Go Back   Home > Forums > >
Home Forums Rules Articles diyAudio Store Blogs Gallery Wiki Register Donations FAQ Calendar Search Today's Posts Mark Forums Read

Tubes / Valves All about our sweet vacuum tubes :) Threads about Musical Instrument Amps of all kinds should be in the Instruments & Amps forum

Seeking advice on "Circlotron" cross-coupled feedback OTL design
Seeking advice on "Circlotron" cross-coupled feedback OTL design
Please consider donating to help us continue to serve you.

Ads on/off / Custom Title / More PMs / More album space / Advanced printing & mass image saving
Thread Tools Search this Thread
Old 6th September 2004, 04:06 AM   #1
arteitle is offline arteitle  United States
diyAudio Member
Join Date: Dec 2003
Location: Toledo, OH
Default Seeking advice on "Circlotron" cross-coupled feedback OTL design

I'm still a tube design newbie, but lately I've been learning all I could, reading Morgan Jones and Bruce Rozenblit's books, and experimenting with PSpice. I've been playing with a design for an OTL amp using 6C33Cs, and I was hoping that the more knowledgable here would take a look and give me their opinions.

John Broskie's Tube Cad Journal from a few years ago included a partial schematic for an OTL amp he called "Circlotron/Horizontal push-pull amplifier with cross-coupled feedback" which piqued my interest. It uses a 6DJ8 to drive a push-pull pair of 6AS7s, with the cathodes of the 6DJ8 tied to those of the 6AS7 and the output. As he says, "If the output does not follow their cathodes, their cathode-to-grid voltage will change in value as a consequence," and "So what this circuit does is apply a quick, short feedback loop across the 6AS7s. This feedback greatly reduces the output impedance and lowers the distortion of the amplifier."

Attached below is the version I've been tinkering with, currently using a 6SN7 (or possibly a 7N7 for its lower capacitances) to drive a quad of 6C33Cs (one parallelled pair for push, one for pull). The preceding gain/phase splitter stage is a long-tailed pair made from a 6SL7 (or 7F7) with a CCS cathode load. Originally I started out with a 12AX7 LTP and 6DJ8 driver, then tried changing the driver to a 5687, and then the 6SN7, which I understand is known for its linearity. However, I was dismayed at seeing all those great parallel lines on the plate characteristics graph getting smushed together at the southeastern end of my load line, so I tried CCS plate loads instead of resistors, which simulates great, but makes my schematic look pretty FET-heavy. As far as I can tell, Miller effect in the 6C33Cs is negligible due to their sub-unity gain. The 100 ohm grid stoppers are just token values at the moment, and the other overly-precise resistor values were chosen to get idle voltages and currents as close to integers as possible.

It's meant to run in class AB1 only. With an 8 ohm (resistive) load, input sensitivity is 1.5 Vp-p, and max output power is 100 W peak. With a 4 ohm load, sensitivity is 1 Vp-p and power is 74 W peak. Bandwidth is within 1 dB from around 7Hz to 90 kHz, and THD is about 0.5%. Of course these are idealistic results with a purely resistive load, but you get the idea.

Does anyone see anything glaringly wrong with this design, with my choice of tubes, or anything else? Are the CCS anode loads a bad idea? I'm very curious to hear any comments you might have, and to learn all I can. Thanks very much.

Aaron T.

(Please view the full-sized schematic here )
Attached Images
File Type: jpg otlampschemsmall.jpg (80.8 KB, 889 views)
  Reply With Quote
Old 6th September 2004, 09:21 AM   #2
Arcolette is offline Arcolette  Germany
diyAudio Member
Join Date: Jun 2004
Location: Germany
Default Circlotron OTL design

Hallo Aaron,

the biasing of your OTL design is very critical. That is not a particular problem of your design. That is typical for almost all circlotron based OTL-designs from my knowledge. The 6C33C has a large variation of anode current during warm-up. And after large amplified signal (i.e. dynamic peak in a classic Symphony) this anode current will have another value than before. This causes a large unsymmetry of the bridge and a rather high DC voltage occurs at the load (loudspeaker). Another problem is the runaway of the working points of the output tubes (6C33C). Like with the Graaf GM-20 circuit the stabilizing of the DC working point does not operate sufficiently well. Graaf selects the 6C33C tubes quite hard for lowest drift of anode current and best matching. You will need about 40 tubes to find 4 pieces which might work. But there is no guarantee that your amp will work after 50 hours of operation (if it or your loudspeaker survives this time span...). So, all in all it is not recommendable from my point of view to build such an amp.

Greatings Arcolette
  Reply With Quote
Old 6th September 2004, 09:47 PM   #3
smoking-amp is offline smoking-amp  United States
diyAudio Member
smoking-amp's Avatar
Join Date: Dec 2001
Location: Hickory, NC
Default DC Servo

I have heard of this bias or operating point runaway for 6C33 before. A simple way to fix this would be to use a couple of Op. Amp. DC servos on the output tubes. These just monitor the DC current levels and control the DC bias levels, the audio signal is still AC coupled from the tube drivers. A little more tricky to sample the current in the Circlotron configuration, maybe have to use instrumentation differential amps to measure floating voltage on a small sampling resistor. Unfortunately the sampling resistors raise output impedance some.
Another way would be to use a couple of center tapped inductors (or just one, if you can find one with two center tapped windings) and just one non-floating power supply. Inductor 1 connnects across the 6C33 plates and inductor 2 connects across the load (or cathodes) and gets its center tap grounded. The power supply connects between the two center taps. Current sampling resistors can be put into the separate wires forming the grounded center tap. Then put the final power supply filter caps across the inductor ends in the same way the floating power supplies would have been connected. By connecting the caps to the non grounded end of the sampling resistors, the resistors will be deleted from the audio current path, so won't affect output impedance. And, the sampling resistors will have one end grounded for easy use with a couple of DC bias servo Op. Amps.
The inductors will have balanced DC currents if the DC servos are working correctly, so no air gap required for the inductors. The all-in-one inductor design has to have the phasing between windings correct. Have to make sure the inductor(s) have enough inductance to avoid loading the amplifier and heavy enough wire to handle the DC currents.

  Reply With Quote
Old 29th December 2009, 04:47 PM   #4
tiefbassuebertr is offline tiefbassuebertr  Germany
diyAudio Member
tiefbassuebertr's Avatar
Join Date: Jul 2005
Location: D-55629 Schwarzerden
in this case this thread could be of interest
Graaf GM20 mods - help!
  Reply With Quote
Old 29th December 2009, 05:23 PM   #5
tyu is offline tyu  United States
diyAudio Member
Join Date: Jul 2009
Default Circlotron EZe,s amp i ever Diy,ed

Bias is in the front end,i used 6AS7
Attached Images
File Type: gif m60-signal.gif (18.7 KB, 582 views)
  Reply With Quote
Old 29th December 2009, 06:51 PM   #6
kevinkr is offline kevinkr  United States
diyAudio Moderator
kevinkr's Avatar
Join Date: Sep 2004
Location: Boston, Massachusetts
Seeking advice on "Circlotron" cross-coupled feedback OTL design
You might want to read my 6C33 otl saga (you can find it on my website) many of the issues raised by others on this thread can be dealt with by careful burn-in and aging of the 6C33 before you put them into service.

There are a number of members here who have been running 6C33 based circlotrons for many years without difficulty. It is all down to design and properly selecting the output tubes. Hopefully one of these guys will weigh in and give you some guidance.

I never observed the odd dynamic behavior Arcolette describes in his post, and bias drift was most significant in the first 15 minutes of operation, and I think can easily enough be compensated for in a good design without the use of a bias servo.

Cathode follower or source followers direct coupled to the grids of the 6C33 seem to work better in my experience as they do draw some grid current even at modest power levels. I would try to keep things simple, I think you have over complicated things. (I certainly did in my design which is why it is not available today)

A few key points:

Filament currents are extremely high and will destroy most socket contacts, some mil surplus Soviet era sockets are rugged enough.

Convection cooling around the sockets and tube bases will improve tube and socket life.

Keep anode dissipation to 45W or less.

Higher currents at lower plate voltages seem to work better. IMHO don't go over 140V, and 110 - 120V is probably most reasonable.

100W into 4 ohms is unrealistic even with 4 x 6C33, ok into 8.

Burn in all output tubes for a minimum of 48hrs in a jig built for the purpose. Start by running just the filaments for 30 minutes or longer and then apply bias and plate voltage in that order. Ramp up the plate current over a couple of hours to provide 45W of dissipation. Make sure whatever you do is safe!
"To argue with a person who has renounced the use of reason is like administering medicine to the dead." - Thomas Paine
  Reply With Quote
Old 5th March 2010, 12:06 AM   #7
j.michael droke is offline j.michael droke  United States
diyAudio Member
Join Date: Jan 2009
Seeking advice on "Circlotron" cross-coupled feedback OTL design

A few key points: [re: OTL power amp using 6c23 tubes]...

"Keep anode dissipation to 45W or less.

Higher currents at lower plate voltages seem to work better. IMHO don't go over 140V, and 110 - 120V is probably most reasonable. ...kevinkr"

100W into 4 ohms is unrealistic even with 4 x 6C33, ok into 8....

Hi there: As mentioned by "kevenkr" in the above post, with 110 to 120v dc for the B+, a simplified SS diode power supply would be possible, eliminating the load of iron transformer for the four 6c23's, cost and space saving? (6.3v filiment transformer(s) would still be needed).... regards, Michael
  Reply With Quote
Old 5th March 2010, 07:22 PM   #8
ulibub is offline ulibub  France
diyAudio Member
Join Date: Nov 2006
Location: Lorraine
Default 6c33-design-comments

Hi Aaron,

did you already see this thread? Might be interesting to you....

6C33 design comments?

This thread tells also a lot about experience with 6C33 Circlotron amps...

I have built one, not easy and certainly not a beginner's project, but I use it since several years now regularly without any bigger problems - but surveillance of the bias currents is recommended

all the best

  Reply With Quote
Old 5th March 2010, 10:55 PM   #9
atmasphere is offline atmasphere  United States
diyAudio Member
atmasphere's Avatar
Join Date: May 2007
If you really want to build an OTL, do yourself a favor and avoid the 6C33. Its a great sounding tube, but if you don't drink the cool-aid from the experience posted above, you will toast the sockets, and the tubes themselves are not exactly cheap.

The 6AS7 is a lot easier to set up and overall less expensive, certainly more reliable.

If you do stick to the 6C33s, I strongly recommend inserting some 10watt 3 ohm resistors in series with each cathode! The tubes will hold up better and you will reduce distortion as well. I would also consider installing a fuse in series with each cathode, about 1.5A for protection against arcs in the tubes.

**Do not** capacitively couple to a the grid of a 6C33. If you do you will be dealing with getting bass and transparency out of the unit and juggling that against bias stability. IOW you will be trying to put the smallest cap in the unit that still plays bass and is transparent, and when you reach the limits there then the bias network will get 'reworked' to a higher impedance to allow the timing constant to play out. Its a battle you can't win! Using a cathode-follower driver direct-coupled to the grids eliminates this problem.
  Reply With Quote


Seeking advice on "Circlotron" cross-coupled feedback OTL designHide this!Advertise here!
Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump

Similar Threads
Thread Thread Starter Forum Replies Last Post
Conversion of Hiraga's "Le Monstre" to a Circlotron tiefbassuebertr Solid State 13 30th May 2015 03:31 AM
A little "legal" advice on a design similarity kneadle Multi-Way 20 17th April 2006 06:39 PM
Seeking a "pure digital amplifier" ... MarkZ Class D 10 28th December 2005 03:58 AM
Effect of cross-coupled "local" FB on plate load of previous stage ray_moth Tubes / Valves 23 27th April 2005 09:55 AM

New To Site? Need Help?

All times are GMT. The time now is 03:21 PM.

Search Engine Optimisation provided by DragonByte SEO (Pro) - vBulletin Mods & Addons Copyright © 2018 DragonByte Technologies Ltd.
Resources saved on this page: MySQL 15.00%
vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2018 DragonByte Technologies Ltd.
Copyright ©1999-2018 diyAudio