interesting article about "transistors vs tubes" in audio

[unclejed613]

please don't start any "flame wars" over this. i tend to agree with the writer of the article, each of the two technologies has it's purpose. each also has it's pro's and con's. it's like my dad used to say..."the right tool for the right job". i just thought it was interesting. the article is "part 1" of a series, and i will post links as the other parts are published. it also has a few links to other articles about components and the way they behave in an audio system.

oops, ya can;t get there from here without having my yahoo ID..... let me see if i can get a copy of it somehow...

**broken link removed**
that should work...

 

[KeepItSimpleStupid]

This is going to get interesting. I still say CD's don't have enough bandwidth. I'll bet if the BW was about 100K they would sound a lot better. It's OK to BW limit on playback to about 40 Khz.

 

[crutschow]

This is going to get interesting. I still say CD's don't have enough bandwidth. I'll bet if the BW was about 100K they would sound a lot better. It's OK to BW limit on playback to about 40 Khz.

Then you should like the "Superaudio CD" (SACD) from Sony. It has a 50kHz upper frequency limit.

If it's OK to BW limit on playback to 40kHz, then what possible improvement could a 100kHz CD bandwidth have? Why stop at 100kHz?

 

[unclejed613]

the extended bandwidth would play nicely through a Lirpa-1 amplifier with a frequency response of DC to Daylight.

 

[KeepItSimpleStupid]

Neat. My audio amp does have an unrolled off -3db point of 800 Khz. It's built like an RF amp with a ground plane and metal film resistors. But it is rolled of to 40 Khz.

My reasoning for high bandwidth is this: It takes about 7 sine waves to make a square wave using Fourier theory, I believe.
This if you can hear a 3 Khz tone easily, then 21 Khz would be necessary to reproduce a square wave at that frequency. That is attack time. e.g. Listening to a hammer hitting the strings on a piano.

We can obviously hear past 3 Khz and even at 10 KHz, that's a 70 Khz bandwidth requirement.

So, the 20-20 Khz frequency requirement is bogus in my opinion. My professional audio stuff goes from 0-100 KHz.

 

[unclejed613]

there's information in audio, that while you can't "hear" it, your brain does seem to process it. this ultrasonic information is part of what the author is calling "sound stage".

 

[bountyhunter]

Then you should like the "Superaudio CD" (SACD) from Sony. It has a 50kHz upper frequency limit.

If it's OK to BW limit on playback to 40kHz, then what possible improvement could a 100kHz CD bandwidth have? Why stop at 100kHz?

There's a job for you somewhere in marketing......

 

[bountyhunter]

there's information in audio, that while you can't "hear" it, your brain does seem to process it. this ultrasonic information is part of what the author is calling "sound stage".

I have been hearing about "presence" and "clarity" and "natural sounding" and every other BS term I can't remember forever...... All I know is, every time they do double blind tests, the "experts" can't accurately and repeatedely detect the differences even so they have to claim that the "switch contacts" affected the test sound. There is no doubt tube amps sound DIFFERENT than solid state and people do get used to things........ but I am unconvinced they actually sound "better" if there is such a thing anyway.

BTW: the "transduceres" (speakers) introduce about 50 - 100X as much distortion as all other components combined... especially woofers that have all sorts of bizarre effects like cone flex. I have seen THD as high as 3 - 4% on low bass notes (like 45 Hz) on "name brand" high end speakers.

A truly good set of headphones are as close as you'll get to actual sound.

 

[bountyhunter]

My reasoning for high bandwidth is this: It takes about 7 sine waves to make a square wave using Fourier theory, I believe.
This if you can hear a 3 Khz tone easily, then 21 Khz would be necessary to reproduce a square wave at that frequency. That is attack time. e.g. Listening to a hammer hitting the strings on a piano.

We can obviously hear past 3 Khz and even at 10 KHz, that's a 70 Khz bandwidth requirement.

So, the 20-20 Khz frequency requirement is bogus in my opinion. My professional audio stuff goes from 0-100 KHz.

This is analogous to "TIM" distortion which is another name for slew limiting. Any true square wave rises faster than an amp can follow, so as the amp slews the control loop is not engaged... and as it settles on the "top" of the sq wave, there is a finite time until all the feedback settles. This is old news, many better amps test and publish TIM specs, and it is probably the one place a sound argument can be made for the superiority of FET or tube amps which use LESS feedback to attain the same THD levels... so degrade less from slew distortion. But IMHO, it's not something 99.99999% of people would ever hear.

 

[unclejed613]

i think the best "challenge" so far was Bob Carver's. he built a solid state amp that emulated a McIntosh amp EXACTLY. he did it by measuring the transfer characteristics of the McIntosh, and emulating them in the feedback loop. there were a few other amps he was able to emulate this way, including a Mark Levinson amp (which was much more expensive than the amp Carver made). i think they were available as the Model 1.5t (the t for Transfer characteristic)

 

[bountyhunter]

I remember that.

 

[KeepItSimpleStupid]

My AMP is specifically a low TIM design with not s lot of feedback. When you have the ability to derivative process the signal with a 4bx (Impact Restoration settings), you can hear the difference. When you substitute a low quality amp for a fast rise time signal, the sound difference is very pronounced.

Good quality, high dynamic range music doesn't play well on an iPod. When the media matches the playback device then we appear to be satisfied.

When you put a full range sounding speakers on a very cheap stereo of the 70's or so, what you hear is hiss. Put crummy speakers on the system and it sounds reasonably well.

At least with an LP, you could have two concentric tracks. It was apparently done on an album and really confused a few people.

 

[Mr RB]

Then you should like the "Superaudio CD" (SACD) from Sony. It has a 50kHz upper frequency limit.

If it's OK to BW limit on playback to 40kHz, then what possible improvement could a 100kHz CD bandwidth have? ...

I agree! The CD itself has actual data stored as 44.1kHz 16bit stereo samples. So the actual waveform of the music is at 44.1kHz, and the closer the amp gets to reproducing any frequencies above 20kHz the only real thing that happens is that the aliasing errors (distortion) of the 44.1kHz data start to be reproduced by the amp.

A really PERFECT amp for playing CDs (or a perfect CD player) would have flat response to exactly 20kHz and completely attenuate any and all frequencies above 20kHz.

"Audiophiles" can be a funny bunch. Most of the new solid state amps have phenomenally low distortion, and accurately reproduce the CD sound, but all speakers have massive distorion and colouration. What people are really listening to are the SPEAKERS. And their own distorted ears.

 

[unclejed613]

CDs and MP3s should be rolled off at 20khz, but if you listen to vinyl or have any magnetic media, i'd keep the extended response available. with a well designed amp, using relatively good transistors, TIM shouldn't be a problem. any audio source material available will generally produce signal risetimes well below the slew rate of a good amplifier. TIM was a real problem up until about the mid to late 70's before the causes of it were understood well enough to do something about it. i worked for an audio company in the late 70's through the early 80's that made a very nice amp. the slew rate was something like 30V/us. the frequency response was flat to 30khz, and there was a portion of the final test procedure that ran the amp at 100khz to make sure the output transistors didn't go into common mode conduction. as a matter of fact, the burn-in test for the amp ran it at 70% full output into a capacitive load overnight at 50khz.

 

[KeepItSimpleStupid]

Mr. RB:

Yep, match the source and playback to the input material.

 

[Nigel Goodwin]

I agree! The CD itself has actual data stored as 44.1kHz 16bit stereo samples. So the actual waveform of the music is at 44.1kHz, and the closer the amp gets to reproducing any frequencies above 20kHz the only real thing that happens is that the aliasing errors (distortion) of the 44.1kHz data start to be reproduced by the amp.

Except that CD Players have a steep roll-off filter for that exact reason, as is standard in D2A systems.

 

[DJDAudio]

Love this topic, marking for future read.

I have both classic, and modern transistor amps as well as tubes, even very old 10+ tube mono preamps. Each have there own characteristics and unique applications.

I even built a 'push pull' amp from a pair of matched modified automotive alternators. As well as built a 'saturable core' Magnetic amplifier for low frequency.

Like my username implies I love this stuff haha


Oh I have seen ultra high end gear $500K+ setups that are 100% Analog, from the marble slab record player to the tube amp. So no digital conversions or sampling losses.
Even herd rumor of laser pickup Players.

 

[tvtech]

Ultra high end stuff comes to mind. Krell afaik is leading edge silicon technology. Audio Research is leading edge tube technology.

Audio Research here: https://www.audioresearch.com/
Krell here: https://www.krellonline.com/

Both these companies were around when I was selling high end audio in the early Eighties. And they are both still doing well.

Different horses for different courses.

On a truly subjective note from what I can remember back then, the Krell KSA 100 blew me away (literally) with it's total control of any speaker/s it had to drive. Rock solid and controlled bass and basically everything else. So smooth yet plenty punch.

The Audio Research never impressed me as much as the Krell. Too smooth. It's a VERY subjective argument depending on what your choice of music is.

Cheers,
TV TECH

 

[tcmtech]

I even built a 'push pull' amp from a pair of matched modified automotive alternators. As well as built a 'saturable core' Magnetic amplifier for low frequency.

I would be very curious as to how that is done! Got any schematics, reference material and or pictures of those things?

 

[DJDAudio]

I would be very curious as to how that is done! Got any schematics, reference material and or pictures of those things?

**broken link removed**
NOTE: This was drawn as the theory before I actually built it, the final circuit was close but with some extra components.


I will find the the Video of it in operation

The Distortion was crazy the fidelity was low, but it one of those things I built because I wondered if it was possible.

The power came from a 2HP AC motor that would belt drive both alternators at the same RPM.

The faster it was spun the more the 'amplification' haha.

Will also dig up some of my old Mag Amp projects.

I will start a new thread for these.