Hmmmm Star ground

[Chris Hornbeck]

Solving ground issues inside an amplifier is often enough to keep everybody happy, but occasionally the outside world intrudes, in one of two ways. First is the ground loop formed by safety earths between chassis and power source. This part is well understood and more elaborate designs include internal isolation between signal "ground" and safety "earth".

The second is sneakier, and only applies to stereo within a single chassis. The interconnections between chassis using single-ended ("RCA"/"phono") will have a common ground at both ends so, even in the best case of signal ground isolated from safety earth, signal returns through the interconnecting cables are split evenly between both channels' return wires (commonly also the shields, but true even with telescoping grounds). This damages noise immunity whether coax or twisted pair.

Often not a "problem" but things to think about.

All good fortune,
Chris

 

[Shoog]

Before:



Star ground or ground spaghetti monster?

After:



There are a few other changes, like cleaned up tone stack / gain control wiring. And the big honkin 1R resistor. And the heater wiring for the 9 pin socket not only exists but has a switch to toggle between 12ax7 and 6n2p.

Just like to point out that your original build was not a star ground - it was a round bus. It was almost guaranteed to create some hum.

Shoog

 

[Wavebourn]

Its vital you keep currents in successive stages away from each other or you will introduce noise.

It's all about the ground, AKA Common Wire.

 

[ericj]

Just like to point out that your original build was not a star ground - it was a round bus. It was almost guaranteed to create some hum.

Shoog

I am skeptical that it makes much of a difference if the center of the 'star' is a stud or a small ring. i doubt, for example, that soldering a bit of sheet metal across it would have made any difference.

 

[Shoog]

I am skeptical that it makes much of a difference if the center of the 'star' is a stud or a small ring. i doubt, for example, that soldering a bit of sheet metal across it would have made any difference.

Its the very definition of a loop.

If you want to see if it makes a difference to do it properly then do it onto a solder blob or a bolt with tags and report back. Your experience isn't that informative until you do.

This is exactly what I was saying about people hitting problems when they diverge from the correct methodology

Shoog

 

[Wavebourn]

I am skeptical that it makes much of a difference if the center of the 'star' is a stud or a small ring. i doubt, for example, that soldering a bit of sheet metal across it would have made any difference.

The center of the star have to be at the output of power supplies, where main filtering ends. Look again at the picture I posted above. You see blue big film caps that shunt 800 B+, caps that shunt filament, bias, regulated 270 and 400V, all them have common point of interconnect by shortest traces, and multiple posts for ground wires that go from each stage here, as well as from input, output sockets, chassis, and power ground. Wires can be quite thin, since voltage drops caused by big currents are not applied to input ground wires.