• WARNING: Tube/Valve amplifiers use potentially LETHAL HIGH VOLTAGES.
    Building, troubleshooting and testing of these amplifiers should only be
    performed by someone who is thoroughly familiar with
    the safety precautions around high voltages.

Dual Single Ended Amplifier? - (not PSE)

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Hi,

Wondering if anyone has seen, built, or heard anything like this..

Two SE output stages per channel, signal at each grid out of phase WRT the other. Two SE output transformers per channel, one of which has its secondary winding reversed so that the waveforms are recombined in phase.

There are some interesting things that supposedly occur in that there is a combined function of the two output transformers.

All of which is outlined in the diagrams and description at;
Dual single ended amplifier by Cohen, Graeme John ? AU1996042023

I'm interested in pretty much anything anyone cares to offer in response.


Thanks,
Shane
 
Hi Shane,

yes, I did experiment with this setup. I built the prototype, basically a PP amplifier, with 2 single ended transformers per channel and swapt in a PP transformer for comparison. Indeed the 2 single ended transformers performed much better, sonic wise. Buy may also be due to the quality of the iron: the SE transformers were Silver Rock, while the other was a German 'Engel' (plain EI core)

Of course that for the same amount of power a PP transformer is much smaller than 2 SE transformers, but I had no problems with space in my hobby room.

Regards, Erik
 
The only advantage I can see is that the OPTs will not have to do a flux zero crossing, which may or may not create distortion in a PP OPT.

this may be a nice approach. I wonder if the benefits of this approach are best suited for low power, both because the amp spends more time near the zero crossing and because the SE trafo's would be huge if high power were needed.

some imbalance between the two SE trafo's would be likely, so maybe this would play more like a SE at lowest power.
 
I have done this from another perspective. Small SE OPT's are relatively common and cheap. So are tubes and circuits for them in the less than 10 watts per amp range. SE transformers and tubes in the 30+ watt range are not common or cheap. Why couldn't we wire 4 10 watt SE amps in parallel to make a 40 watt amp. Well, it works, and if you wire the two amps out of phase as you suggest.....it IS a push pull amp. Now granted it may have less than perfect coupling between the two halves, but it is a class A push pull amp. I noticed that when the 4 amps were paralleled in phase (two stereo SSE's) the maximum power output was roughly equal to 4 times a single channel. When two of the amps are wired out of phase and driven with a phase splitting transformer, I could push the amps harder and get nearly 6 times the output power. Why? Because it is a push pull amp and you can now drive one side into cutoff without overall clipping.

So, why couldn't I wire push pull amps in parallel....well that works too. I have a good supply of a certain P-P OPT rated at 80 guitar amp watts. I have seen 500 watts come out of 4 P-P amps running sweep tubes through these OPTs wired in parallel.

When you parallel 2 amps you need to drop the load impedance by half. If you have an 8 ohm speaker drive it with two 16 ohm taps in parallel. You can series parallel OPT's but pure series doesn't work as well since the coupling between the two amps isn't direct since they are connected only through the speaker.
 
so you could take a stereo SET and wire the outputs in parallel and use it as a mono, or invert one channel and put them in parallel and use it as PP mono

Yes, I and a few builders have taken my SSE board which is a two channel SE amp, driven the channels out of phase, connected a normal P-P OPT and created a push pull amp. An SSE usually makes 12 to 14 WPC when run hard using KT88's at 450+ volts in UL mode. Using the same power supply I get 60+ watts mono from the same KT88's wired in pentode with a P-P OPT (no UL taps). No changes are needed to the SSE amp other than the OPT and a differential drive source (my guitar preamp). Makes a little terror of a guitar amp.

does it mean that no biasing is needed beyond the phase splitter stage

Not sure what you mean by the words "no biasing". Most tubes must have some kind of negative bias applied to its grid to control the tube current, whether it has signal voltage applied or not.

I assume that you are reffering to operating the output tubes where the idle current is zero, or very near zero. Zero current operation would be true class B operation which yields high efficiency, high power output, and high distortion especially at low volumes. It was used for high power PA amplifiers where 10% distortion was acceptable.

If tubes, mosfets, BJT's or whatever were completely linear devices all the way down to zero current, this would be a perfectly valid approach. Unfortunately, they are not. Linearity is best in the middle or their operating range and degrades at each end of the curve. To avoid high distortion near the region where the signal transitions from one device to the other (known as crossover distortion) in a push pull pair, we apply enough bias to move the devices out of the region of worst linearity. This results in an overlap region where both devices are operating.

There has been talk, and even a few commercial products, of building "screen driven amplifiers" using TV sweep tubes (I have built a few). In this case the tubes can be operated very close to true class B with idle currents as low as 3 mA. This is because a TV horizontal (line) output tube was designed for use as a grid (G1) controlled CCS with the screen (G2) as the current controlling element. The curves of G2 voltage VS tube current are very linear and hold flat over a wide plate voltage range, down to very low currents. Screen drive has other issues like high drive requirements, tube arcs and spurious oscillations if high power levels are attempted, probably preventing mainstream popularity.
 
so you could take a stereo SET and wire the outputs in parallel and use it as a mono, or invert one channel and put them in parallel and use it as PP mono
I think the correct answer is yes and no. Two signals in phase will be additive in parallel. Two signals, with one out of phase, will cancel each other. A push-pull signal becomes additive in the primary of the output transformer by the way it is constructed. This doesn't happen if the SE secondaries are simply paralleled. You would have to series them and refigure output impedance.
 
Two signals, with one out of phase, will cancel each other.

If you parallel the inputs, you can just parallel the outputs and drop the load impedance. You will still have a SE amp that makes about twice the power.

If you feed the inputs out of phase, you must wire the outputs out of phase. This can be done by wiring the outputs in series observing the proper phase, and increasing the load impedance.

The outputs can be wired in parallel, but one set of secondary wires must be swapped. For most OPT's this means you are connecting the black wire to the green wire....twice...and then wiring the speaker across the black/green pairs. You must still reduce the load impedance and you now have a P-P amp.

You also have the option of simply using a P-P OPT since it is now a P-P amp.
 
Hi,

Wondering if anyone has seen, built, or heard anything like this..

Two SE output stages per channel, signal at each grid out of phase WRT the other. Two SE output transformers per channel, one of which has its secondary winding reversed so that the waveforms are recombined in phase.

There are some interesting things that supposedly occur in that there is a combined function of the two output transformers.

All of which is outlined in the diagrams and description at;
Dual single ended amplifier by Cohen, Graeme John ? AU1996042023

I'm interested in pretty much anything anyone cares to offer in response.


Thanks,
Shane

It sounds like you are bridging the two SE amps. This is commonly done in pro audio for very high power needs. There are advantages and disadvantages to this approach. One advantage is that each of the SE amps will only have to produce half of the required voltage to achieve the same output from the speaker. This means each SE amp would produce less distortion. One disadvantage is that each SE amp now sees basically half the load, doubling the current required. So, if you've got enough current available in your output stage, then this can work great.
 
It sounds like you are bridging the two SE amps.

Bridging usually refers to driving the two channels out of phase and tying the load across the two hot output terminals, effectively operating the amp outputs in series. This is indeed common in the solid state amp world. It is not always recommended with tube amps.

Solid state amps tend to be designed with excess open loop gain and considerable negative feedback This provides a very low output impedance and good channel to channel gain balance.

Most tube amps do not use lots of feedback, in fact many designs have none at all. This means the individual channel gain is determined by the tubes themselves and can vary by a db or two fron channel to channel. The output impedance is much higher than a SS amp, often an ohm or two. Wiring two channels in series can cause considerable output from one channel to be dissipated in the other channels output stage. This can blow stuff up.

I have experimented with bridging and paralleling tube amps. All amps are of my own design, and I don't mind blowing stuff up! I haven't destroyed anything yet, but a slight channel imbalance in a bridged pair does cause extremely high plate voltages in one channel especially if one amp clips before the other. I have successfully paralleled 4 125 Watt amp channels wiring all the 16 ohm outputs in parallel and connecting a 4 ohm load. Before doing this you need to test the gain of each channel at low, mid, and high frequencies, and swap tubes around until they all match. In this manner one amp clipping will tend to force the others into clipping without excessive plate voltage spikes.

These are my own experiences, and others may see something else, but many of our tube amps were never intended to be operated in this manner. Approach with caution and monitor plates with scope probes before cranking things up!
 
Hi Shane,

yes, I did experiment with this setup. I built the prototype, basically a PP amplifier, with 2 single ended transformers per channel and swapt in a PP transformer for comparison. Indeed the 2 single ended transformers performed much better, sonic wise. Buy may also be due to the quality of the iron: the SE transformers were Silver Rock, while the other was a German 'Engel' (plain EI core)

Of course that for the same amount of power a PP transformer is much smaller than 2 SE transformers, but I had no problems with space in my hobby room.

Regards, Erik

Hi Erik, did you then try just a single SE stage in comparison to both?. What I'm wondering is - and assuming power output is not an issue - would the harmonic profile represent that of a SE or PP amplifier, or neither?

Thanks to all,
Shane
 
would the harmonic profile represent that of a SE or PP amplifier, or neither?

Both.....Since you are essentially making a P-P amp, it works much like a P-P amp, but since you now have two OPT's and the summing happens at the secondaries, there will be less cancellation of the even harmonics. The usual large amount of second harmonic in the SE amp is caused by the "bending" of the I/O transfer function inherent in an SET. Much (but not all) of this will cancel since the two transfer functions are now "bent" in opposite directions yet tied together. The same thing happens to a greater extent in a class A P-P triode amp. The distribution of the higher order harmonics is dependent on the bias current. A hegher current lowers the higher order harmonics due to the shift away from the non linear region at low currents.



At least here at work, none of the 10 year old schematics show up. I saw some talk of a differential output stage with a CCS in the tail? If so, a 3 stage fully differential amp using 3 LTP's (including the output stage) was one of the sweetest sounding creations I ever made (low power though). Got to hook it back up if I ever get the FH-3's finished.....nearly done after 2 1/2 years.
 
In case you prefer the sound characteristics of SE amp, I would suggest just go with SE amp and get more efficient speakers so that the low power output of SE can be compensated.

Speakers based around GPA604 so 3W SE is fine.

Preference, well, I built a DHT PP amp using SE to PP grid interstage and reconfigured to compare with DHT PSE. The PSE was more to my liking, but it was/is direct coupled (removed the IT) and that might account for some of the preference/difference. There were some things that the PP did better (especially with a CCS tail) - it was a 'bigger' sound, mostly wider, but just doesnt have the tone of the PSE, well most of the tone and you wouldnt know it was lacking until you play the PSE.

The PSE is of course noiser with AC filaments.. something that I expect would be cured with the design in question, and also the opportunity to run the output pair with less Pd for the same output power is appealing.

The comment about less 2H cancellation compared with PP is encouraging. I suspect that I might need to buy the four OPT's and build it - and compare with a SE version of the same. If SE wins out, I can try biamping the duplex speaker utilizing the second OPT pair.


Thank to All,
Shane
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.