Bacon665,
Exactly! Any Circlotron output stage is a traditional bridge circuit; each side of the bridged output requires a separate and independent power supply that "floats".
The FET Circlotron project adheres to the same requirements; the document calls-out the need for independent/floating power supplies for each half of the output-stage circuit on page 8 of the PDF:
http://www.passdiy.com/pdf/Build The Amazing FET Circlotron.pdf
The need for the twin "floating" power-supplies for each power-amplifier channel is the "pound of flesh" extracted in exchange for the elegant symmetry of the Circlotron push-pull output stage. In the case of the M-60, the four 6AS7's (eight triode sections) in each "phase" of the output stage share a "floating" +/-140VDC power-supply.
Last edited:
Basically, the OTL problem using vacuum tubes is that the plate impedance of the tube (s) must be matched to the impedance of the reproducer (speaker) for maximum power transfer. Tube plate impedances can be made as low as 1000 Ohms, for low voltage, high current operation, or tubes in parallel. Few speakers will show impedance greater than a few Ohms. The output transformer therefore operates as an impedance matching device.
OTL amplifiers using transistors are more practical, because collector impedance of a transistor more closely matches speaker impedance.
I have seen one tube amplifier using the single-ended tube as a cathode follower. This gets the output impedance down closer to the impedance of the speaker, although such a circuit brings up several other design problems, such as second harmonic distortion, and some circuit-protection issues.
Floyd Carter
OTL amplifiers using transistors are more practical, because collector impedance of a transistor more closely matches speaker impedance.
I have seen one tube amplifier using the single-ended tube as a cathode follower. This gets the output impedance down closer to the impedance of the speaker, although such a circuit brings up several other design problems, such as second harmonic distortion, and some circuit-protection issues.
Floyd Carter
Thanks. Finally it's starting to make sense. The notation on the schmatics were scr**ing it up for me. I think it's time to start reading up on powersupplies. 
And it kind of makes me wonder if something similar could be applicable in a buffer?
It really is beautiful in it's simplicity.
A famous scientist once said "in simplicity lies the genius". (Or something along those lines?)
I'll look into the Bridge circuit (circlotron) link as well.
And it kind of makes me wonder if something similar could be applicable in a buffer?
It really is beautiful in it's simplicity.
A famous scientist once said "in simplicity lies the genius". (Or something along those lines?)
I'll look into the Bridge circuit (circlotron) link as well.
Thanks. Finally it's starting to make sense. The notation on the schmatics were scr**ing it up for me. I think it's time to start reading up on powersupplies.
And it kind of makes me wonder if something similar could be applicable in a buffer?
It really is beautiful in it's simplicity.
A famous scientist once said "in simplicity lies the genius". (Or something along those lines?)
I'll look into the Bridge circuit (circlotron) link as well.
MarkusG,
Glad to help; Heaven knows I've needed help in "clearing the fog" on a number of occasions. Bootstrapping each other is one of the reasons we participate in these public forums...
You could certainly implement a circlotron/bridge buffer. Take the M-60 circuit and just chop-off everything before the coupling capacitors that carry the signals from the cascoded input stage; just tie your buffer inputs to the coupling capacitors (along with an input resistor to the ground potential, possibly 100K-ohm resistors). The driver stage and the output stage are simple followers; no voltage gain, just current gain. With just one pair of 6AS7's (each powered by one of the two "floating" power-supplies) and the 6SN7 driver tube (powered by a separate +/-300VDC power-supply), you'd have a "power follower" that could deliver up to 600mA peak current to drive any circuit input of your choice! Talk about a kick-butt buffer!!
Floyd,
Thanks for the thoughts; however, the OTL circuit under discussion is a cathode-follower. The thread (at least at this point) is focused on developing a DIY implementation of the Atma-Sphere M-60, which is a Circlotron OTL cathode-follower DC-coupled output configuration.
Yes, solid-state OTL power-amplifiers (such as the FET Circlotron on the Pass DIY Web site) are certainly more practical, but overt practicality is somewhat beside the point for this project. The music is the primary goal, so we're looking to build a fire-breathing, hair-shirt, vacuum-tube OTL power-amplifier...
Come along for the ride, it should be fun!
The passivists claim no buffer is the best buffer. Maybe it's time to challenge those beliefs?
But first things first...
For those limited situations in which an unbuffered passive preamp can successfully connect the source component to the power-amplifier, that configuration is optimal. But, in my experience, it's an incredibly rare set of circumstances in which the unbuffered passive preamp cannot be bettered by an attenuator feeding the source component signals into a buffer/line-driver. And what better way to answer that requirement than a fire-breathing, bat-out-of-hell, vacuum-tube Circlotron balanced buffer --- it would have to be positively orgasmic, right??!!!
since ive already mentioned it and a couple others have as well its going to show up in the search results so i might as well add that the transistors for the fet circlotron are hard to find. I sourced parts for it and i think ebay was the only source i found. A redesign might be in order with a new thread.
The IRFP240 power MOSFET's are readily available from Mouser and Digikey. The Toshiba 2SJ74BL JFET's have always been a specialty item in somewhat limited available to individuals, but they're also readily available from B&D Enterprises (B&D Enterprises: Part #2SJ74BL).
I first became aware of B&D from noted DIYaudio forum member, Grey Rollins. B&D makes a number of those hard-to-get Toshiba JFET's (apparently constructed of "unobtanium"...
) available to the DIY community --- IMHO, a great resource. As far as duplicating the power supplies goes....
If you were to opt for some odd voltage could you find a transformer large enough to power ALL of the B+/B- then from that place isolation transformers(1:1) so you essentially have 2 3 or even 4 of the odd power transformer from which to build your floating supplies.
If you were to opt for some odd voltage could you find a transformer large enough to power ALL of the B+/B- then from that place isolation transformers(1:1) so you essentially have 2 3 or even 4 of the odd power transformer from which to build your floating supplies.
Bacon665,
I don't think that would be possible; even if you could use that approach for the pair of "floating" output-stage +/-140VDC power-supplies, you'd still need a separate +/-300VDC power-supply for the input/driver stages. The Antek AN-15475 1500VA power-transformer could provision sufficient current capacity to power twice the number of 6AS7's as the Antek AN-8475 800VA power-transformer that I had identified to power the 60-watt M-60 monoblock output stage; however, once you've purchased sufficiently robust isolation transformers to "share" the single power-transformer, I don't see where you would gain any useful advantage.
Why the aversion to just accepting the trio of power-transformers as the cost of implementing an advantageous topology? I'm already budgeting for all of the power-supply "iron" that I'll need to build a pair of the monoblocks...
I wouldnt call it an adversion. the SMPS was just something i threw out there. as for the iso's i was thinking on a smaller scale where this approach could actually save some money. But looking at the prices of matching transformers this application doesnt exist unless you are building a VERY small amp as a driver for an XLR cable. My personal intentions are not to build the M-60 as I've never heard one or even one quite similar to it if i dont like the sound id simply be throwing money down the tube (for tubes ). Ive got my own (mostly own) OTL design in the beginning stages of being cooked up. Its posted in another thread here but atm as i dont have simulation software im stuck unless somebody happens to notice something that needs to be corrected.
). Ive got my own (mostly own) OTL design in the beginning stages of being cooked up. Its posted in another thread here but atm as i dont have simulation software im stuck unless somebody happens to notice something that needs to be corrected.I wouldnt call it an adversion. the SMPS was just something i threw out there. as for the iso's i was thinking on a smaller scale where this approach could actually save some money. But looking at the prices of matching transformers this application doesnt exist unless you are building a VERY small amp as a driver for an XLR cable. My personal intentions are not to build the M-60 as I've never heard one or even one quite similar to it if i dont like the sound id simply be throwing money down the tube (for tubes
I certainly understand that standing-up a project on the scale presented by the M-60 is a serious financial commitment. Unfortunately, an inherently simple signal-path always requires a disproportionately massive power-supply infrastructure, so there's really no way to side-step the need to sink the money in order to hear how the circuit sounds when properly implemented.
If you can provision a Linux-based system (possibly as a second-boot on your PC), you could make use of the GPL'ed suite and toolkit of Electronic Design Automation tools (gEDA).
These tools can be used for electrical circuit design, schematic capture, simulation, prototyping, and production. Currently, the gEDA project offers a mature suite of free software applications for electronics design, including schematic capture, attribute management, bill of materials (BOM) generation, netlisting into over 20 netlist formats, analog and digital simulation, and printed circuit board (PCB) layout. Be advised that the learning-curve is wicked steep, but the tools offer a very robust personal EDA environment.
M-60 OTL Power-Supply Considerations
Now that we've identified some likely power-supply transformer candidates, the next power-supply design task is to determine the energy-storage needs for the various power supplies. Rather than employ the usual array of often conflicting engineering design formulas for calculating various power-supply characteristics (output ripple voltage, etc.), I've developed a "rule-of-thumb" formula over the years (based on the energy-storage of the capacitors (specified in joules), then transformed into values of work over a period of time) for directly estimating a rational range of capacitor values necessary for high-quality audio performance. For those interested in the calcuation method, I'll include that information in a later post.
The first-level approximation for total power-supply capacitance values for each power-supply rail is:
I've "second-sourced" these capacitor value calculations by reviewing the schematics for other OTL power-amplifier designs with similar power output specifications and these projected power-supply capacitance values seem to be in the ball-park. Perhaps we can ask our technology patron for this thread, DIYaudio forum member, "atmasphere", to comment on the validity of these power-supply capacitance values for the M-60?
NOTE: Power-supply decoupling capacitors for each stage are always a good idea; 50microFarads for decoupling the 300VDC rails, 200microFarads for decoupling the 140VDC rails. The use of film-type capacitors for these roles often pays a significant sonic benefit.
NOTE: The input/driver stage power-supply could be provisioned exclusively with film-type capacitors, potentially paying sonic dividends through the use of higher-quality power-supply capacitors; I would recommend this approach.
I'm planning to provide an initial power-supply schematic, but the best time-frame that I can commit to is some time in the next few weeks since my "day job" (large project getting underway) and family life will largely consume my available cycles during that time; sorry, it's that whole work/life balancing act that we must all perform...
Now that we've identified some likely power-supply transformer candidates, the next power-supply design task is to determine the energy-storage needs for the various power supplies. Rather than employ the usual array of often conflicting engineering design formulas for calculating various power-supply characteristics (output ripple voltage, etc.), I've developed a "rule-of-thumb" formula over the years (based on the energy-storage of the capacitors (specified in joules), then transformed into values of work over a period of time) for directly estimating a rational range of capacitor values necessary for high-quality audio performance. For those interested in the calcuation method, I'll include that information in a later post.
The first-level approximation for total power-supply capacitance values for each power-supply rail is:
Input/Driver Stage: 100uF-220uF (for each 300VDC power-supply rail in the +/-300VDC power-supply)
60-watt Output Stage: 2200uF-3400uF (for each of the four 140VDC power-supply rails, grouped into a pair of "floating" independent +/-140VDC power-supplies)
Vacuum-tube Filaments: no power-supply capacitors needed; we can power the heaters directly via the 6.3VAC transformer secondaries
60-watt Output Stage: 2200uF-3400uF (for each of the four 140VDC power-supply rails, grouped into a pair of "floating" independent +/-140VDC power-supplies)
Vacuum-tube Filaments: no power-supply capacitors needed; we can power the heaters directly via the 6.3VAC transformer secondaries
I've "second-sourced" these capacitor value calculations by reviewing the schematics for other OTL power-amplifier designs with similar power output specifications and these projected power-supply capacitance values seem to be in the ball-park. Perhaps we can ask our technology patron for this thread, DIYaudio forum member, "atmasphere", to comment on the validity of these power-supply capacitance values for the M-60?
NOTE: Power-supply decoupling capacitors for each stage are always a good idea; 50microFarads for decoupling the 300VDC rails, 200microFarads for decoupling the 140VDC rails. The use of film-type capacitors for these roles often pays a significant sonic benefit.
NOTE: The input/driver stage power-supply could be provisioned exclusively with film-type capacitors, potentially paying sonic dividends through the use of higher-quality power-supply capacitors; I would recommend this approach.
I'm planning to provide an initial power-supply schematic, but the best time-frame that I can commit to is some time in the next few weeks since my "day job" (large project getting underway) and family life will largely consume my available cycles during that time; sorry, it's that whole work/life balancing act that we must all perform...
I've been away over the weekend but now that I'm back I can do nothing but applaud this initiative.
I plan to sit down with a friend and talk it out regarding power supplies some time this comming week. Hopefully he can fill in some of the blanks for me.
I would love to see some example psu's. I understand the concepts and topologies, kind of... but I'm no-where near being able to perform the required calculations and design decisions necessary.
I plan to sit down with a friend and talk it out regarding power supplies some time this comming week. Hopefully he can fill in some of the blanks for me.
I would love to see some example psu's. I understand the concepts and topologies, kind of... but I'm no-where near being able to perform the required calculations and design decisions necessary.
The power-supply is certainly a key aspect in the build-out of an M-60, so a clear understanding of the unique role of the power-supplies in explicitly serving as fundamental elements of the Circlotron bridged output stage will be crucial.
We'll harness the power of the collective ("resistance is futile; you will be assimilated...")
Seriously, before we can begin defining detailed power-supply designs, we need to determine the core component values. We're working to develop an acceptable power-supply design that is predicated upon off-the-shelf components. Once we have defined these major components, we can get down to the business of providing a complete power-supply schematic for the group. At that juncture, we'll also have a clearer view into the costs associated with building-out the M-60 and the physical space required to house all of the pieces...
We're gradually developing an M-60 power-supply design, "peeling the onion" layer-by-layer. We've identified off-the-shelf power-transformers that are capable of driving the M-60's power needs. We've also determined the total capacitance/energy-storage needs of the power-supplies; I've begun screening the usual DIY component suppliers to identify specific candidate power-supply capacitors.
Power-supplies for a "beast" such as the M-60 usually employ "pi" filters (C-L-C) to isolate the minimalist circuitry from the vagaries of the AC-mains. Unfortunately, my initial estimation of the desired combination of inductance values and current capabilities do not seem to be available in off-the-shelf components; I've sent e-mail quote requests to a couple of toroidal transformer vendors to determine how much it would cost to have custom inductors manufactured at reasonable cost.