OTL Asylum

OTL, Output Transformerless Amplifier User Group.

Return to OTL Asylum

Classical Circlotron - A superb performer

77.43.10.3
Posted on November 4, 2009 at 02:51:42
6AS7_6SN7
Audiophile

Posts: 98
Joined: September 10, 2009



It is generally accepted that the key features of OTL circlotron topology are:
- no even order harmonics
- vanishingly small distortion at low level
- predominat third harmonic at high level

The above appears to be confirmed by the following LTSpice simulations.

I know that most seasoned engineers think that SPICE 'does lie'.
Even though I am seasoned too I think that SPICE is a very useful tool for DIY design of HI-FI tube amplifiers.

As it is not fair to simulate the circuits of others I have used this simple one of my own creation.

And in order to retain some correlation with the results of those inmates who do not like feedback at all the components values have been adjusted for 3 dB feedback - 1V peak input for 1.8A peak @ 8 Ohm.

The simulated spectra follow.

Best Regards
Luca Cellai
ecc230
Not sure I would call that 'classical', posted on November 4, 2009 at 08:57:57
Ralph
Manufacturer

Posts: 1156
Location: Minnesota
Joined: April 24, 2002
Contributor
  Since: April 1, 2002
 The original 'classic' Circlotron circuits date from the 1950s and look nothing like this.

There are some missed bets in the voltage amplifier circuits. You might want to run your simulations again after installing effective constant current sources for the differential amplifiers- most semiconductor designs have them, why not tubes? I would also consider running a bipolar supply for the driver circuit. It will help in enhancing the performance.

FWIW, though I find that most CCS circuits used in transistor amps are ineffective. You might want to look in the archives here for a better CCS design.

I know, I know, posted on November 5, 2009 at 00:32:36
6AS7_6SN7
Audiophile

Posts: 98
Joined: September 10, 2009
 Dear Ralph,

It seems that in the asylum everybody is concerned with names.
Just to make myself clear.
Classical in this context does neither refer to the initial Electrovoice patent nor to any particular realization.
It is only a practical way to indicate a topology that most closely resembles the published OTL circlotron schematics.
As far as the constant current sources (CCS) are concerned I regret to say that I am not interested in complicating the schematic simply because of my personal philosophy, that is, "good enough is good enough".
I do not think that a further potential 10 dB reduction of distortion in exchange for an higher parts count is worth the effort.
As the simulations already show, the prospected distortion of this sample circlotron (of the order of 1-2% third harmonic) occurs only at high current levels. High power levels occur in music only for a tiny fraction of time, when the aural masking effect is very strong.
My technical opinion is, if more 'undistorted' power is needed, it is better to add output tubes.

Best Regards
Luca Cellai

ecc230
RE: I know, I know, posted on November 5, 2009 at 13:55:33
Ralph
Manufacturer

Posts: 1156
Location: Minnesota
Joined: April 24, 2002
Contributor
  Since: April 1, 2002
 The first Circlotron patent was by Cecil Hall. The second by Bongiorno and the the 3rd and 4th by Atma-Sphere. There was never an EV patent.

A lot of the distortion you are showing is likely coming from your driver circuit, not the output. If you tinker with the cathode resistor values of the driver (and use a bipolar power supply) you will see what I mean. CCS circuits will clear this up a lot. I would also explore what happens with different values of current limiting resistors in the output section.

The means for driving the power tubes is another area of weakness of this approach (which has been used by at least 3 manufacturers that I know of, only one of which is still in business). In practice, an amplifier like this will exhibit slow overload recovery, bias instability and/or weak bass, possibly both.

RE: I know, I know, posted on November 6, 2009 at 21:13:03
Tre'
Industry Professional

Posts: 4845
Joined: February 9, 2002
U.S. Patent 2,705,265, C. T. Hall, "Parallel Opposed Power Amplifiers" was filed on June 7, 1951 and granted on March 29, 1955.

Finnish Patent 27332, Tapio Köykkä, (original amplifier) was filed on September 2, 1952 and granted on November 10, 1954.

U.S. Patent 2,828,369, A. M. Wiggins, "High Fidelity Audio Amplifier" was filed on March 1, 1954 and granted on March 25, 1958.

Finnish Patent 29642, Tapio Köykkä, (improvements to the amplifier) was filed on September 30, 1955 and granted on April 10, 1958.


.
Have Fun and Enjoy the Music
"Still Working the Problem"
RE: I know, I know, posted on November 9, 2009 at 09:06:00
Ralph
Manufacturer

Posts: 1156
Location: Minnesota
Joined: April 24, 2002
Contributor
  Since: April 1, 2002
 And all this time I thought the Wiggins was just a variant. Thanks for the looky-look.

You're welcome (nt), posted on November 9, 2009 at 12:06:55
Tre'
Industry Professional

Posts: 4845
Joined: February 9, 2002
 .
Have Fun and Enjoy the Music
"Still Working the Problem"
RE: I know, I know, posted on November 6, 2009 at 18:58:11
WCT
Audiophile

Posts: 27
Location: central coast, California
Joined: July 18, 2002
 "The second by Bongiorno"


just curious... is that Bongiorno as in James Bongiorno of Great American Sound(and others) fame?
RE: yup. nt, posted on November 9, 2009 at 09:02:03
Ralph
Manufacturer

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

Quick check, posted on November 6, 2009 at 06:56:55
6AS7_6SN7
Audiophile

Posts: 98
Joined: September 10, 2009
 Dear Ralph,

I have provisionally replaced common cathode resistor with current generators (ideal CCS) of the same value of the bias currents.

The overall result is negative, that is, no major change of output spectrum.

Then I have checked the spectrum at the output of the first stage and at the output of the second stage. Here I have seen the likely culprit.

There is an unexpected (and fixed) -55 dBc third harmonic at the output of the first stage in both cases (that is, with and without CCS).

The likely problem is that computer simulations with 'simple' tools (of the level of free LTSpice), due their inherent numerical errors, do not deliver reliable results below -50 dBc (0.3% distortion).

As a consequence, as it is not easy for me to unmask the effect of the third harmonic that is unexpectedly present at the output of the first stage, and that is likely due to numerical rounding errors, it is not possible to ascertain with LTSpice the prospected effect of your recommended modifications.

We can only trust your professional experience.

I have been told that other SPICE tools, whose license cost more than several thousand dollars, could deliver improved accuracy (down to 0.1% distortion or even better), but that kind of investigation is too expensive for me.

Best Regards
Luca Cellai

ecc230
RE: Quick check, posted on November 6, 2009 at 09:17:39
Ralph
Manufacturer

Posts: 1156
Location: Minnesota
Joined: April 24, 2002
Contributor
  Since: April 1, 2002
 I don't know how spice models CCS circuits, but I find that most CCS designs commonly seen in tube and transistor amps are inadequate designs. If you can, look through the ASOG archives at the bottom of this page- somewhere in there is a schematic of a more effective CCS. The results we have been getting out of it have been transformation.

In this field you are the master, posted on November 6, 2009 at 04:01:39
6AS7_6SN7
Audiophile

Posts: 98
Joined: September 10, 2009
 I am a hobbist.
I got your message.
I will take your suggestions as new food for brain.

Best Regards
Luca Cellai
ecc230
Spectrum @ 1750mA peak 8 Ohm, posted on November 4, 2009 at 02:59:37
6AS7_6SN7
Audiophile

Posts: 98
Joined: September 10, 2009



Spectrum @ 1750mA peak 8 Ohm
ecc230
Spectrum @ 1400 mA peak 8 Ohm, posted on November 4, 2009 at 02:58:04
6AS7_6SN7
Audiophile

Posts: 98
Joined: September 10, 2009



Spectrum @ 1400 mA peak 8 Ohm
ecc230
Spectrum @ 1050 mA peak 8 Ohm, posted on November 4, 2009 at 02:57:10
6AS7_6SN7
Audiophile

Posts: 98
Joined: September 10, 2009



Spectrum @ 1050 mA peak 8 Ohm
ecc230
Spectrum @ 700 mA peak 8 Ohm, posted on November 4, 2009 at 02:56:05
6AS7_6SN7
Audiophile

Posts: 98
Joined: September 10, 2009



Spectrum @ 700 mA peak 8 Ohm
ecc230
Spectrum @ 350 mA peak 8 Ohm, posted on November 4, 2009 at 02:54:34
6AS7_6SN7
Audiophile

Posts: 98
Joined: September 10, 2009



Spectrum @ 350 mA peak - 8 Ohm
ecc230

Page processed in 0.052 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 ]