Hi,
I'm looking for some guidance on good design starting points for a multichannel SET amplifier.
Priorities include low noise (assuming a sufficiently quiet power supply) and simplicity. The final amplifier does not need to provide wattage higher than 2 or 3 watts continuous, but I would like to be able to push at least 4x power during transient peaks.
Low frequency performance is not important.
Thanks in advance.
bob
I'm looking for some guidance on good design starting points for a multichannel SET amplifier.
Priorities include low noise (assuming a sufficiently quiet power supply) and simplicity. The final amplifier does not need to provide wattage higher than 2 or 3 watts continuous, but I would like to be able to push at least 4x power during transient peaks.
Low frequency performance is not important.
Thanks in advance.
bob
A property to consider is the power bandwith of an amplifier. In low frequencies there is no headroom above the measured maximal (undistorted) output voltage. If you wish to amplify 10W peaks, you need a 10W amplifier. In SE design that's aready expensive territory 
Is your budget a concern?
Is your budget a concern?
Last edited:
I should mention that I'm also interested in avoiding transformers if possible.
I have attached a schematic of an early version of the MA1 OTL amplifier.
If I were to remove a significant portion of the tubes, would I run into issues with the output impedance of the amplifier as it relates to damping factor (more so than simply exponentially lower power)?
Imgur: The magic of the Internet
I have attached a schematic of an early version of the MA1 OTL amplifier.
If I were to remove a significant portion of the tubes, would I run into issues with the output impedance of the amplifier as it relates to damping factor (more so than simply exponentially lower power)?
Imgur: The magic of the Internet
Build the circlotron if you want to, but be prepared for the potential problems along the way.
The first thing I notice about your schematic is that there are not individual resistors in each 6AS7 cathode for current sharing, that is asking for one or more tubes to take over, and destroy itself.
Don't get me wrong, OTL tube amps can be very nice, depending on the design, and the factors you are willing to trade off.
How many Ohms are your loudspeakers?
Not just the manufacturer's nominal rating, but the impedance minimum(s) at various frequencies. A typical 8 Ohm loudspeaker may have minimum impedance(s) of 4 Ohms.
10 Watts rms into 8 Ohms requires 1.58 Amps peak.
10 Watts rms into 4 Ohms requires 2.24 Amps peak.
That is a lot of current from 12 cathodes of 6 6AS7 in series with 12 more cathodes of 6 6AS7 tubes.
OTL and circlotron requires many output tubes. The cathode impedance of a 6AS7 is 143 Ohms, for 12 cathodes (6 tubes), the cathodes are 12 Ohms. The circlotron schematic you linked to has 12 Ohm + 12 Ohm output impedance = 24 Ohm impedance. That is because the set of 6 tubes is in series with the other set of 6 tubes.
For an 8 Ohm loudspeaker, the signal will be developed across 12 Ohms cathodes, + 8 Ohms, speaker + 12 Ohms cathodes.
If there is 10 Watts into 8 Ohms, the signal power distribution will be thus:
8 Ohm loudspeaker 10 Watts
24 Ohm tube cathodes impedance 30 Watts
Of course there is much much more total power dissipated in the tubes (the plate dissipation), I just listed the signal power dissipated in the amplifiers output impedance.
The damping factor of 24 Ohms output impedance into 8 Ohm loudspeaker is 0.33 (yes, the damping factor is less than unity). The frequency response is almost entirely dependent on the loudspeaker impedance versus frequency.
Negative feedback (that does not exist in your schematic) could reduce the output impedance, but there would still be the same total power dissipated in the tubes.
The circlotron amp is very inefficient at providing power to the speaker, but is very efficient at heating the room.
A tube amp that drives an output transformer from the plates is more efficient than your schematic.
For a 10 Watt tube amp that can work with a variety of loudspeaker impedances (both nominal and minimum, an output transformer is a very practical thing.
A PSE 300B amp (Parallel Single Ended) could easily put out 14 Watts.
A typical damping factor of a 300B amp could be 3.3.
That is 10 times better damping factor than the schematic you posted.
The first thing I notice about your schematic is that there are not individual resistors in each 6AS7 cathode for current sharing, that is asking for one or more tubes to take over, and destroy itself.
Don't get me wrong, OTL tube amps can be very nice, depending on the design, and the factors you are willing to trade off.
How many Ohms are your loudspeakers?
Not just the manufacturer's nominal rating, but the impedance minimum(s) at various frequencies. A typical 8 Ohm loudspeaker may have minimum impedance(s) of 4 Ohms.
10 Watts rms into 8 Ohms requires 1.58 Amps peak.
10 Watts rms into 4 Ohms requires 2.24 Amps peak.
That is a lot of current from 12 cathodes of 6 6AS7 in series with 12 more cathodes of 6 6AS7 tubes.
OTL and circlotron requires many output tubes. The cathode impedance of a 6AS7 is 143 Ohms, for 12 cathodes (6 tubes), the cathodes are 12 Ohms. The circlotron schematic you linked to has 12 Ohm + 12 Ohm output impedance = 24 Ohm impedance. That is because the set of 6 tubes is in series with the other set of 6 tubes.
For an 8 Ohm loudspeaker, the signal will be developed across 12 Ohms cathodes, + 8 Ohms, speaker + 12 Ohms cathodes.
If there is 10 Watts into 8 Ohms, the signal power distribution will be thus:
8 Ohm loudspeaker 10 Watts
24 Ohm tube cathodes impedance 30 Watts
Of course there is much much more total power dissipated in the tubes (the plate dissipation), I just listed the signal power dissipated in the amplifiers output impedance.
The damping factor of 24 Ohms output impedance into 8 Ohm loudspeaker is 0.33 (yes, the damping factor is less than unity). The frequency response is almost entirely dependent on the loudspeaker impedance versus frequency.
Negative feedback (that does not exist in your schematic) could reduce the output impedance, but there would still be the same total power dissipated in the tubes.
The circlotron amp is very inefficient at providing power to the speaker, but is very efficient at heating the room.
A tube amp that drives an output transformer from the plates is more efficient than your schematic.
For a 10 Watt tube amp that can work with a variety of loudspeaker impedances (both nominal and minimum, an output transformer is a very practical thing.
A PSE 300B amp (Parallel Single Ended) could easily put out 14 Watts.
A typical damping factor of a 300B amp could be 3.3.
That is 10 times better damping factor than the schematic you posted.
Like everyone else says, cyclotron would be the way to go for an OTL amp. Although I would use 6C33C-B tubes instead of 6as7s.
You might even be able to try something like a super triode connected OTL amp?
However, if you are not budget constrained and you only want between 8-12 watts, I would probably advise you to build something with output transformers. If you only really need 8 watts, I would just build a bunch of single ended 300b amps. If you really need more power, then I would look at a bunch of single ended GM-70 amps.
Not to mention that it is a LOT easier to match a handful of 300bs to each other than it is to match a few dozen 6as7s/6C33C-B tubes.
10 Watts in 8 Ohms needs 9V 1.1 Amps peak.
Few tubes will pass over 0.1A at moderate voltage. "Moderate" still being 10X more than the 9V you get into the load, so efficiency is way down the toilet.
A different view. For good power transfer you want the "valve resistance" much less than the load impedance. Like 1 Ohm. 6080 is a 280 Ohm resistance. 8 in parallel is still 30X more than load impedance. A heater with a small side-effect of sound. There are no <10r tubes or reasonable tube-teams. (You need like a dozen, each side, to come close.)
Few tubes will pass over 0.1A at moderate voltage. "Moderate" still being 10X more than the 9V you get into the load, so efficiency is way down the toilet.
A different view. For good power transfer you want the "valve resistance" much less than the load impedance. Like 1 Ohm. 6080 is a 280 Ohm resistance. 8 in parallel is still 30X more than load impedance. A heater with a small side-effect of sound. There are no <10r tubes or reasonable tube-teams. (You need like a dozen, each side, to come close.)
I assumed that by SET you meant single-ended triode. Was I mistaken?
If LF is not important, I'm assuming you are looking for the an amp that will drive mids and tweeters in a bi-amped set-up. Is that correct?
A "SET" implies a Single Ended amplifier using a Triode tube. The Circlotron is neither "Simplest" or Single Ended, although it is often built with triodes.
The MA1 type design could be built, but as stated you would run into peak current limitations with only two output tubes. The 6AS7 and it's Russian equivalent have a reputation for handling brief overcurrent peaks, but your speakers are at risk if one fails.
It is possible to build a simple single ended amp without an output transformer, but you will still need a huge choke or CCS load, efficiency would be terrible and you would likely wind up with a capacitor in series with your speaker. It would also need one very big triode, or several smaller ones wired in parallel.
I had this idea several years ago when I found several 5 Henry 5 amp chokes in a surplus yard. They were large and weighed about 50 pounds each. My plan was to use a few triode wired TV sweep tubes in parallel. The chokes did not make the trip when I had to move 1200 miles, so the idea was abandoned.
Note, the data sheet lists a suitable big triode for the job.....good luck finding them for a reasonable price. It would be easier to wire a bunch of 6AS7's or 6336's together.
Hoooooooolly crap. Now thats what I call a tube. It even looks more linear than a 6as7.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.