Groundplane layout theory

[Richard Ellis]

Wouldn't a groundplane set up as a circle like this, as either a solid copper panel or thick style PCB..circular, set off the chassis two inches...the ground connects would be a short one and a half inches...all components would be situated "around" this groundplane...all connects would be short....Ideas??
___________________________________________Rick... ........

[JesseG]

WHY?

[Richard Ellis]

I guess coming from an antenna design background...can you say collinear array?...., I see E & I fields everywhere...it would even out ground patterns as each component would be equadistant(sp?) from each other, the green ground from the power-cord would pop up in the very middle....no one component would see a different ground...I'm seeing fields (Dead people?) everywhere........this setup would eliminate that potential variable.
___________________________________________Rick... ........

[Miles Prower]

Originally posted by Richard Ellis
Wouldn't a groundplane set up as a circle like this, as either a solid copper panel or thick style PCB..circular, set off the chassis two inches...the ground connects would be a short one and a half inches...all components would be situated "around" this groundplane...all connects would be short....Ideas??

I already do something similar, though not exactly like that. The last project I did had four subsystem circuit boards: front end and driver, positive rail supply, negative rail supply, and active screen volatge regulator. Each one used a "modified dead bug" construction where connection pads are etched into the copper cladding, which serves as a local ground plane. Each board is insulated from the metal chassis. The ground planes of each board are connected individually directly to the DC neutral, which is then connected to the chassis via diodes wired anti-parallel (i.e. "69" style) and paralleled by a 10R resistor and 0.1uF capacitor. The chassis being connected to the third wire safety ground.

This semi-star grounding works very well, and makes for dead quiet. Only way you can tell if it's on is by the glow of the tubes.

[JesseG]

This is certainly interesting.

I have tried doing spectral analysis (with VERY limited test gear) on my hand-built amp with both complex and simple signals in the audio band to see if there is any non-linearity being introduced by external fields (I was curious)

I couldn't find anything besides the signal, with identifiable distortion. However I did not look above 40Khz.

Miles: I like your description of your grounding scheme. I have made some tube ccts that are pretty quiet, but not so quiet that you coul'n't hear some hum if you put your ear to the speaker. I guess I have just accepted that as the price of using H/V power for tubes.

it would even out ground patterns as each component would be equadistant(sp?) from each other

Richard: what I find interesting about your idea is that is echos something I read in a review of the Lyra Conniseur preamp (HiFi+ calls it the best preamp ever made). Apparently, the active circuits are arranged in a cube with precise spacing between every component. measured capacitances and inductances are factored into the component selection. Seems they are using 0.1% components and they claim this results in the lowest distortion (and the pricetag of $20,000!).

Jess

[EC8010]

Originally posted by JesseG
Apparently, the active circuits are arranged in a cube with precise spacing between every component. measured capacitances and inductances are factored into the component selection. Seems they are using 0.1% components and they claim this results in the lowest distortion (and the pricetag of $20,000!).

Well, they need to do something to justify a price tag like that.

Seriously, ground planes are a great idea and well worth the trouble.

[arend-jan]

Ground planes are good because they have very low inductance/resistance and very small loop area. The disadvantage could be increase in capacitance. You certainly don't want that at some places.

[Miles Prower]

Originally posted by JesseG
Miles: I like your description of your grounding scheme. I have made some tube ccts that are pretty quiet, but not so quiet that you coul'n't hear some hum if you put your ear to the speaker. I guess I have just accepted that as the price of using H/V power for tubes.

I didn't say there wasn't any hum. I o'scope about 6.0mVp-p of a 60Hz sine wave at the output. Really, who cares? I certainly do not listen with my ears right up against the speeks. I suppose I could've gotten 30 car batteries, connected them in series for the HV, and eliminated all the hum, but really...

[lochness]

I guess coming from an antenna design background...can you say collinear array?...., I see E & I fields everywhere...it would even out ground patterns as each component would be equadistant(sp?) from each other, the green ground from the power-cord would pop up in the very middle....no one component would see a different ground...I'm seeing fields (Dead people?) everywhere........this setup would eliminate that potential variable.

Off topic but I now want a shirt that says "I see E & I fields".

Other than that this topic is beyond me. I just use a simple star ground...no problems yet.

[arend-jan]

Originally posted by Richard Ellis
...it would even out ground patterns as each component would be equadistant(sp?) from each other, the green ground from the power-cord would pop up in the very middle....no one component would see a different ground.

I prefer not to think about 'ground'. As far as the operation of the amplifier is concerned, ground and the green wire from the power cord are rather uninteresting. It's better to think current loops.

Lynn Olson on current loops

Thinking current loops would have probably prevented the mistake this DIY builder made:
spot the mistake


Groundplanes can get you good current loop control because the AC return currents are free to choose the path of minimum induction, and the induction of a wide flat conductor is minimal. Of course there are plenty of opportunities to shoot yourself in the foot with a ground plane too.

Ground (bounce) and more on current loops
And more on ground planes / current loops

Originally posted by Gesse
I have tried doing spectral analysis (with VERY limited test gear) on my hand-built amp with both complex and simple signals in the audio band to see if there is any non-linearity being introduced by external fields

What would be a mechanism by which an external field could cause non linear distortion?