If you focus on the looks of it, it may be an “art”. If proper functionality is in focus, then it is a learnt art, i.e. a craft requiring knowledge and skill in applying reasoning.
If (*) proper functioning of a circuit is affected by the 3D morphology of it’s physical implementation, then circuit layout should be considered by the electrical circuit designer an inherent part of his job.
George
(*)I am a bit sarcastic here. It is well known that it is affected. The question is if the “parasitic” capacitances and inductances due to component positioning are considered much more important than some parasitics due to component’s own construction.
I think the comment was that split prim sound better.
The only thing I can think of is the winding geometry. With split prim, you
need a layer of insulation between the two windings due to the voltage
gradient in parallel (120V) mode.
So split prims will have less 'tight' winding geometry = slightly more leakage
inductance, but it must be pretty small difference.
I think here Ed is referring to voltage gradient across the prim when
distributed / coupled to 2 secondaries can cancel.
Measuring - Good idea.
And FWIW you can easily go less than 38puf (without ES shield) if you are
careful with construction.
negative and positive are never symmetrical in draw, thus a split tap causes unequal loading in the dynamic or instantaneous impedance considerations at the the given moment of complex draw.(one rail vs the other)
To avoid this non-linearity under dynamic draw in the micro moment sense, we go to two transformers.
The art of perfection, not the art of good enough.
That art of perfection sees the problem and then demands the solution.
Last edited:
There is little you can do about intrinsic issues, except control them as best possible and then work to minimize any potential for unwanted extras. Thus, layout is high on the list of things to do right.
Like a samurai coming at you, at the least defensible position, you must do the same with signal. Place the signal, in 3d, in the place where it is least interfered with. This, according to the hierarchy of it's potential for interference and in conjunction with it's hierarchy of importance in unimpeded generation.
Working on passive crossovers, re-doing ones that one feels are good and even redoing excellent ones (in an act of single cause analysis) is an excellent lesson to prepare for such things. (IME)
I recently found and amazing cure for some of these issues (it's a secret!
)(for the moment) (fluoride free to boot, no artificial sweeteners) (JC told me the prelude to it, then I took it a step further)
Last edited:
Hey, they use vibration to anneal large weldments, why not copper??
I'm just not sure if 4 aura shakers mounted to the plate is sufficient to force quantum tunneling through the boundaries..
j
true enough. One thing that shut the tongue waggers down in the world of audio, and audio rags and audio threads..is when John Atkinson related a story. First in the magazine (stereohile) and then on the stereophile forum.
The nay-sayers where screeching hard that day...
A story about RF engineers 'tuning' the copper bus in order to match/tune correctly. (out at a given transmitter)
They did it, this micro adjustment, by whanging on the copper bus, with a hammer.
"Yeah, but that's RF", they began to whine.
So I told them it was not a simple signal that they listened to with their incredibly complex ears (too bad about the meat between), it was and is a highly integrated and highly complex series of harmonics.
Which add up being a requirement for subtlety and correctness that exceeds the need for correctness of inter-related signals/flow in a RF chain..an RF chain and layout... that might need tuning of a bus bar with a hammer. This, Easily so.
But they continued to screech, as they did not understand. They had the screeching bit totally down, though. Gotta give them that.
(It is possible to write some of that off as oxide in threading and such, so it's not totally a clear case)(not enough data to provide exception of all other possibilities...)
Let's put it this way: Even a leaf can completely change the course of a bullet. Relational deltas become everything.
Last edited:
Yes, we do a 3D extraction of bonds and package for giga-Hz IC's. No hammers needed, but when needed an Estwing framing hammer makes the most beautiful ringing sound..
Last edited:
Huh?? that was tongue in cheek...nothing true about it at all..
That on the other hand, is an accurate depiction of rf and solid copper busswork. And it wasn't oxide in threading...rf doesn't penetrate to the bolts, it's all surface.
They were changing the geometry of the conductors to alter impedance and prop velocity, that's all.
Same thing happens in solid state microwave hybrids. Since the small signal parameters of the semi's can only be close, the traces many times require tuning, either stubs or actual blobs of silver epoxy planted on the traces in historically useable locations. That technique was unfortunately, an "art", as it was not possible to apply a smith chart to the problem.
j
Gpapag, I think you got closest to my philosophy, that it is, at least partially, the FLOW of the audio information through a circuit, that makes a serious difference.
When I try to make a WORLD CLASS product, I look at every possible factor that 'might' improve the sound quality. LAYOUT, I have found, is comparable to every other factor, including topology, and parts quality and selection.
That is why we called our company CTC. It could be interpreted as: topology, layout, parts selection, because they are the respective area of each of us, who contributed to the design and construction of the CTC Blowtorch.
In the past, we only had CT, with Vendetta. That is: topology and layout. Before that, only C, or topology, with marginal layout, and surprisingly scary (in retrospect) parts selection.
Heck, I started the Vendetta circuit with cheap, marginally designed layouts, encased in a plastic box, that was sprayed internally with conductive paint for shielding, and tantalum caps in the power supply. It would still work today, but it would not compete with a later SCP-2, any more than an early Porsche 911 would compete successfully with a 30 years later Porsche 911 in a road race.
To make a WORLD CLASS product, every aspect must be refined, or any design will lose its place in history, almost overnight, to someone who has continued to refine THEIR product.
When I try to make a WORLD CLASS product, I look at every possible factor that 'might' improve the sound quality. LAYOUT, I have found, is comparable to every other factor, including topology, and parts quality and selection.
That is why we called our company CTC. It could be interpreted as: topology, layout, parts selection, because they are the respective area of each of us, who contributed to the design and construction of the CTC Blowtorch.
In the past, we only had CT, with Vendetta. That is: topology and layout. Before that, only C, or topology, with marginal layout, and surprisingly scary (in retrospect) parts selection.
Heck, I started the Vendetta circuit with cheap, marginally designed layouts, encased in a plastic box, that was sprayed internally with conductive paint for shielding, and tantalum caps in the power supply. It would still work today, but it would not compete with a later SCP-2, any more than an early Porsche 911 would compete successfully with a 30 years later Porsche 911 in a road race.
To make a WORLD CLASS product, every aspect must be refined, or any design will lose its place in history, almost overnight, to someone who has continued to refine THEIR product.
Last edited:
That is my philosophy as well. Back in the 80's, I was designing half nanosecond differential test pin receivers for large scale ATS...I would take the schematic and color the differential signal paths green and blue, red for supply plus, black for supply minus, yellow for ground..then sketch out the layout of the ic's and transistor chips complete with chip rotation.
Then I'd give the layout guys my crayola enhanced sketches and schematics, and explain how to concentrate on making the signal paths straight and short, ground hard, and supplies hard to ground.
Of course, my hybrids were not quite as complex as your hardware, but the exact same philosophy used at rf/micro works exceedingly well for low impedance audio.
j
The art starts when you finish drawing last version of layout that you think is optimal. Then go to Frys to buy copper clad fiberglass for prototype PCB and find that longer than 10" boards are unavailable, then start thinking again to find that you can do better layout if to remove input transformers from PCBS that contain voltage stabilizers.
It happened yestarday.
It happened yestarday.
The spice of life!The art starts when you finish drawing last version of layout that you think is optimal. Then go to Frys to buy copper clad fiberglass for prototype PCB and find that longer than 10" boards are unavailable, then start thinking again to find that you can do better layout if to remove input transformers from PCBS that contain voltage stabilizers.
It happened yestarday.
not Exactly - audio doesn't have localized image current in planes, near complete isolation between routing layers separated by planes
audio frequency currents diffuse through planes - are largley resistance controlled
explicit design of low loop area/common centroid traces is required for heavy current/power in audio - can't rely on planes minimizing loop area for audio return currents
Last edited:
When it was point-to-point with tube sockets and terminal lugstrips, my father designed an FM multiplex receiver for use in background music systems. The long rectangular chassis had a flow of signal that was mostly unidirectional. The manufacturing was done in Los Angeles, and the alignment of the IF strip by one technician, who was with the various companies for many years.
When people would seek employment as assemblers they would usually be rejected if they had any knowledge of electronics. George Wood knew that they'd be tempted to change the routing of connections. He wanted people who could do the best job of precisely duplicating a sample. Otherwise all would be lost, especially ahead of the IF stages.
That is why I do my own layouts it is sometimes faster. Although when I say the problem was the damned layout guy Nelson then points out it was me.
Actually, rf and low impedance audio are virtually identical.
Localized image currents are part and parcel a problem with analog...they are responsible for analog paths seeking the lowest impedance return path. The analog 20hz path may not be the analog 10Khz path, especially if the ground plane is extensive. Image currents are responsible (lenz's law) for causing this. So low impedance audio currents are absolutely NOT controlled by the resistive path, because as the frequency goes up, it's the impedance that controls it.
How many analog boards require enough layers to require planar isolation? And do not forget, an isolation plane in 2 or 3 oz copper is darn near transparent to 20 hz current generated magnetic fields.
We are in agreement here..loop area is controlled by where the current centroid lies.
I will grant you, not many understand it well enough to realise this.
ps...you do realize I did not mention planes. There was a reason..
pps. The blue hilited statement is one of the most interesting errors some people fall for. For example, ground loop induced currents going from an input rca jack to the IEC ground bond will take the lowest impedance path at any frequency other than zero where it will take the lowest resistance one. As the loop current rises in frequency, the lowest impedance path may not be the most direct. It is actually this effect which Bill Whitlock missed in his papers.
j
ppps...we are still fighting vacuum tube understandings of high z circuitry..
Last edited:
I remember a story, on radio plant in Riga, Lituania, they got a party of receivers that did not work at all. Investigation revealed that a lady in warehouse who worked there many years tried what she believed to correct an error: new capacitors were of different color, she thought it is wrong, so she replaced them by other capacitors from different bin.
She knew nothing about picofarads...
- Status
- Not open for further replies.