Yes, the harmonics. But are you saying the 2nd harmonic distortion of SE amps improve the timbre, although this would be minimal at low level? What happens to the sound if you turn the volume up to produce say 0.5 to 1% distortion, doesn't the IMD start to muddy it at that point?
what about Berglund's "asymetrical" PI?
Does it really simulate (and ,if yes,how close?) a SE behaviour of a PP AB1 output stage?
Does it really simulate (and ,if yes,how close?) a SE behaviour of a PP AB1 output stage?
I wonder whether the single ended topology has become more popular. Most of the stages claimed to be single ended have push-pull characteristics. The amplifier circuit is simple, but it requires a non-simple power supply.
It's true that the sound deteriorates at higher levels. But there are other factors such as my output tubes have a bit under 18W dissipation. If I listened louder I'd want to take that up to 22-25W dissipation per tube. I generally listen at fairly low levels. I can't deny that when you turn up the volume with SE amps you start to need more and more sensitive speakers. Mine are 89db.
As an aside, I see you're right next to Portmeirion. Exotic stuff for a boy from Bridgend!
.
Last edited:
Then what do you do to the high driver generated distortion? Just leave it there? Why should we choose higher overall acoustic distortion? 😕
The driver isn't a problem. I know of many amps that sound excellent, superb.
Do you aim for the zero distortion amp? Why not go to the big audio stores, they will be glad to sell you the newest Sony, Pioneer, Yamaha, you name it.
All those companies have strived for decades for amps that produce near zero distortions. Go, audition them. Are they the golden shrines of amp building?
Nope. They are awfull sounding, like my PC board soundcard. Build with dozens of SMD operational amplifiers (IC) format. Have beautifull data, big power outputs and nearly zero distortions.
It depends, as we say. You can have bad pp amps and bad designed SE amps.
If an amp is properly designed, it just doesn't matter what principle.
The PP amps have more output power and the SE amps less power. Thats all.
Choose the meal you need and prefer.
Yes. I wondered how it is though that loudspeaker distortions are not such an issue as amplifier distortions. There is a nice summing up in post #11 of this thread Geddes on Distortion perception
> As posted above, SE is only good for certain kinds of music - "girl and guitar". It is not good for complex music, particularly orchestral and choral.
>My principal objection towards SE too. What is worse, ownership of a SE amp / single driver speaker, further limits the listener's taste to really lame music the system likes playing. The descent from Diana Krall into Norah Jones is eventually unavoidable
There are many designs for SE amps circulating around here that probably fail to do justice to loud and complex music. I can only imagine those are what you have heard. The recycled 20-30 year-old 300B-SE circuits that use small power supply caps, cheap OTs, weak driver design with 6SN7 reappear with depressing regularity.
But loud and complex 20th century choral and symphonic pieces for big orchestras are exactly my taste. Mahler №6, Shostakovich №4, Schnittke 'Concerto for Choir', Britten War Requiem are some of the works I can't do without. My 300B-SE plays these with a huge stereo presentation, and conveys the exilaration and terror that make these pieces what they are.
It can be done, but it does require design, and not blind imitation.
I started in the 1990s with circuits that everyone used then. But quickly found almost any carefully designed driver stage is better than 6SN7s, and worked up to my own design: the Shunt Cascode driver Rod Coleman Shunt Cascode - Power Tube Driver. This driver is characterised by instant recovery from overload, and low distortion right up to clipping, with a most natural cadence of harmonics (see linked measurement).
My filament supply everyone knows, but the Anode supply also needed big changes from the blessed norms - especially away from small capacitors and low current-delivery. The Sowter SA08s (25W @ 20Hz 1%) OT completes the built-for-loud design.
>My principal objection towards SE too. What is worse, ownership of a SE amp / single driver speaker, further limits the listener's taste to really lame music the system likes playing. The descent from Diana Krall into Norah Jones is eventually unavoidable
There are many designs for SE amps circulating around here that probably fail to do justice to loud and complex music. I can only imagine those are what you have heard. The recycled 20-30 year-old 300B-SE circuits that use small power supply caps, cheap OTs, weak driver design with 6SN7 reappear with depressing regularity.
But loud and complex 20th century choral and symphonic pieces for big orchestras are exactly my taste. Mahler №6, Shostakovich №4, Schnittke 'Concerto for Choir', Britten War Requiem are some of the works I can't do without. My 300B-SE plays these with a huge stereo presentation, and conveys the exilaration and terror that make these pieces what they are.
It can be done, but it does require design, and not blind imitation.
I started in the 1990s with circuits that everyone used then. But quickly found almost any carefully designed driver stage is better than 6SN7s, and worked up to my own design: the Shunt Cascode driver Rod Coleman Shunt Cascode - Power Tube Driver. This driver is characterised by instant recovery from overload, and low distortion right up to clipping, with a most natural cadence of harmonics (see linked measurement).
My filament supply everyone knows, but the Anode supply also needed big changes from the blessed norms - especially away from small capacitors and low current-delivery. The Sowter SA08s (25W @ 20Hz 1%) OT completes the built-for-loud design.
> There is no need for individual DC supplies for each filament in a PP DHT. AC works fine for outputs and drivers, it just has to be clean filtered AC, and in fully differential stages all filaments can be tied together. SE indeed requires DC regulators, as there is no common mode cancellation. Another disadvantage of the topology.
You can connect the filaments of PP halves in parallel, and it will operate. But the difference in gm at the two ends (caused by the filament voltage skewing the effective bias - Vgk) means that there is a music-signal voltage at the filament ends too. Simply shorting these together is not conducive to good performance - don't forget that the anode currents and hence the differential voltages are in anti-phase for PP pairs, and forcing them to cross-talk degrades the sound.
As for AC-heating: PP may well reduce the hum to levels you can tolerate, but the intermodulation-distortion elephant remains stomping around the room.
No need to take my word for it: the attached picture is a measurement taken by euro21 here:
Sound of 300B SET by Satoru Kobayashi
The intermodulation in the picture is taken at 1W output. The IMD level gets worse rapidly with increasing power, and is a completely unnatural product of the music against a series of 50/60Hz and 100/120Hz (and higher) cross-products. It is easily audible as muddle, fuzz and confusion in loud passages - and badly degrades the presentation of big symphonic pieces.
You can connect the filaments of PP halves in parallel, and it will operate. But the difference in gm at the two ends (caused by the filament voltage skewing the effective bias - Vgk) means that there is a music-signal voltage at the filament ends too. Simply shorting these together is not conducive to good performance - don't forget that the anode currents and hence the differential voltages are in anti-phase for PP pairs, and forcing them to cross-talk degrades the sound.
As for AC-heating: PP may well reduce the hum to levels you can tolerate, but the intermodulation-distortion elephant remains stomping around the room.
No need to take my word for it: the attached picture is a measurement taken by euro21 here:
Sound of 300B SET by Satoru Kobayashi
The intermodulation in the picture is taken at 1W output. The IMD level gets worse rapidly with increasing power, and is a completely unnatural product of the music against a series of 50/60Hz and 100/120Hz (and higher) cross-products. It is easily audible as muddle, fuzz and confusion in loud passages - and badly degrades the presentation of big symphonic pieces.
Attachments
Rod Coleman,
you should be careful when interpreting meaningless graphs and numbers. Intermodulation is not representable.
you should be careful when interpreting meaningless graphs and numbers. Intermodulation is not representable.
It is not 'meaningless' information. It identifies the presence of distortion spectrum that was not present in the recording.
Why is it taken as a given the SE amps have more 2nd H distortion than PP amps? My SE amp on the bench has its 2nd H 30db down from the fundamental. A PP amp with a very slight transconductance tube mismatch can easily have as much or more 2nd H distortion than a SE amp.
I'm late to the party, but agree with Tony T and a few others. Single ended has always been around, it's the original amplifier. The single ended amp with a DHT is father of them all. It's the P-P amp that's the newcomer.
I also agree with Andy. A properly built SET with a robust power supply, well adapted driver tube (often pentode) and excellent OPT isn't restricted to small music. It can do whatever you throw at it, if the speakers are up to the task. That usually means rather large. 😛
I also agree with Andy. A properly built SET with a robust power supply, well adapted driver tube (often pentode) and excellent OPT isn't restricted to small music. It can do whatever you throw at it, if the speakers are up to the task. That usually means rather large. 😛
Rod - I'm interested in this. As you know, I like to use DHTs. Can this be used with lower current DHTs like 01A (4mA) or 26 (7mA) or is it better to use higher current DHTs like 10Y or 2P29L? Given that the gain of these DHTs is around x9 (mu=9), what would be the total gain of the stage using your board? Would a Ri of 9K or so be a problem or not?