The Secret of Tube Amplifiers Revealed - and much more!

[ETM]

Have you ever wondered why all loudspeakers are fed solely by voltage, never caring a jot what the current through the voice coil happens to be? If so, you have really queried the very right thing.

Introducing the missing link in the pursuit of natural sound:

CURRENT-DRIVING OF LOUDSPEAKERS:
Remedy to the Fundamental Fallacy of Sound Reproduction Technology

by Esa Meriläinen



Softcover, 342 pages, available at Amazon.com.

Why have the basic laws of electrodynamics been ignored in the design of all loudspeaker operations? A frank exposure of the flaws of voltage drive and a thorough, groundbreaking guide to the only physically justified way of operating electrodynamic loudspeakers - current-drive. The author argues that the sound quality of virtually all existing speaker systems has been severely impaired by the manifoldly indefinite electromotive forces induced in the voice coil, that corrupt the flow of current; and what's dramatic - the measurable evidence is plain and overwhelming! These inherent EMFs in themselves can never be suppressed by any kind of amplifier, but their detrimental effect on current and hence on sonic performance can be eliminated by adequate source impedance. Thus, the secret of valve amplifiers also becomes apparent. Written for all from the academic community to lay hobbyists. Besides new concepts for amplifier and speaker design and demonstrative projects, the book also features novel ideas for modelling, filter design, measurements, and protection; and a useful tutorial on analogue linear systems. For more details, please visit www.current-drive.info.

If you have an account on Amazon.com, you are able to view some part of the content by the search inside feature.

Questions and comments are welcome either here or through my website.

Esa M.

[UnixMan]

Interesting topic, though not new. It have been discussed several times in many places, including this forum (try searching for "current drive" here and you'll find plenty of results). I wonder if the book contains any new info...

[ETM]

I'm afraid the book contains new and surprising information more than any other loudspeaker related book one can find.

As is said in the back cover:
Current-controlling loudspeakers is nevertheless not an all-new concept. Some of the benefits it brings, like decreasing of harmonic distortion and expanding of frequency range, have been known at least by specialists. However, there prevails great ignorance about the many disastrous interference mechanisms that traditional voltage drive inevitably creates and that are fully or partially eliminable by correctly operating current-drive. Also, current-drive does not necessitate active speakers any more than voltage drive does, but amazing improvement in sound quality is achievable by simple means. Based on properties of linear systems and investigation, plenty of other flawed conceptions prevalent in the field of audio technology are also pointed out.


And still, even if one is determined to hold on to ones prejudices about current drive despite any scientific facts or evidence, the book remains still a wealth of nowhere-else-found practical information and ideas on loudspeaker operation, modeling and simulation, filter design, and measurements; and teaches the general properties of analogue signals and systems.

[anatech]

Hi ETM,
May I humbly suggest that you investigate this a little more?

For instance, using current drive means that you must design the speaker in a totally new way. Otherwise, you'll end up with something that has a frequency response that closely follows the impedance curve of each driver. Damping for this speaker will be another casualty.

I do know that Nelson Pass has hands-on experience with this technique. I don't see it working for a multi-driver system that uses a crossover, no properly anyway.

To really make a system that will sound far better than the norm is simple enough. Use a multi-amp, electronically crossed over system. You could even add thermal correction from the drivers magnet. Now that has the capacity to sound amazing while remaining on solid ground. The use of a passive crossover is only a "make it go" solution. So 99% of the systems out there are not operating to their potential.

Remember, most current (sic) speaker drivers on the market today are designed with amplifiers that approximate a voltage source. You really need drivers designed for current drive if you expect really good sound quality.

-Chris

[planet10]

Originally Posted by anatech

For instance, using current drive means that you must design the speaker in a totally new way. Otherwise, you'll end up with something that has a frequency response that closely follows the impedance curve of each driver. Damping for this speaker will be another casualty.

Chris, you are speaking from the status quo. most speakers today and for the last 50 years* have been designed and built with the assumption that all amplifiers are voltage amplifiers.

*(since the big right (?) turn Nelson talks about)

It is certainly not the only way, and if we had been designed for current drive amps for 50 years the idea of designing a speaker for voltage drive would seem odd.

Talk to an old Western Electric engineer (if you can find one -- i got the opportunity about 10 years ago) and they'd think you a bit strange if you didn't have an amp with an output impedance - speaker impeadance.

Many of my speakers are designed for amps that tend toward current drive (the Fonken + F2 are supposedly a match made in heaven).

If current drive amplifiers become more common, speakers with inherently good damping will become more common.

dave

[planet10]

PS: my copy arrived from Amazon today. Won't get to look at it for a couple weeks.

[anatech]

Hi Dave,

Chris, you are speaking from the status quo. most speakers today and for the last 50 years* have been designed and built with the assumption that all amplifiers are voltage amplifiers.

That is my entire point, restated. But since we live in the real world where what I have said is true, what's your point?

*(since the big right (?) turn Nelson talks about)

I did mention that for completeness. I thought my response reflects the current state of the art. I also pointed out that there are fundamental differences between current and voltage drive type transducers. Yes?

It is certainly not the only way, and if we had been designed for current drive amps for 50 years the idea of designing a speaker for voltage drive would seem odd.

That goes without saying, doesn't it? But just think about your suggestion for a moment. Since loudspeakers were designed long before transistors appeared, wouldn't you think that a current driven loudspeaker would have matched the high output impedance of vacuum tubes? Therefore, the early loudspeaker and amplifier designers worked together to seemingly do things the hard way. It even added a very, very expensive component called an output transformer. The one component that is responsible for limiting the performance of all tube amplifiers. I fail to see why the industry would take the most expensive route possible that also limited the performance of thier products! Also note that the earlier loudspeakers had a common impedance of 16 ohms, not 8 ohms. A 4 ohm driver would have been a rare animal indeed! It was BTW.

The engineers of the early years (say 1935 even) were quite advanced in amplification and transducers. Most of the designs back then were only limited in the quality of components for building sound reproduction systems. They were more than intelligent enough to have eliminated the output transformer and operate in current mode to drive loudspeakers. They didn't, probably because there were very compelling reasons to go the way they did.
P-P outputs with transformer coupling. The only thing we see today are people who decide to go the direction of the audio dodo. It's maybe just different (perfectly acceptable), and a way to differentiate a product that will have a different sound. Selling product on this idea is irresponsible in my view. To play and experiment - fine. However, the title of the book alone is misleading.

Talk to an old Western Electric engineer (if you can find one -- i got the opportunity about 10 years ago) and they'd think you a bit strange if you didn't have an amp with an output impedance - speaker impeadance.

Naw, I haven't done that. My discussions tended to involve a fellow from Westinghouse and quite a number of technicians raised and lived on tubes alone. Their very common sense viewpoint on these things is hard to ignore.

Many of my speakers are designed for amps that tend toward current drive (the Fonken + F2 are supposedly a match made in heaven).

Well, of course! That is the design direction that you took. I haven't heard your speakers, but I have heard many other high efficiency, single driver systems. The old ones found in churches (infinite baffle with a 10" ~ 15" driver) were very efficient. The rather large room they played into also helped. Just a little P-P 10 ~ 14 watt amplifier was enough. Still, not what I would ever call accurate sound! For me, an optimum B4 type enclosure is a happy thing. Driven with a voltage source type amplifier of course!

If current drive amplifiers become more common, speakers with inherently good damping will become more common.

Is this a chicken or the egg thing? 'Cause I can't see that happening any time soon. What we have now is well entrenched and can perform extremely well. Still better if we go with active crossovers and one amp per driver. That is actually a bit like what you wish to do, except that driver limitations have been recognized. By limiting the bandwidth supplied to the driver, we optimize each driver for it's frequency range and can apply full damping - which is (I think) the one thing that gives you the character you like about a single driver / no crossover system. A passive speaker crossover seems to be what we both wish to get away from here. The only problem seems to be the added expense of two more amplifiers and an electronic crossover. In your case, the frequency extremes are ignored and Doppler effects are accepted. I know this because the Doppler effect is a problem with two way systems.

Nothing wrong with experimenting and having fun as long as this is not oversold. That would be my only issue with single driver systems that are not for background music or paging only. (think "Musak")

-Chris

[ETM]

Originally Posted by anatech

Hi ETM,
May I humbly suggest that you investigate this a little more?

For instance, using current drive means that you must design the speaker in a totally new way. Otherwise, you'll end up with something that has a frequency response that closely follows the impedance curve of each driver. Damping for this speaker will be another casualty.

Hi, Chris
The book contains at least thee whole chapters that deal with designing the speaker in a totally new way. So why are you so humbly advising me on this?

Damping of the speaker is factually only a matter of shaping the frequency response at the low end, and there is a multitude of ways to accomplish this in current-drive as well as in voltage drive. Again, a whole chapter is devoted to this. It is also possible to keep the bass driver closer to voltage mode near the resonance region and shift to current-mode for higher frequencies.

Originally Posted by anatech

I do know that Nelson Pass has hands-on experience with this technique. I don't see it working for a multi-driver system that uses a crossover, no properly anyway.

Nelson Pass has been on the right track with his F1, which has got rave reviews even with speakers that are not originally designed for current-drive. However, this here is not a proper way to design crossover filters for current drive because the impedance seen by the drivers is not taken into account.

With current-drive as the target, crossovers do work just as well or as badly as when voltage drive is the target. There is full dualism between these two modes of operation.

Originally Posted by anatech

To really make a system that will sound far better than the norm is simple enough. Use a multi-amp, electronically crossed over system. You could even add thermal correction from the drivers magnet. Now that has the capacity to sound amazing while remaining on solid ground. The use of a passive crossover is only a "make it go" solution. So 99% of the systems out there are not operating to their potential.

Multi-amping with voltage drive is the worst method because the drivers see zero impedance at all frequencies and thus the spurious interference currents generated by the electromotive forces (both the motional EMF and inductive EMF) are at maximum for all frequencies.

Originally Posted by anatech

Remember, most current (sic) speaker drivers on the market today are designed with amplifiers that approximate a voltage source. You really need drivers designed for current drive if you expect really good sound quality.

Of course the design considerations differ somewhat, but even with existing types it is possible to reach amazing results. This is also addressed in the book.

Originally Posted by anatech

They were more than intelligent enough to have eliminated the output transformer and operate in current mode to drive loudspeakers. They didn't, probably because there were very compelling reasons to go the way they did.

Such "compelling reason" may well have been that tubes are not able to deliver enough current to drive a loudspeaker directly.

At least with modern amplifier technology, there is not any compelling reason, or any valid reason whatsoever, to operate a speaker by voltage. Nobody has ever presented such a (scientifically valid) reason, and nobody ever will because such a reason simply does not exist. (And reasons relating only to frequency response shaping are not valid reasons.)

Originally Posted by anatech

However, the title of the book alone is misleading.

A is said in the preface:

"The subtitle "Remedy to the fundamental fallacy of sound reproduction technology" may at first sound exaggerative or inflated, which it, however, is not; for as will show up especially in chapter 4, the issue is honestly about nothing lesser than that.

We have namely been deceived - not so much with intent but by negligence or ignorance anyway - by providing to our use solely audio power equipment that disregards the realities of physics, and by establishing, as the backing of the practice, yet odd myths that do not stand up any closer examination. Most regrettable in the present practice is that the manifoldly indefinite electromotive forces of the loudspeaker are allowed to freely mingle with the reproduced signal."

In short, we have these two choices: Either the voltage is controlled whereby the current becomes severely corrupted (as is summarized here) or then the current is controlled whereby the voltage becomes severely corrupted. As a speaker driver responds only and solely to current, according to the law F=Bil, it really makes no sense whatsoever to choose the first-mentioned.


By the way, the price is now 21% off, $27.76 at Amazon, so it's now very affordable for anyone.

Here is a link to the news release at mi2n.

[kikey]

Hi everybody, this is my first post here.

Interesting subject!

For people interested in current drive I can recommend the following links:
Current driving
Effects Of Source Impedance on Loudspeakers
http://www.passdiy.com/pdf/cs-amps-speakers.pdf

[ETM]

My comments about the above sites:

Current driving

Though they have grasped the idea of current-drive in a way, there is so much technical inaccuracies that as an engineer I cannot recommend this page or their designs to anyone. Their last update also seems to be from behind 10 years. (A little bit of knowledge can be a dangerous thing.)

Effects Of Source Impedance on Loudspeakers

This is a very good example of the popular fallaciousness of rationale upon which the arguments laid out against current-drive are based: They pick up a conventional bass loudspeaker whose frequency response is designed only for voltage drive, then they apply a current signal to that loudspeaker and find out that the bass response differs unfavorably from that obtained by voltage drive. From this they then draw the conclusion that current-drive must be inherently bad. So sad, but this is just how the reasoning generally goes: if the bass frequency response of one's favorite (sub)woofer suffers from current-drive, then the whole concept must be thrown away! And the real sufferer is all of humanity.

Some facts which I have also tried to point out in the book: The frequency response and transient response of any linear system are not separate things but one and the same thing only viewed in different ways. If any two linear systems (e.g. a subwoofer so long as its behaviour can be considered linear) have the same frequency response (amplitude & phase), then their transient responses (for any input) must also be equal - there cannot be any difference. This stems directly from the Fourier Transform and properties of linear systems.

Therefore, if the frequency response of a loudspeaker that is designed for current operation is made equal to the frequency response of a loudspeaker designed for voltage operation, then all the transient properties of the two speakers are also automatically equal, and in this respect one cannot be better than other.

The essential difference between the driving methods lies not in any frequency response matters, as frequency responses (and hence time responses) can always be tailored well enough in both cases. The essential and crucial difference is instead in the huge amount of distortion (mostly IM) and interference that can be eliminated or at least substantially attenuated by the correct driving technique.

http://www.passdiy.com/pdf/cs-amps-speakers.pdf

These experiments are at least of some real interest. However, from the standpoint of suppressing the EMF interference currents of the driver, using high-efficiency drivers in not a good starting point. This is because the higher the sensitivity of the driver is, the stronger are also the EMFs (both motional and inductive EMF) generated by the driver and the higher the source impedance must be for a given level of interference suppression. The fuzzy nature of these EMFs also becomes clearly visible in the shown impedance curves as peaks and roughness.

Also, it may have gone unnoticed by the experimenters, but when an impedance of only a few ohms (R1-C1-L1) is placed in parallel with the driver, it can hardly be said to operate in current-mode any more. Instead, at the upper mids and the treble, the actual operation will be (with the tabulated values) closer to voltage drive than current drive.