the input connectors.

Often an internal shield is only attached to the chassis at the signal cable's connector. Because the Return Signal conductor is handled in some other way. So the shield's only task is to be a shield.

 

[RickS]

The Safety Ground/Protective Earth should be attached to the metal chassis near where the AC cord enters the chassis.
All cable shields should be attached to the chassis at their connector. (not near the AC cord)
Audio circuit common should be attached to the chassis near the input connectors.

Often an internal shield is only attached to the chassis at the signal cable's connector. Because the Return Signal conductor is handled in some other way. So the shield's only task is to be a shield.

Agree, my earlier divergence from this is based on the challenges around making good grounds. A low resistance wire is likely better than a questionably exposed case surfaces with anodized screws. If you want to be sure, need to measure the resistance of each connection.

 

[Speedskater]

The important point is:
The cable shield needs to be attached to the chassis very near its connector.
Say about a 1/2 inch or 1 cm.
Longer wires are interference antennas.

 

[RickS]

The important point is:
The cable shield needs to be attached to the chassis very near its connector.
Say about a 1/2 inch or 1 cm.
Longer wires are interference antennas.

Interesting, have not heard the term before. Please elaborate on what an interference antenna is and how it could negatively impact the amplifier performance.