jneutron said:
I'd place this a lot higher on the to-do list of design than sub 100db distortion. In fact it's effect is very hard to measure because the many things circulating on the grounds become part of the reference for the test equipment; very easily heard but very difficult to measure.
Funny thing, I got interested in this thread because I saw the picture posted early on of the Blowtorch and thought to myself, "self" the layout and grounding are all wrong in this thing... I'll just see if I can strikeup a conversation and discuss what I see and how I do it from my experience... Might learn something, might share something (good for the soul).
Pretty much, I'm an idiot and the conversation here keeps recycling itself. Lots O luck pointing out helicopters.
Mike.
john curl said:
I wasn't talking about the three wire ground, I'm talking about the way the innards are interconnected.
I've studied this for thirty years and my take is that I would have done it different. No big deal, there's no right or wrong here.
There is a difference between experimenting allot and strange.
This is actually funny, the first time I tried to suggest a different approach I was attacked by the safety ground crew... I kept thinking to my self, this is DIY Audio, what the frick are we worrying about the safety ground. If I shock myself I'm stupid. What actually sounds good? I'll worry about that later. I'm not trying to sell it, just amuse myself with music.
Mike.
As the signal is separated from the source, to a degree,(*) at the given point of generation...it is then is subject to the following circuit proper. Since it is 'loosed' to a degree, it is then on it's own. Like a set of waves on the surface of a pot of water, it can be manipulated, but shares it's energy (in it's differing forms) within the given 'area' of transmission. Adding signal to 'repair' to return to a better replica is not the best idea--With regards to distortions, the first line of defense is control of the gating of the inevitable losses.
*The sharper the separation(change) - the greater the reflection - the greater the distortions.
These things are both obvious-and not.
*The sharper the separation(change) - the greater the reflection - the greater the distortions.
These things are both obvious-and not.
lumanauw said:
Define share.
Sometimes ground is a return path and sometimes it's only a reference point. Mixing the two purposes is a problem. It also helps to define where the true reference is between circuits and actually between different components (amp, preamp, CD), not to forget the path of the return currents from all of the regulation and RF bypassing.
Then there's the loop to consider, and the impedance of the gain stages these returns and references service. I keep hearing that "oh yeah, the layout is important, but then I see the examples of "good" layout and think, someone needs to give this a bit more thought.
Then what happens when one throws in an alternate (unintended) path, a ground plane, a chassis, a heatsink, etc. what's the attraction and where's the path back to the (still undefined) reference? What's the result of a less than well chosen placement for a bypass return?
As for the ground of the output of a preamp, there is only reference needed. It has only one purpose and it has nothing to do with the bias currents, RF bypassing, etc. The FB should be referenced to it since the FB is just sampling the output and feeding back a voltage to the input stage. After sampling, the loop currents back through the input stage are extremely small and very unlikely to disturb the device bias currents.
The output ground has more to do with the ground reference of the power amp (source, etc.) than it's own circuitry or power supply, think difference between return and reference: after all it's a voltage source not a power amp (even one designs a low Z output to drive ridiculous cable designs).
Just thought I'd try to add some food for thought rather than just being disagreeable.
Regards, Mike.
I think Mike has something to offer if not easy to explain or clarify. I have been working to keep currents away from ground for a long time, to the extent that I don't use a connection between the supplies and ground, deriving the relationship instead.
It might be easier to think of the ground system a a network of resistors with significant resistance in series with the supposed reference, and any circulating currents degrade the quality of that reference. And a milliAmp across a milliOhm is a microVolt, a voltage that may be higher than the noise or distortion floor of the design. One milliOhm of trace resistance is a very very short trace. A single ended design with the power supply return flowing across ground would be a real challenge to keep clean.
It might be easier to think of the ground system a a network of resistors with significant resistance in series with the supposed reference, and any circulating currents degrade the quality of that reference. And a milliAmp across a milliOhm is a microVolt, a voltage that may be higher than the noise or distortion floor of the design. One milliOhm of trace resistance is a very very short trace. A single ended design with the power supply return flowing across ground would be a real challenge to keep clean.
- Status
- Not open for further replies.