[jneutron]

Even more suprising is the continued disparity between tube and solid state "sound", when the reasons are so easy to see.

You trying to ignite WW3? As if wires were not contentious enough
This is the part where I usually get the death threats, so I think I'll bow out now before the eruption.

All the best,

- AJ

Hey...as far as I've seen, you've been rather civil.  You pointed out how far the thread got without the word moron.

I'm not afraid of tubers, nor discussing them...and they are, after all....jewelry, aren't they?

Cheers, John

[BrianM]

Even more suprising is the continued disparity between tube and solid state "sound", when the reasons are so easy to see.

Hi, can you explain what exactly you mean by this? Thanks..

The bulk of the circuitry internal to a tube amplifier is high impedance.  Low currents..I recall my 6l6 beast with 6600 ohm ct transformer impedance.  At these impedances, very little magnetic field is created even though the output is a whopping 55 watts rms..

Contrast that with a solid state beast.  The full speaker currents are inside the chassis, free to do what they will....and they will.

If you apply amperes law to the output rail currents, you can derive the magfield caused by it.  This magfield is time varying...apply faradays law of induction to any circuit loops within the chassis, and you can derive the induced voltage that will permeate the circuits.  I did this a while back, don't recall which forum I posted it on...but with a 10 ampere output current, induced voltages were in the millivolt range at the square centimeter area, at a khz or so.  This kind of induced error will not be found using an FFT, as it's only phase error, and slight.  But it gets complex when there is a smaller hf signal riding on a large bass one, as the induction back to the front end of the amp will be dependent on the quadrant the output stage is running...the positive rail and the negative rail of the output source will never broadcast the same.

Star grounding techniques are absolutely great for tube stuff, as well as any high impedance circuits.  But for low z solid state, it sucks.  I've had to re-think quite a few chassis and widgits because star grounding is not the panacea all think it is.

If one were to look at standard rf techniques, one could see how they've solved the problem.  But even at low frequencies, like audio, I have felt the "pain" of induced voltages causing errors.  Many RF techniques are extremely useful in audio, but not many understand how.

Cheers, John

I was hoping for English, but thanks anyway.

[jneutron]

I was hoping for English, but thanks anyway.

Ah sheesh, sorry about that.

Whenever a wire carries current, it makes a magnetic field.  Audio currents make a magnetic field that changes with time.

A loop of wire will see a changing magnetic field, and it will generate a voltage as a result of that.

Audio amplifier designers don't particularly watch out for the magnetic fields that are caused by the high currents, and they don't watch out that the input side of the amp can be affected.  Some just stuff the wires in the box.

With tubes, the currents are much lower.  It's only on the other side of the output transformer that the currents are big.

Hope that helped more..

Cheers, John

[BrianM]

I was hoping for English, but thanks anyway.

Ah sheesh, sorry about that.

Whenever a wire carries current, it makes a magnetic field.  Audio currents make a magnetic field that changes with time.

A loop of wire will see a changing magnetic field, and it will generate a voltage as a result of that.

Audio amplifier designers don't particularly watch out for the magnetic fields that are caused by the high currents, and they don't watch out that the input side of the amp can be affected.  Some just stuff the wires in the box.

With tubes, the currents are much lower.  It's only on the other side of the output transformer that the currents are big.

Hope that helped more..

Cheers, John

A little, thanks.  Do you favor some kind of hybrid design then, or am I misreading?

[jneutron]

A little, thanks.  Do you favor some kind of hybrid design then, or am I misreading?

No.  I favor the elimination of as much of the internal magnetic fields as is possible.  For example, when the amplifier is pushing the speakers with positive voltage, the positive power supply wire is carrying the current, and the chassis or ground wire carries it back.  Because this is a loop, it makes field.

Good design can get rid of most of the loops.

Cheers, John

[BrianM]

A little, thanks.  Do you favor some kind of hybrid design then, or am I misreading?

No.  I favor the elimination of as much of the internal magnetic fields as is possible.  For example, when the amplifier is pushing the speakers with positive voltage, the positive power supply wire is carrying the current, and the chassis or ground wire carries it back.  Because this is a loop, it makes field.

Good design can get rid of most of the loops.

Cheers, John

Thanks, was just trying to tie it back to your comment about tube vs. solid state *sound*, which I'm still not sure what to make of.

[DaveC113]

I think he expressed it pretty well right there. Yes, they make a difference but what is the difference and is it something you want? Perhaps you can EQ anything a cable does for a lot less money and a lot more convenient experimentation.

What about the time alignment of the signal? Could too much capacitance in an IC have an effect here?  

Decent ICs don't need to cost much. Zu Gede is very good at $199/meter. I'm sure there are others.

That seems expensive to me... but I can see how a cable can cost that much if you're using complicated geometries and expensive materials.

I think simpler solutions can perform just as well and cost a lot less money. If you're spending more than a few hundred on ICs, you're getting taken for a ride, IMO.

[Russell Dawkins]

And then there's the mixdown at the studio...sigh..
Cheers, John

I'd be curious to know your concerns regarding this.

[Scott F.]

John,

Great series of posts 

Question (and I'm going to show my arse here so be kind), how do you go about laying out an amp or pre to reduce the effect of the magnetic fields? Is it as simple as twisting your B+ and B- and grouping your passive parts close together throughout? Or is it more just knowing that you have a magnetic field potential and making sure that you have minimal passives located within that field?

That leads to the next question (or observation), your power supply transformer (one would think) puts off one heck a a magfield. Would it make more sense to locate your PS transformer and Pi filter in the back of a chassis (I'm sure a separate box would be even better) and then have your output transformers towards the front of your chassis and finally locating your inputs (the smallest signals) somewhere away from both sets of transformers?

Then I guess the next question would be, how effective would it be to use shielded wire on the internals and draining the shield? Or would that only drain RF?



PS, I'm glad this thread has turned into something educational rather than the same old "you're deaf...no you're stupid" thread.

John, thanks for taking the time to teach rather than preach 

[Daygloworange]

If you're spending more than a few hundred on ICs, you're getting taken for a ride, IMO.

Depending on what your personal value system is. The same could be said of amplifiers. There are amps that sell for a couple of hundred dollars that will perform respectably, all the way up to 10's of thousands of dollars. Same with CDP's, Pre-amps, and of course speakers.

At the end of the day, it's really is a personal thing.

Cheers

[DaveC113]

If you're spending more than a few hundred on ICs, you're getting taken for a ride, IMO.

Depending on what your personal value system is. The same could be said of amplifiers. There are amps that sell for a couple of hundred dollars that will perform respectably, all the way up to 10's of thousands of dollars. Same with CDP's, Pre-amps, and of course speakers.


Cheers

Heh, well... amps and most other components have a lot more parts than ICs. Even Ultra Pure OCC silver isn't that expensive in sizes useuful for making ICs out of ($4/ft at VHAudio). I can comprehend spending a few grand on parts to make a top of the line amp, but materials for an IC just aren't going to add up that much.     

[jneutron]

And then there's the mixdown at the studio...sigh..
Cheers, John

I'd be curious to know your concerns regarding this.

We humans determine where a sound is coming from using two things.  Difference in amplitude between ears, and difference in time delay between ears.

If you adjust the tape head angle on a cassette player as it plays into headphones, you can get a very good alignment by adjusting it until all the image is in the center of your head, and not to one side or the other.  This is an adjustment of time delay with constant amplitude.

If you use the balance pot to move the image of your sound from one side to the other in the headphones, that is amplitude adjustment with no change in time delay.


Humans use BOTH.  Not one, or tuther...but both.

Putting an image in a precise location requires adjustment of both.  A studio mixdown setup does not adjust the time delay from one channel to the other.  If that were done, the resulting product would not be mono compatible.  Any interchannel delay would comb like heck..

So we make do with amplitude only panning.  Unfortunately, we also have a frequency based dependence on amplitude panning...some frequencies require more delta amplitude than others for the same spacial location.  For some flavor on this, look up Griesinger..pan laws pdf..it's a start.

True spacial construction requires a lot more horsepower at the studio...and, it will only be applicable to one speaker placement geometry.  What does one use as the standard?

Me, I just enjoy the music.

Cheers, John

[jneutron]

John, Great series of posts 

No prob.

Question (and I'm going to show my arse here so be kind), how do you go about laying out an amp or pre to reduce the effect of the magnetic fields? Is it as simple as twisting your B+ and B- and grouping your passive parts close together throughout? Or is it more just knowing that you have a magnetic field potential and making sure that you have minimal passives located within that field?

I attack both.  For the power rails, triaxially constructed wire is the absolute best, with the center shield being the ground.  Use this to get the current from the capacitors to the output stage.  No matter what the output current is, this triax wire will not have an external magnetic field at audio frequencies.  Flat copper ribbon can also work well, but once it is made, it can't be bent much as the copper will kink on the inner corner of the bends. Flat copper braid will work very well here.

As a bonus, if you choose the sizes rignt, you can keep the supply feed 8 ohm nominal impedance, and the rail inductance will be in the 10 nanohenry per foot range, with the feed rail at 300 pf per foot.

Good layout of the passives helps too.

That leads to the next question (or observation), your power supply transformer (one would think) puts off one heck a a magfield. Would it make more sense to locate your PS transformer and Pi filter in the back of a chassis (I'm sure a separate box would be even better) and then have your output transformers towards the front of your chassis and finally locating your inputs (the smallest signals) somewhere away from both sets of transformers?

For tubes, it's not as critical.  Lotsa people talk about how to arrange the transformers for minimal hum pickup.

Then I guess the next question would be, how effective would it be to use shielded wire on the internals and draining the shield? Or would that only drain RF?

At audio frequencies, shields are only good for electrostatic fields.  Magnetic fields tend to go right through shields.  At 60 hz, you'd need shielding about two inches thick.  Best thing is to not trap flux in any loops.

PS, I'm glad this thread has turned into something educational rather than the same old "you're deaf...no you're stupid" thread.

Yah, they tend to get tiresome..

Cheers, John

how does one attach an image??

[Daygloworange]

True spacial construction requires a lot more horsepower at the studio...and, it will only be applicable to one speaker placement geometry.  What does one use as the standard?

Cheers, John

Correct. Sadly, there is no industry standard.

Heh, well... amps and most other components have a lot more parts than ICs. Even Ultra Pure OCC silver isn't that expensive in sizes useuful for making ICs out of ($4/ft at VHAudio). I can comprehend spending a few grand on parts to make a top of the line amp, but materials for an IC just aren't going to add up that much.    

That's a moot point in regards to what I'm talking about.

My point is ratio of dollars spent vs sonic improvement. Not justifying in terms of parts count or complexity. Although I do appreciate your point, you forgot to factor in economy of scale on amps and so forth vs high end PC's, IC's, SC's which are often times handmade one pair at a time by small business entities.

jneutron,

Thanks for the learnin' ! 

Cheers

[BrianM]

Nice discussion, all that's lacking would be relating it to the real world somehow.

[jneutron]

Nice discussion, all that's lacking would be relating it to the real world somehow.

he he...welcome to my world.

This discussion is certainly real world.  Perhaps not yours, but certainly mine.  It will eventually impact what you buy, what you listen to.  Not now, maybe ten years from now.

I'm used to that.  What I do normally is kinda esoteric, and quite beyond normal experience.  Geek heaven..

Cheers, John

ps...be back sept 6..

[dado5]

Just for the record, I used the term 'moron' to refer to a very small group of people. And moreover it is a reference to their activity outside of the audio realm.

As I said, folks are free to hear what they want, buy what they want and talk about what they want. It's those few who try to use the monopoly of violence that is the state to stop this freedom of activity, speech and association who rightfully deserve the title 'moron'.

clear 'nuff then.