Capacitor radiation

[Sebastiaan]

Dear,

In various books and topics I read about keeping your VAS isolated and as far as way as possible from power devices and parts.

Let's say you want to keep your power lines really short, and you mount your elco capacitor bank close in front of the power stage. On vertical axis, in fron of this capacitor bank you place your VAS circuit. This way you have a very shirt power supply path, and you separate the VAS from the power. However, the elco's are in "between". How far will those elco's radiate and magnetically couple energy to the VAS?

With kind regards,
Bas

[wakibaki]

Originally Posted by Sebastiaan

How far will those elco's radiate and magnetically couple energy to the VAS?

They won't. Magnetic components such as transformers or chokes may interfere with nearby components. Don't be too paranoid about the whole issue. Try your layout out roughly before committing to building the whole thing into a box if you can. Be careful you don't short anything and destroy what you've built.

w

[Bigun]

I worried about leakage from the main trafo, high current charging pulses between trafo and rectifiers and rectifiers to psu caps. But even when I 'waved' my amp pcb in front of these things there didn't seem to be a problem. However, getting the signal input anwhere close to the amplifier output was a sure way of getting nasty oscillations and a blown fuse ! All I had were the two wires running parallel to each other 1 cm apart for about 10 cm and it was off the races - thankfully only a temporary set up.

[Sebastiaan]

Originally Posted by wakibaki

They won't. Magnetic components such as transformers or chokes may interfere with nearby components. Don't be too paranoid about the whole issue. Try your layout out roughly before committing to building the whole thing into a box if you can. Be careful you don't short anything and destroy what you've built.

w

Dear Wakibaki,

I am not paranoid But I discovered the obvious. That a good PCB layout is almost everything in an amplifier, and much more important then the endless search for "high-end" components.

Low band-witdh audio circuit are general very forgiving, so it will often simply works and produce wonderful sounds, However during many experiments making different PCB's and keep measuring I found out that there is so much to improve to the edge.

I discovered for example

1: If your "source" and "sink" traces from your VAS to your EF stage run to close to each other, THD+N get worse

2: If you place the VAS in front of the output devices, THD+N get worse

3: If you place the whole circuit to close to the Bridge rectifier THD+N get worse significant

4: If you put the positive and negative power supply traces all the way to the edge of the left and right side of the PCB (in mirror) THD+N get worse. Instead If you put them together in the middle of the PCB and run the + and - traces close together THD+N improve.

5: A ground-plane as split between the output trace and the power supply trace, and the line traces improve THD+N

THD+N on paper is nothing for some, but it shows that big improvements can be made.

Next step is to further experiment with a solid unbroken groundplane for ALL grounds.

With kind regards,
Bas

[wakibaki]

OK Bas

My 45 as a years as a musician, design experience of GSM and 3G basestations, VOFDM broadband communication systems, GPS based SOTDMA, low phase-noise oscillators, antenna design, control systems, instrumentation, EMC test and general electronics are probably nothing compared to your experience of baseband amplifiers so just forget what I said...

w

[Sebastiaan]

Originally Posted by wakibaki

OK Bas

My 45 as a years as a musician, design experience of GSM and 3G basestations, VOFDM broadband communication systems, GPS based SOTDMA, low phase-noise oscillators, antenna design, control systems, instrumentation, EMC test and general electronics are probably nothing compared to your experience of baseband amplifiers so just forget what I said...

w

Dear Wakibaki,

Why so offended? Did I stated anywhere you are wrong? I didn't. I simply state that more carefull PCB layout can improve performance. I don't see what is wrong with that.

And for your record, I am a musician myself and not and engineer (graduated cum laude conservatory), and I live for music only. Therefor my fanatic search for the best possible audio designs out of love for music. I love music and music theory much more then any electronic theory, but I am fascinated by sound reproduction, so therefor it is a must to learn more theory.

With other words, I am here to learn from the experts and that means also from you.

However I am a very practical minded person and I measure and experiment a lot to learn. I did nothing more then share my experiments in above post, so again I don't understand why you feel offended You can simply tell me where you think I am wrong and or right, and I accept your criticism.

With kind regards,
Bas

[wakibaki]

Your question in the first instance was virtually meaningless:

'How far will those elco's radiate and magnetically couple energy to the VAS?'

How long is a piece of string?

What I'm saying is: your question is that of a novice seeking reassurance. That is what you got, reassurance. Your concerns are exaggerated and misplaced.

While capacitors are capable of radiation the effect is negligible. Particularly at the frequencies involved in audio. You will find many references in electronics texts detailing the measures necessary to reduce the interference to other components by magnetic components, but none referring to capacitors.

If I had known that all you were seeking was an opportunity to lecture me on the positioning of power traces on a PCB, I would not have bothered to reply in the first place.

w

[Sebastiaan]

Originally Posted by wakibaki

Your question in the first instance was virtually meaningless:

'How far will those elco's radiate and magnetically couple energy to the VAS?'

How long is a piece of string?

What I'm saying is: your question is that of a novice seeking reassurance. That is what you got, reassurance. Your concerns are exaggerated and misplaced.

While capacitors are capable of radiation the effect is negligible. Particularly at the frequencies involved in audio. You will find many references in electronics texts detailing the measures necessary to reduce the interference to other components by magnetic components, but none referring to capacitors.

If I had known that all you were seeking was an opportunity to lecture me on the positioning of power traces on a PCB, I would not have bothered to reply in the first place.

w

Dear Wakibaki,

Gut oh gut.... what short toes. I wasn't lecture you at all, I shared my experience, in order to get some feedback. (a challenge to correct me) You total got my intention wrong, and see it as a personal attack, which is a pity.

My whole point was to indicate, that with placing VAS on different locations change my measured results. My attempt was to indicate where my question came from. And I did read in the books about capacitor radiation, and all I was wondering about if it has any influence in audio as well. Which isn't a stupid question at all.

Especially knowing that there people here who mold capacitors into bamboo pipes, remove foil from them and so on.

And please keep your manners and don't put stamps on me! I keep your in your value as well. You don't know anything about me, neither my experiences.

With kind regards,
Bas

[wakibaki]

OK Bas, pax

Here's the one (kind of) exception I can recall with regard to capacitors, which has to do with hum pickup, rather than radiation:

Morgan Jones - Valve Amplifiers - P218

Polystyrene capacitors often have one end delineated by a red or yellow band. This does not mean that they are sensitive to polarity, but that the marked end is connected to the outer foil. This is significant because the marked end of the capacitor may be connected to a less sensitive part of the circuitry than the other. For instance, if a small polystyrene capacitor was used as part of an active crossover network, and connected as a series coupling capacitor (high-pass filter) then the banded end should be connected to the source to reduce induced hum. Alternatively, if one end were connected to ground, then the banded end should go to ground to reduce stray capacitance to active signals (strays to ground rarely cause problems, but Miller effect can cause other strays to be significant).

---------------------

In general, capacitor radiation is not significant.

Although many people here like to think that squeezing the last 0.001% THD out of a layout is worthwhile, I think it is more important to focus on the music. I have had a lot of enjoyment out of listening to music, in the past sometimes played on systems which were not too good, and I often get impatient with people who I think are too focussed on equipment (and often how much it cost). You don't need 'the best' equipment, you can hear the music past the system if you are not obsessive.

I like to try to encourage DIY beginners to make some progress and build a working system, instead of worrying too much about the sound. Most amplifiers nowadays have a perfectly good sound, good enough that the speakers have a much bigger effect on the final result than the amplifier.

Anyway, that's basically where I'm coming from.

w

[Sebastiaan]

Originally Posted by wakibaki

OK Bas, pax

Here's the one (kind of) exception I can recall with regard to capacitors, which has to do with hum pickup, rather than radiation:

Morgan Jones - Valve Amplifiers - P218

Polystyrene capacitors often have one end delineated by a red or yellow band. This does not mean that they are sensitive to polarity, but that the marked end is connected to the outer foil. This is significant because the marked end of the capacitor may be connected to a less sensitive part of the circuitry than the other. For instance, if a small polystyrene capacitor was used as part of an active crossover network, and connected as a series coupling capacitor (high-pass filter) then the banded end should be connected to the source to reduce induced hum. Alternatively, if one end were connected to ground, then the banded end should go to ground to reduce stray capacitance to active signals (strays to ground rarely cause problems, but Miller effect can cause other strays to be significant).

---------------------

In general, capacitor radiation is not significant.

Although many people here like to think that squeezing the last 0.001% THD out of a layout is worthwhile, I think it is more important to focus on the music. I have had a lot of enjoyment out of listening to music, in the past sometimes played on systems which were not too good, and I often get impatient with people who I think are too focussed on equipment (and often how much it cost). You don't need 'the best' equipment, you can hear the music past the system if you are not obsessive.

I like to try to encourage DIY beginners to make some progress and build a working system, instead of worrying too much about the sound. Most amplifiers nowadays have a perfectly good sound, good enough that the speakers have a much bigger effect on the final result than the amplifier.

Anyway, that's basically where I'm coming from.

w

Dear Wakibaki, peace and sorry from my side if in any way I offended you.

I am fully agree with you. Once you get absorbed by music it all doesn't matter anymore. And there is no bigger kick then listen to your own constructed amplifier! But I think it is in our nature to always searching for the "unknown". Hence that is why "audiophiles" change their equipment once in a while. The same for many musicians. We are always searching to obtain a certain sound out of our instrument. One like a Ken Smith bass, others like a Fender Jazz Bass. We try to find a certain sound we can identify ourselves with.

I am also agree with you as well (at least I trapped myself on it) that sometimes you can get to obsessed or absorbed by a certain design aspect, and as designer there will be moments that u get to obsessed by chasing a certain spec, that you forget the rest. I had such a moments, I tried everything to obtain a high slew-rate and perfect square-wave reproduction. Working for weeks (if not months) on that aspect unconsciously I was creating a sterile bad sounding product that totally lack in so many other aspects.

The listening experience when you give a listen to various circuits often answer in 5 seconds already what is is a go and what is a no go.

Now I am stucked in such an obsession. I am fortunate that my circuits are reliable, working well, and yes even pass the EMI and EMC tests and simply sound good. Then you get curious to new facets in design. and try other ways to improve (or degrade ) things. I am such a person who never can't be satisfied and always think "it can be be done better"

Often I called some audiophiles fools, and a few years ago I thought all those chip-amp fans are fools for example. Shame on me! that close-mind attitude of me made me miss a lot of beautiful opportunities in HIFI. And since a few times I got so surprised by things that I considered foolish, I open my mind. That is how I came to such a questions here on this very respectable forum.

I realize something. Audio equipment is no medical equipment, and audio and music is emotion. Often we hear things we can't explain. It are forums like this where I can gratefully discover electronic issues audio related and sound quality orientated. All by the forum members experiences. Things that are often hard to find in the books.

Btw. The reason I came with my first question was after seeing a Halcro Amplifier, where all those circuits are shielded in separate chambers. I just often become curious why certain designers (besides marketing) choose certain solutions.

Thank you for your answers and have a nice evening

With kind regards,
Bas

Ps. This wasn't a lecture but a share of my personal thoughts