Star grounding the power supply?

[Deke609]

Another question for the experts. I'm wiring up the BeePre rebuild and need to decide on the final details of the grounding scheme. Like the stock BP, the rebuild has a single, separate star ground for each channel, with each star connected directly to the chassis. 

Question: can the B+ power supply be star-grounded? I.e., have the negative lead of each cap connected to star ground, rather than  serially connected?

I ask b/c my CLCLC raw B+ power supply is a bit spread out, so running wires to star ground isn't much more effort, and wouldn't use all that much more wire, than serial/daisy-chain connection.

Now, I more than half suspect that my former hum problem will return with the all metal chassis (non-humming prototype rebuild was all wood) and that I will need to tinker with the ground scheme. So this likely isn't my last kick at playing with grounding in my Beepre rebuild.  But if star grounding individual components of the B+ power supply is a complete howler and a no go, I'd really appreciate someone letting me know.

MTIA, Derek

[Paul Birkeland]

The BeePre does not use a star ground.

The serial connections are a ground bus, which is what the BeePre uses. 

Since you're building from scratch, you may have to experiment to see what works best.  Thankfully the BeePre is rather forgiving.

[Deke609]

Thanks PB. Right. My bad. I was focusing on the fact that for each channel the power supply 0V and the signal 0V meet at common ground reference somewhat close to the center of the chassis. But a point with 2 branches does not make a star.

But no obvious violations of electronics principle or wisdom if I star ground each B+ power supply?

many thanks, Derek

[Paul Joppa]

There are a lot of ways to achieve the same ends. I like Blencowe's article because it talks about what those ends should be, as well as his way of achieving them.

http://www.valvewizard.co.uk/Grounding.html

His way is sequential nesting  (daisy chain). A star ground IMHO just eliminates the need to trace all the current paths - if it is low enough resistance.

[Deke609]

Many thanks PJ. My biggest fear is ground loops - and avoiding those is what attracts me to the star ground scheme. 

I had read the Blencowe article before posting, as well as the section on grounding in Henry Ott's "Electromagnetic Compatibility" book - and came away with conflicting understandings. I suspect that more reflects my need to spend some time thinking through what they've each written than an actual conflict between their writings.

But there is one thing that definitely confuses/confounds me. Blencowe suggests that the virtues of star grounding are defeated if the wiring to the star is long -- b/c of resistance and inductance. I don't understand this. I would have thought that if the wires to star ground are of approximately equal length (and diameter/composition), they provide an approximately equal resistance/impedance to ground, regardless of whether they are 0.1 inches or 10 inches. 

 I think PB's suggestion to just try it and see is the way to go. I can switch from serial to star grounding relatively easily and quickly. I just wanted to make sure I wasn't doing something totally crazy or damaging.

cheer and thanks, derek

[Thermioniclife]

I find it helpful to recite this prayer at the start of a new project.
Blackest of the black voodoos, give me the power to ground my amp and the cunning to do so safely. I summon you to smother hum beneath the multitude of your slithering sable tendrils and to untwine my ground loops as if they were the entrails of philistines. For thine is the impedance, the silence, and the Safety Earth, now and forever.
on a more serious note this guy has a lot of good info on his site. https://wtfamps.com/grounding/

[Deke609]

Thanks Lee. Hah! Until I followed the link, I figured you came up with that off the top of your head, thinking, "Wow, this guy can write". I even had to look up the meaning of "sable" -- means "black". 

cheers, Derek

[Thermioniclife]

Derek,
His explanation of plotting load lines is about the only one to make sense to me, almost.

[Deke609]

I agree. He does a really good job of boiling things down to the basic working principles. I just now read his explanation of how a basic triode-based shunt regulator works and it actually makes a bit of sense to me. That's a neat trick of connecting a cap across the plate and grid to use feedback to cancel AC ripple.  I'm sure it's a commonplace design for those in the know, but for someone just starting the learning process, it's pretty cool.

[Edit - and I am guessing that the same scheme works with an IC regulator. Just looked again at PJ's LM1085 fil reg board in the BeePre. UNtil 2 minutes ago I had (embarassingly) always read C3 as "CRIPPLE" - thinking that "cripple" must be some term of art. Pretty sure now that should be read as C-ripple -- i.e., the ripple cap.  Hah!]



cheers, Derek