OTL Asylum

OTL, Output Transformerless Amplifier User Group.

Return to OTL Asylum

The OTL speaker that does not exist (yet)

77.43.10.3
Posted on September 23, 2009 at 23:49:08
6AS7_6SN7
Audiophile

Posts: 98
Joined: September 10, 2009
 Let's imagine a full range speaker with two standard insulated voice coils (4, 8 or 16 Ohm).

Let's imagine what happens if two voice coils replace the two half primaries of the output transformer of the recently published 'Stonehenge Schematic' (see the DIY section of the asylum).

Don't you see?
A non standard OTL topology (maybe even novel, I do not know) comes out, with push-pull action at voice coil level.

I am quite interested in Ralph and Lew opinion, as if the idea gained sufficient consensus there is room for asking to Lowther, Fostex, Ciare (and the other manufacturers) to add such a new speaker to their catalog.


ecc230
RE: There are dual voice coil drivers..., posted on September 24, 2009 at 09:07:28
Ralph
Manufacturer

Posts: 1156
Location: Minnesota
Joined: April 24, 2002
Contributor
  Since: April 1, 2002
 Mostly they are woofers, not full-range devices. You'd have a bit of current to deal with through each voice coil; there does not appear to be any means of eliminating the DC current in the Stonehenge circuit (but the schematic was a bit fuzzy...)

Fertin might be the people to talk to, but they offer 32 ohms already (optional), ideal for smaller OTLs. You could probably get by with only 2 power tubes per channel. Their regular stuff is 16 ohms, and that already is an excellent impedance value.

Full range is the key ingredient, posted on September 25, 2009 at 00:45:49
6AS7_6SN7
Audiophile

Posts: 98
Joined: September 10, 2009
 Under normal bias conditions (100-300 mA) the DC current through the coils is well within the power handling capability of most commercial 4-8 Ohm speakers.
And with the referenced schematic the power supply is much simpler than the one that is necessary for Circlotron.
Circlotron is in my opinion the most elegant schematic ever conceived for balancing the two arms of push-pull OTL amplifiers but Circlotron needs at least three independent HT power supplies, and this implies more mass and volume.
The Stonehenge amplifier that I have recently built has a weight of less than 8 Kg and if the use of the output transformers might be avoided at all (by a suitable dual coil full range speaker) another 1-2 Kg could be spared and the volume could be further reduced by 20%.
By using a computer switching power supply for filaments (and direct mains rectification as in the referenced schematic) a tube amplifier of less than 5 Kg appears feasible.
As the tag price of a tube amplifier is more or less directly related to its mass, more economic OTL products are theoretically possible. And I think that the availability of commercial OTL amplifiers (to be connected to dual coil speakers), that are competitive even with respect to low cost Chinese tube amplifiers, would be beneficial to penetrate new market segments (for manufacturers of speakers and OTL) and to stimulate the DIY market too.

ecc230
RE: Full range is the key ingredient, posted on September 25, 2009 at 09:59:50
Ralph
Manufacturer

Posts: 1156
Location: Minnesota
Joined: April 24, 2002
Contributor
  Since: April 1, 2002



I think you will find that continuous current through a voice coil will have undesirable effects on the speaker over time. Magnets do not like heat.

With regards to weight, FWIW if weight were the determining factor in a tube amp I'm sure things would look a lot different in the high end audio world today! Its nice to think about how light a tube amp could be (I play in a band and the bass player has a 400-watt tube amp; I don't think its a good idea to pick that thing up without help).

Switching supplies certainly can help in that area as long as they are engineered properly. There has been a lot of advance in that field in the last 20 years- so its more a matter of getting them built in quantity. So far conventional transformers work out a lot better if cost vs weight has anything to do with it. There is something to be said for the longevity and serviceability of conventional power supplies- I've seen a lot of semiconductors go obsolete over the years; imagine having to repair a switching supply and having difficulties finding the semiconductor needed.

It is possible to derive the driver power supply from the output section power supplies of a Circlotron, so if one wanted to, you could get by with 2 supplies rather than the three we use. However, even in a transformer-coupled amp, a separate power supply for the driver has the same advantage of reducing IM distortion that it has for us now.

FWIW, if you did have a dual voice coil driver, you could hook it up in a more conventional fashion, joining the two coils together to form a center tap tied to ground, the other sides tied to your tube banks. You would have the same DC offset issues, but the circuit could be simpler and lower output impedance. You're not going to get any serious voltage gain out of the output section anyway, and running the ground current through the speaker rather than the plate current might be a little safer, but the danger to the operator would still be present. This might be an easier way to set up the Stonehenge amp, as you already have an OPT for that.

BTW, years ago (1985?) we built an amp that put the power tubes in a circle. This amp made 100 watts into 8 ohms. They are rare; we made 7 of them.

A few flash comments, posted on September 26, 2009 at 22:47:52
6AS7_6SN7
Audiophile

Posts: 98
Joined: September 10, 2009
 1) Magnets heating - to me this is not a meaningful problem as 200 mA-DC @ 8 Ohm = 0.32 W/coil (0.64 W for two coils).

2) The 'more conventional fashion' (to hook the voice coils) you are referring to will certainly destroy the perfectly balanced PP operation of output tubes.

3) Safety - The safety risks of the assembly (i.e. potential danger to the operator) are exactly the same as those of an hair dryer (or any similar home appliances): can this possibly be a real problem?

4) HUM - The proposed topology exhibits the minimum sensitivity to residual HT ripple, that is, push pull operation of voice coils (as well of all transformer half primaries) cancels out residual HT hum about 10 dB more that with 'center tap tied to ground'.
ecc230
RE: A few flash comments, posted on October 6, 2009 at 09:59:11
Ralph
Manufacturer

Posts: 1156
Location: Minnesota
Joined: April 24, 2002
Contributor
  Since: April 1, 2002
 The danger to the operator is **significantly** higher than the use of a hair drier!!

If such an amp is on with no speaker is attached, it is dangerous. The shock hazard will be equal to the power supply voltage, since humans are usually high impedance. If someone had a cut on their finger, this could easily be lethal.

Regarding P-P balance- there is no difference there with regards to any P-P output section.
RE: The OTL speaker that does not exist (yet), posted on September 24, 2009 at 07:47:15
Lew
Audiophile

Posts: 2685
Location: Bethesda, Maryland
Joined: December 11, 2000
 I am too lazy and busy to go look up the Stonehenge circuit, but it sounds like you are thinking about an analogue to direct drive of an ESL, using an OTL. (No audio transformer is necessary on the ESL side, if the audio signal can be delivered to the ESL at sufficiently high voltage.) Only in your idea the transducer will be an electromagnetic type rather than an ESL. It's a natural implementation with an ESL, which is an inherently balanced transducer. In your conception, you want to transform the electromagnetic transducer to balanced operation. Am I on the right track?
Lew
You got the point, posted on September 25, 2009 at 00:05:43
6AS7_6SN7
Audiophile

Posts: 98
Joined: September 10, 2009
 The idea is that electromagnetic full range transducers, operated in push-pull (balanced) mode, can effectively replace the output transformer of tube amplifiers. A sufficient efficiency will only be possible with low impedance tubes (like 6AS7), but this is not big new for OTL fans.
The two coils (in the referenced schematic) are in series, same current but with reversed polarity (under no signal conditions), so that the net force applied to the cone is zero.
With audio signal one coil will push while the other will pull.
And if the bias current through the coils is of the order of a few hundred mA the DC power dissipation is well within the power handling capability of commercial spekers.
ecc230
I really like it !, posted on September 25, 2009 at 21:33:22
Satelliteman
Audiophile

Posts: 101
Location: Napier - The Art Deco City of the world
Joined: May 21, 2006
 What an brilliant concept ....this topic has really woken me up !
True outside of the box thinking - I really hope this idea goes somewhere - id love to be able to build a prototype if I had the shop tools to do it . No doubt some of you have these.

Thank you, your comment is my best reward!, posted on September 26, 2009 at 22:58:18
6AS7_6SN7
Audiophile

Posts: 98
Joined: September 10, 2009
 Open mind and intellectual honesty, that are my goals.
ecc230
Just curious: are you cellai@space.it ?, posted on October 19, 2009 at 12:22:38
LineArrayNut
Audiophile

Posts: 1921
Location: Tenn.
Joined: December 10, 2002



seems like it...
I *like* the loudness button!
RE: Just curious: are you cellai@space.it ?, posted on October 20, 2009 at 11:41:54
6AS7_6SN7
Audiophile

Posts: 98
Joined: September 10, 2009
 YES
ecc230
RE: The OTL speaker that does not exist (yet), posted on September 24, 2009 at 07:40:26
Slick
Audiophile

Posts: 308
Location: Maryland
Joined: April 10, 2002
 I'd just be happy with a 32 ohm Maggie........
"My God, it's full of stars"
RE: The OTL speaker that does not exist (yet), posted on October 7, 2009 at 09:35:39
WALACE
 Just wonder that any problems you have of running DC in a transfornmer is being shifted to to the loudspeaker . . (thinking in terms of saturation etc).

I got slightly excited by the idea of using the two matching tranformers on a pair of stacked quad ESLs in series with the centre earthed - but I don't want to run DC through them at all. . ..
A word of caution, posted on October 9, 2009 at 02:39:28
6AS7_6SN7
Audiophile

Posts: 98
Joined: September 10, 2009
 If the DC bias current (let's say less than 250 mA) passes through two (coaxial) voice coils of the very same loudspeaker (under the indicated conditions) there is no static resulting force applied to the diaphragm and only some minor heat dissipation.

But if you are thinking of passing the DC bias current, whatever it is, through the voice coils of two separated identical loudspeakers you get an extended sound source where the diaphragm of one is pushing while the diaphragm of the other is pulling the air, but at different spatial positions. The resulting dipole radiation pattern is likely to exhibit an intolerable spatial/frequency dependent interference pattern.

I am quite convinced that the proposed idea is only viable by using a (dual coil) full range speaker.
ecc230
RE: A word of caution, posted on October 29, 2009 at 07:54:42
WALLACE
Audiophile

Posts: 24
Location: London, UK
Joined: July 31, 2008
 Silly question - maybe - but why are loudspeakers generally low impedance ?

Must have always been a bit of a pain to match even with a transformer . .


W.

Page processed in 0.094 seconds.

[ Top Top of Page ] [ Contact Us ] [ Support/Wish List ] [ Copyright © 2009, Audio Asylum, all rights reserved ] [ Go Green ] [ Take5 ]
[ General ] [ Speakers ] [ Tubes ] [ Vinyl ] [ Digital ] [ Hi-Rez ] [ Video Asylum ] [ Cables ] [ Tweaks/DIY ] [ Music ] [ Films ]