The Secret of Tube Amplifiers Revealed - and much more!

[ThorstenL]

Hi,

Originally Posted by poptart

Thanks for the explanation. Yes I was definitely thinking in a different ballpark than a resistor when you said "mechanical" damping.

My suggestion is electrical. But it has the merit that existing drive units are easily adapted.

The original Eckmiller driver had a classic Spider (that is one with legs - where the name originated - not a flat preformed colth disk) and each leg of the spider was damped using a grease-pot style damper with the addition of aperiodic damping of the rear cone by having a basket a lot like a sieve.

Modern polymers can probably offer other options of resonance damping in a moving system as well.

The point I was trying to make however is that we do not need to use the Amplifier to damp the cone oscillation at resonance and other then reducing drive unit cost doing so (as is common) actually does not provide any performance benefits, but many dis-benefits.

Yet the practice being established we are caught in a vicious circle.

Practically all Speaker makers make speakers designed for voltage drive. So if I make Amplifiers and wish to stay in business I need to make Amplifiers that drive these speakers well. Anything else is financial suicide. But with all Amplifiers being made to offer voltage drive, they will not drive a speaker designed for current drive well at all. As there no amplifiers to drive such speakers making any would be strongly prejudicial to a company's satying in business. So.... <return to the beginning of the paragraph and keep reading>

This is why I gave my various experiments in the 80's on how to improve speakers "the active way" (read MFB, Current Drive etc.) eventually and instead worked on how to get good sound in the context of the systems as they are, as opposed to "how they should be".

Ciao T

[keantoken]

I finally found a speaker I ripped out of a karaoke machine today, and decided to see how it sounded. I also decided to test out series resistors.

I was surprised by the quality of the speaker. It has a black cloth surround, treated with what I guess is some sort of drying oil, like boiled linseed oil. Black dust cap (mylar?).

The surround is large enough to accommodate decent excursion, and as such toms sound good. Deeper bass than you would expect from a 3.5" speaker.

I wouldn't hesitate to say it was still a piece of crap if it didn't sound better than my $80 Kevlar cones... Unfortunately, I only have one because the karaoke machine only had one speaker...

Using a 25 ohm rheostat as a series resistor, it seems to sound less militant. The sound doesn't bite so much, things sound smoother. I don't know if this is because of equalization changes. Speech is more legible. It is easier to relax while listening. (this is out of a single speaker, mono)

Instruments don't obscure each other so much. Bass instruments have less bite.

For this speaker at least, I would consider current drive because series resistors, to me, improved the sound.

- keantoken

[ETM]

Originally Posted by ThorstenL

If we have a driver with a slotted Alu voice coil former we could link across the slot with a resistor of around 6 Ohm to give us the same damping as if we had the Amplifier producing voltage drive.

Actually, if we are to use this method the get the same damping as the voltage amplifier provides, one turn is not enough, but it would need a shorted second coil with the same number of turns as the primary coil.

However, if we do so, the damping coil generates just equal electromotive forces (motional EMF and inductance EMF) as the active voice coil does, and these EMFs would then bring back the same detrimental effects that are just avoided by current drive.

So, it is the mechanical damping supported by suitable cabinet filling materials that are useful, and I also think it would not take long to develop these if only some effort were put to it. Even now there are rubbers that yield free air Qm values of around 1.5; and according to my tests with copious cotton cloth stuffing, the value can yet be considerably lowered from this.

Often it is not even necessary to reach to the 0.7 since with a slightly higher value, the mild boost that develops in the 100 Hz region can be used for benefit to compensate some part of the baffle step.

With existing gear, it is feasible to employ some degree of electrical damping for the resonance region (a series RCL network in parallel with the driver) and keep the impedance level high elsewhere, as the benefits of current drive appear mostly in the mid and treble regions.

However current drive removes one absolutely major and one minor source of performance impairment that is unneccesary

I have sorted out and treated these effects:

• Voice coil acting as a microphone for the sound waves reflecting from the cabinet interior and passing through the diaphragm

• Voice coil acting as a microphone for the sound waves from adjacent drivers

• Mechanical and pneumatic non-idealities of the moving parts causing unpredictable EMF-effects

• Bl-variation causing modulation of impedance's angle and hence phase modulation of current at middle frequencies

• Position-dependent inductance of voice coil causing both amplitude and phase modulation (as in post #66)

• Voice coil inductance depends strongly on current level (without any displacement) causing non-harmonic distortion

• Magnetic coarseness of iron causing harmonic distortion, hysteresis and Barkhausen noise

• Resistance changes caused by temperature variations and manufacturing tolerances

• Program-dependent contact resistance variations in connectors and switches

Additionally, current drive makes driver protection easier and better defined since it is possible to use power transistors as the protection devices.

Originally Posted by keantoken

Using a 25 ohm rheostat as a series resistor, it seems to sound less militant. The sound doesn't bite so much, things sound smoother. I don't know if this is because of equalization changes. Speech is more legible. It is easier to relax while listening. (this is out of a single speaker, mono)

Instruments don't obscure each other so much. Bass instruments have less bite.

For this speaker at least, I would consider current drive because series resistors, to me, improved the sound.

These are pretty similar observations to mine when I made my first experiments in the early 90's and what I've been experiencing ever since. Hold on what you have found.

[keantoken]

I've been able to test series resistors with some hi-fi drivers lately. Results are the same. It is easier to turn up the volume without people complaining. I currently use 25 ohm resistors in stereo with Wild Burro BetsyK's. I can use such a high value since it's an open baffle and I need bass. Response is pretty flat to my ears.

I'm pretty pleased for now. I want to upgrade my amp soon, to exclude the possibility that it's simply reacting positively to the higher load impedance. Higher load impedance would change crossover characteristics, towards a more seamless crossover region AFAIK which would decrease higher order harmonics and improve small-signal linearity (it always bugs me that AB amps distort the most during the first watt...).

- keantoken

[john curl]

ETM, I did buy your book and I find it informative. However, I find it to be a 'double edged sword' sort of like the books written by Doug Self. I like your use of math. This is often presented in too difficult a form in many technical papers, or ignored by many popular writers. As you show, math is important, and with some effort from the reader, understandable.
The other side of the equation is your written opinions of areas outside your general research, such as TIM. Here, you make strong statements, but they really don't help the situation and are not useful in general. I have found this in other researchers who find speaker problems fascinating, and tend to think that electronics amplification is a very poor relation in the contribution to overall sound.
I still recommend this book.

[ETM]

Originally Posted by keantoken

I've been able to test series resistors with some hi-fi drivers lately. Results are the same. It is easier to turn up the volume without people complaining. I currently use 25 ohm resistors in stereo with Wild Burro BetsyK's. I can use such a high value since it's an open baffle and I need bass. Response is pretty flat to my ears.

Nice to hear you are continuing with positive results although 25 ohms doesn't yet make very ideal current drive. I encourage also others to experiment, and if one has access to a graphic equalizer or the like and knows the impedance curve, it may also be possible to eliminate most of the response difference.

Full range speakers usually suffer from cone break-up modes, that cannot be helped very much even by current drive, but even so, reducing the effects of the inductance EMF clears a lot of debris from the highs.

Originally Posted by john curl

I like your use of math. This is often presented in too difficult a form in many technical papers, or ignored by many popular writers. As you show, math is important, and with some effort from the reader, understandable.

One of the goals was just to bring the practical math and especially the use of transfer functions closer to the ordinary, not so highly educated readership, and if I have succeeded in that, I may be satisfied with that part.

In general, however, I was and am hoping more attention to the main points of the book, which I have outlined in my last posting. The effects should also be quite easily reproduced with basic equipment.

The other side of the equation is your written opinions of areas outside your general research, such as TIM. Here, you make strong statements, but they really don't help the situation and are not useful in general.

Concerning TIM, I was able to conclude, based on the results obtained with my simple rise rate meter, that the maximum rate of change in a CD music signal very rarely extends to even half of that reached with a 20 kHz sine wave at the maximum amplitude. (I assume you mean this.) This is also quite consistent with other studies I have come across. While such findings may not be directly related to the main topic, I thought they still give some practical insight into the broader art of distortion hunting.

BTW, there is a review of the book in the July issue of Audioxpress, but, as one can expect, it concentrates mostly on the bass region and resonance damping, omitting the main arguments presented, notably the effects of inductance. More on that later.

[atmasphere]

First I apologize for jumping in this late. This topic is very similar to one that I have dealt with a lot in the past. In looking at the history of the last 50 years, it appears that not all speakers are designed to be 'voltage driven'. Some are what I call 'power driven'. 'Current drive' seems to be a variant of that.
Competing paradigms in amplifier and speaker design

If this is regarded as a hijack, we can cover it elsewhere, but it seems related.

[keantoken]

Originally Posted by ETM

Nice to hear you are continuing with positive results although 25 ohms doesn't yet make very ideal current drive. I encourage also others to experiment, and if one has access to a graphic equalizer or the like and knows the impedance curve, it may also be possible to eliminate most of the response difference.

Full range speakers usually suffer from cone break-up modes, that cannot be helped very much even by current drive, but even so, reducing the effects of the inductance EMF clears a lot of debris from the highs.

In my experiments results seemed positive as long as I had series resistance above 4 ohms. I might go higher than 25, except here the bass becomes noticeably peaked. At 50 ohms and higher, the sound seems to lose life, and I attribute this to highly reduced lower midrange (this is the dip right after the bass impedance spike where AC impedance is close to DC impedance; this region suffers the most).

- keantoken

[keantoken]

Originally Posted by atmasphere

First I apologize for jumping in this late. This topic is very similar to one that I have dealt with a lot in the past. In looking at the history of the last 50 years, it appears that not all speakers are designed to be 'voltage driven'. Some are what I call 'power driven'. 'Current drive' seems to be a variant of that.
Competing paradigms in amplifier and speaker design

If this is regarded as a hijack, we can cover it elsewhere, but it seems related.

Consider Ohm's Law:

V is voltage,
I is current,
R is the load.

V/R=I
V/I=R
IR=V
VI=power

Current and voltage participate in power equally. So saying current drive is a variant of power drive is no more true than saying voltage drive is the same.

In fact, no amplifier can reach true voltage or current drive, though they have made heroic efforts. Any amplifier can only have an impedance between zero and infinity, and so they are all limited to some extreme of power drive. The only factor is the balance between current and voltage drive. If the amplifier has an impedance higher than the load, it is mostly current drive. If the amplifier's impedance is below the load, it is mostly voltage drive. And an amplifier with impedance equal to the load will be balanced voltage/current drive, which is closest to what you call power drive. I'm not sure if it is possible to make an amp which has balanced drive regardless of the load, however you can make a constant power amplifier for a fixed load.

- keantoken

[atmasphere]

Originally Posted by keantoken

Current and voltage participate in power equally. So saying current drive is a variant of power drive is no more true than saying voltage drive is the same.

In fact, no amplifier can reach true voltage or current drive, though they have made heroic efforts. Any amplifier can only have an impedance between zero and infinity, and so they are all limited to some extreme of power drive. The only factor is the balance between current and voltage drive. If the amplifier has an impedance higher than the load, it is mostly current drive. If the amplifier's impedance is below the load, it is mostly voltage drive. And an amplifier with impedance equal to the load will be balanced voltage/current drive, which is closest to what you call power drive. I'm not sure if it is possible to make an amp which has balanced drive regardless of the load, however you can make a constant power amplifier for a fixed load.

- keantoken

Prior to the rise of the Voltage paradigm, no amplifier was able to act as a voltage source. At the time (1955-1968 or so) as the Voltage model was being developed, the attention was focused on flat frequency response from box speakers, and seems to have ignored other speaker technologies whose impedance curve was/is not based on resonance.

Anyway, it is true that the prior art (tubes) had variable output impedance and so there was no 'current source' model. That is why I call it the Power Paradigm. It is still around today in active use- any SET is operating in this way.

Tomick, an Electro Voice engineer, wrote an article about damping that speaks to this issue, and proposed a circuit that allowed you to set the amplifier's response to deal with a wide range of speakers, some requiring a DF of 20:1 and others needing 0.1:1.

Missing Link in Speaker Operation

As we all know, the industry settled instead on the compromise of the Voltage Paradigm. It is a compromise and the Power Paradigm has a lot of advocates as a result! I think the most telling is that the ideal Voltage Paradigm specs that resulted seem to tell you nothing about how the amp will sound, other than once you know what the specs mean, you know the amp will likely sound a little harsh.