So what in solder "sounds" bad, and what can be done about it?

[Mike Sulzer]

Here's a quick and dirty linear approximation:

Ractual = R * (1 - (.0002 * V) - (.0035 * (V/Vmax))
where:

• R = resistor value
• Ractual = calculated resistance at voltage V
• V = applied voltage
• Vmax = maximum continuous voltage, usually 250VDC for 1/4 watt units

The coefficient (-0.00002) is somewhat smaller than what plot in Horiwitz implies for 1000v, but that makes sense for the lower voltages we are dealing with.

That equation looks too complicated for a linear approximation. If we have an R biased by a dc current so that the current always flows in one direction, the we just have Ractual = R - 0.00002V, where V is the ac voltage, that is, a signed quantity, the deviation from the bias voltage of the stage. This causes the resistance to increase for V with one sign, and decrease for the other. This makes second harmonic, etc.. If you put a large ac voltage across an unbiased resistor, the resistance decreases for a swing in either direction, and you have odd harmonics only. You could write that this way: Ractual = R - 0.00002|V|.

[Mike Sulzer]

So far I have seen no hi-end hi-fi or guitar amp guru discover anything that advances the art, only misleading marketing hype, used for the express purpose to mislead customers.

The importance of a well-regulated power supply is discovered through listening. Obviously you can measure the effects, but that will not show you how important they are.

[No Mutant Enemy]

He got it back and reported to me that it lacked weight and mastery. Now what on earth part of the circuit controls that?

Do your amps not have a Mastery control? All of mine do. And they all go up to 11.

[Mike Sulzer]

So the worlds of "sounds appealing" clashes with "sounds accurate" in every step of music production and reproduction. Assigning too much importance to engineering is folly. Engineering is great in its own right, and leads to better devices, more reliable, more predictable and often even cheaper but it is not so important in sound creation.

I could not disagree more. For hi fi amps "more accurate" is better in the sense of "sounds better" once you get past a certain threshold of accuracy, and that threshold was passed long ago. That is why a really good solid state amp is superior to any tube amp: you cannot make the distortion low enough in a tube circuit.

[km6xz]

I was not saying re-production amps are open for such subjective evaluation as production amps. That was my whole point, in a re-production amp there is a real world goal and getting to it or not can be determined objectively. Production amps has no real world reference to compare to, no degree of faithfulness to aspire to. Tubes or any active device perfectly suitable for producing a sound that could be pleasing, it is not trying nor can it be, a faithful representative of a source signal since there is no self complete signal for comparing.
Tubes in hi-fi are not faithful reproducing elements but some people really like their unfaithfulness, how it alters the reference signal. But it really does not fit the definition of high fidelity but closer to the concept of a sound producer than sound re-producer.
My long message was just to show that there are two distinct intents for amplification, one has nothing to reference to and the other one has very well refined references to compare with. Neither one is even remotely a suitable substitute for the other.

[Mike Sulzer]

I was not saying re-production amps are open for such subjective evaluation as production amps. That was my whole point, in a re-production amp there is a real world goal and getting to it or not can be determined objectively. Production amps has no real world reference to compare to, no degree of faithfulness to aspire to. Tubes or any active device perfectly suitable for producing a sound that could be pleasing, it is not trying nor can it be, a faithful representative of a source signal since there is no self complete signal for comparing.
Tubes in hi-fi are not faithful reproducing elements but some people really like their unfaithfulness, how it alters the reference signal. But it really does not fit the definition of high fidelity but closer to the concept of a sound producer than sound re-producer.
My long message was just to show that there are two distinct intents for amplification, one has nothing to reference to and the other one has very well refined references to compare with. Neither one is even remotely a suitable substitute for the other.

I see what you mean now, and I agree completely.

[Steve Conner]

Do your amps not have a Mastery control? All of mine do. And they all go up to 11.

Sure they do! Right next to Douglas Self's "Niceness Knob". But this wasn't one of mine, it was an old Electrocompaniet.

You have to ask yourself where the line between sound production and sound reproduction is, when nowadays you can buy electronic music that was produced entirely out of DSP algorithms inside a laptop, right up to the finished master. Where's the "real event" that is being reproduced?

Also, given room acoustics, sometimes it would take a decidedly weird output from the speakers to produce an "accurate" sound at the listening position. What does accuracy mean then? I guess it means setting your living room up the same as the mastering room where your favourite tracks were mastered.

I prefer the post-modern view that a hi-fi is part of the sound producing apparatus, and you just tinker with it to get a sound that pleases you, on the kinds of music you like to listen to. But I still agree that power amp distortion probably isn't an effect you want to add. (Unless the album was mixed through a Dynaco Stereo 70! )

[No Mutant Enemy]

I guess it means setting your living room up the same as the mastering room where your favourite tracks were mastered.

And adjusting the frequency response of your ears and auditory cortex to match the original. At some point we have to accept that what one person hears may not be what another person hears, and that neither is necessarily wrong.

[salvarsan]

The coefficient (-0.00002) is somewhat smaller than what plot in Horiwitz implies for 1000v, but that makes sense for the lower voltages we are dealing with.

Helps if you get the coefficient right, too. That's 0.02% = 0.0002 = 2.0E-04, okay?

That equation looks too complicated for a linear approximation.

"Linear" meaning "may be plotted in two dimensions", not linear = straight line.
I'm sure you can cruft up a good virial equation for triode circuit simulation.
My intent was to get the ball rolling on this, not to provide a definitive answer, hence the bone-head simple eqn.

If we have an R biased by a dc current so that the current always flows in one direction, the we just have Ractual = R - 0.00002V, where V is the ac voltage, that is, a signed quantity, the deviation from the bias voltage of the stage. This causes the resistance to increase for V with one sign, and decrease for the other. This makes second harmonic, etc.. If you put a large ac voltage across an unbiased resistor, the resistance decreases for a swing in either direction, and you have odd harmonics only. You could write that this way: Ractual = R - 0.00002|V|.

What about that very visible term where the dR/dV increases with voltage? I keep coming back to an eqn like:

Ractual = R (1 - dR/dV*V - d²R/dV²* V)

[salvarsan]

Here's a stupid question for the resident empiricists... how exactly do you measure harmonic content for even order vs. odd order, which harmonic, and whether it is distorted or not? I feel like I'm missing something.

(before you jump on me, bear in mind I prefaced it saying that it was a stupid question)

Not stupid, but important and fundamental.

These days, you put the signal through a spectrum analyzer and look for the harmonics.
Anything that wasn't there at the beginning but appears after passing though your circuit of interest is called distortion.

Tube amps often have a little 2nd harmonic distortion, as in, an A440 tone gets 0-2% A880 added in. Sounds okay.
If the distortion is 3rd harmonic, it is very audible and not always pleasant, gives a guitar a somewhat congested sound like the Arbiter Fuzzface.

Here's an example of oscilloscope traces and spectral traces:
More on spectrum analysis : MIXED-FREQUENCY AC SIGNALS

[km6xz]

Comparing a source signal, normally a pure sine wave, to the output and measuring the amplitude of each of the harmonics individually with a wave meter, or collectively with a distortion analyzer. Spectrum analyzers give amplitude versus frequency domain displays to show it graphically. A sine wave does not need to be used, a complex wave form can also be used by comparing the original signal to the output by normalizing the level and subtracting one from the other. This also gives an indication of a group delay type of distortion where the phase of harmonics are altered which results in different products when adding the component harmonics to get the complex wave form than would be expected if the harmonics are in proper phase relationship with the predicted ideal. We can hear the difference and measure it.
A differential comparator like mentioned above is also a great way to check for the influences of different types of passive components like resistors and capacitors. A battery, a couple resistors and an op-amp is all that is needed to see the change between the input and output of an amplifier.

[FunkyKikuchiyo]

Do your amps not have a Mastery control? All of mine do. And they all go up to 11.

Reminds me of a guitar customer that brought in a classical saying it lacked its "transcendence". I suggested he call a witch doctor. "Which doctor?", he asked. "Yes, witch doctor!" I responded.

Actually, no... I just thought it... not said it... like most witty things I think of.

km6xz... I knew a guitar teacher years ago that would often remind his students that they're playing the amp as well as the guitar, generally to encourage them to practice plugged in rather than unplugged. You're right where he was right, too... electronics with guitars are extensions of the instrument. I like your production/re-production paradigm, too. I may steal that.

I will say engineering is very helpful in instrumentation design, though. Mike said a bit earlier something to the effect that "I hear a difference between two caps because one sounds like a lower value"... he couldn't make that statement if he wasn't both using his ears and his brain - the two informed each other to help him sort things out. In the same way we need to understand individual instruments to hear an orchestra and pick them out, I think the more we understand these basic phenomena, the better our ears become. If we didn't know the difference between a clarinet or a flute, the two playing in unison would just be one sound to our ears. Once we understand the two instruments, our brain can separate them. It is the same way we on this board can hear the difference between a strat pickup and a humbucker, when the strat pickup is in a strat and the humbucker is in a Les Paul. We understand the elements and our brains have figured out how to separate them.

How many people here actually read the hi-fi magazines that get so much grief on this board? I get the feeling that the nuttier than a squirrel turd hi-fi folks are punching bags here and invoked in discussions as an argument ad hominem, but I don't think anyone here really is of that mindset or practice. I really don't think I've encountered this anywhere except peculiar websites.

[Mike Sulzer]

Helps if you get the coefficient right, too. That's 0.02% = 0.0002 = 2.0E-04, okay?

"Linear" meaning "may be plotted in two dimensions", not linear = straight line.
I'm sure you can cruft up a good virial equation for triode circuit simulation.
My intent was to get the ball rolling on this, not to provide a definitive answer, hence the bone-head simple eqn.



What about that very visible term where the dR/dV increases with voltage? I keep coming back to an eqn like:

Ractual = R (1 - dR/dV*V - d²R/dV²* V)

Yes, I got an extra zero in there, sorry. But my statement that the coefficient agrees with the one from the Horowitz figure is right. From the figure one derives about -0.00029 rather than -0.0002, but it is based on a higher voltage, and the effect is not completely linear with voltage, as you are saying. So the number you are using in the linear term (-0.0002) makes good sense. That is all I said. Not sure I understand your reaction.

When we are talking about approximations, linear does mean "straight line": we have a coefficient which we multiply times the voltage (to the first power) to get the change in resistance. The next term would be quadratic, that is shaped like part of a parabola. That term would be this: d²R/dV²* V² . The V² is necessary to make the units work out so that the result of the computation of that term has units of resistance.

A linear approximation is probably good enough to show the effects we are looking for.

[David Schwab]

Reminds me of a guitar customer that brought in a classical saying it lacked its "transcendence". I suggested he call a witch doctor. "Which doctor?", he asked. "Yes, witch doctor!" I responded.

When I get those kind of customers who make statements like that, I ask them to define the word they just used, or explain what the hell they are talking about! Then I can explain what they are, or are not hearing.

I will say engineering is very helpful in instrumentation design, though. Mike said a bit earlier something to the effect that "I hear a difference between two caps because one sounds like a lower value"... he couldn't make that statement if he wasn't both using his ears and his brain - the two informed each other to help him sort things out. In the same way we need to understand individual instruments to hear an orchestra and pick them out, I think the more we understand these basic phenomena, the better our ears become. If we didn't know the difference between a clarinet or a flute, the two playing in unison would just be one sound to our ears. Once we understand the two instruments, our brain can separate them. It is the same way we on this board can hear the difference between a strat pickup and a humbucker, when the strat pickup is in a strat and the humbucker is in a Les Paul. We understand the elements and our brains have figured out how to separate them.

Exactly. You have to think about what you are hearing if you want to figure out why something is doing what it's doing. Sometimes that's not so simple.

How many people here actually read the hi-fi magazines that get so much grief on this board? I get the feeling that the nuttier than a squirrel turd hi-fi folks are punching bags here and invoked in discussions as an argument ad hominem, but I don't think anyone here really is of that mindset or practice. I really don't think I've encountered this anywhere except peculiar websites.

I appreciate high end audio gear, but when people buy $600 wooden knobs for their amp because it's going to tighten up the low end and remove vibrations, you have to question their sanity.

Other nutty things I have seen are a lacquer you paint on your audio gear to improve the tone, another vanish that you paint on plastic cased ICs to make them sound like vacuum tubes, and those holographic stickers to put on your CDs to improve the tone. Also anyone who thinks that vinyl sounds better has hearing issues.

[RedHouse]

...I appreciate high end audio gear, but when people buy $600 wooden knobs for their amp because it's going to tighten up the low end and remove vibrations, you have to question their sanity...

Everybody needs a good knob-job

It's the guys that buy this kind of shite that crack me up:



Wattgate 381 Audio Grade Duplex Receptacle Outlet

If they spend that kind of cash on that kind of equipment expecting their audio quality to rise, well, PT Barnum was looking for them on their way in.