Dynaco ST-70 Series 3 Tube Amplifier Review

[antcollinet]

but it was not confirmation bias.

Almost certainly it was. Or possibly a poor downsample, or incorrectly dithered.

 

[rongon]

Almost certainly it was. Or possibly a poor downsample, or incorrectly dithered.

What would incorrectly dithered mean, if I was using POWr at the time?
Steinberg WaveLab was supposed to be the best reasonably affordable audio editor of its time.
It could be that they were in fact inferior tools. Perhaps you know of a study that confirms that idea?

 

[antcollinet]

What would incorrectly dithered mean, if I was using POWr at the time?
Steinberg WaveLab was supposed to be the best reasonably affordable audio editor of its time.
It could be that they were in fact inferior tools. Perhaps you know of a study that confirms that idea?

I've no idea about those tools but good tools still need to be operated correctly. However, if the down conversion was good, I'd give it 99.9% plus it was confirmation bias.

 

[amirm]

I'll bet not a single person here has ever heard a well produced early 1950s 78rpm record (in good condition) played on a truly 'transcription quality' 78rpm playback setup (a truly top-end radio station turntable such as Garrard 301 or similar) the correct type of cartridge/stylus, phono preamp with correct EQ for the record, etc.

I have heard incredibly heroic attempts at LP playback at audio shows. We are talking both vintage and and modern > $50,000 turntables. Rarely am I impressed with such playback experiences. On the other hand, master tapes from the era are excellent and clearly superior to any digital versions.

 

[DVDdoug]

The 100kHz bandwidth through the transformer is impressive! I did a quick search for the schematic and there is feedback from the secondary and that's probably the "trick". I don't know how common that was/in in tube amps but feedback has the potential of making everything better!

I have what used to be a Dynaco Stereo120 (solid state). I blew it up a few times and then I replaced the "guts" when the parts were no longer available.

 

[rongon]

I was talking about pre-LP 78 rpm disks (with a 'k'). Pierre Clément cartridge/tonearm/turntable as used by Radio France in the 1960s. https://www.itishifi.com/archives//2014/09/pierre-clement-turntable.html

I am amazed at how biased against 78s I was because I'd never heard them played back well.

I'm not surprised if you found $50k LP playback systems at audio shows to be underwhelming. It would be interesting to know what you heard that didn't impress you. I can say that if you take a Garrard 301 or 401, carefully rebuild its motor to reduce vibration to absolute minimum, carefully install a good Frank Schroeder tonearm (installation precision is important here), very very very carefully align a good cartridge (using a USB microscope+software-based alignment tool that includes oscilloscope observation), you can get things sounding pretty darn good. But as I very carefully mentioned previously, it will be objectively inferior to the best digital playback of the best digital source materials. No contest. But a really fine LP playback system should not cost $50k. Mine cost <$1k total (but I did put a lot of work into getting it playing well), and it brings me lots of enjoyment.

I know there is only so much you can do with LP. It's a limited format. It can't hold a candle to the original R2R tape played back on a newly cleaned and aligned tape deck with good electronics. Open reel tape was the deluxe consumer format in the golden age of hi-fi. LPs were for the masses, before cassettes came to be.

My Raspberry Pi digital player must be truly horrible if my sub$1k turntable/tonearm/cartridge setup can bring me any kind of listening pleasure at all. It makes you wonder how bad my hearing is, or what kind of bad taste I must have.

Some 1950s master tapes are quite amazing. Some not so amazing. I wouldn't say all master tapes from that era were excellent, but they were Hi-Fi, ya know?

 

[rongon]

The 100kHz bandwidth through the transformer is impressive! I did a quick search for the schematic and there is feedback from the secondary and that's probably the "trick". I don't know how common that was/in in tube amps but feedback has the potential of making everything better!

I have what used to be a Dynaco Stereo120 (solid state). I blew it up a few times and then I replaced the "guts" when the parts were no longer available.

A feedback loop from the speaker secondary to the cathode of the input stage was how it was done in just about every tube amp sold after 1950. Is that what was done in this now-defunct Dynaco amp?

Negative feedback from a dedicated secondary winding was how McIntosh did it in those battleship amps like the MC225 and MC250. They also used a separate primary winding for a different feedback loop. There was a LOT of engineering thrown into top quality tube amps back before transistors were readily available. NFB from an additional secondary winding wasn't that uncommon.

The harman-kardon Citation II used 'nested' feedback loops, with local NFB around the driver stage, NFB around the output stage, AND a global NFB loop from the speaker terrminals to the input. I scored one of those when I was very young, but didn't have the where-with-all to refurbish it, so I sold it for a profit. Too bad, it turned out to be one of the most valuable of all vintage amps from the golden age of hi-fi. Audio engineers had gotten quite sophisticated with use of NFB in tube amps by the early 1960s.

But 100kHz bandwidth from an output transformer was and still is unusual. In this deceased Dynaco amp, was that closed loop, or before NFB was applied?

I heard quite a few Dynaco Stereo 120 amps back in the day. That was not one of the ones I liked. From those days I thought the h-k Citation 12 and the Hafler DH-220 were better sounding. I'd like to hear a B&K ST-140. Those were supposed to be very good. I remember hearing one in a hi-fi shop back in the 1970s, and I remember thinking it was pretty okay. But I can't tell you how it sounded.

 

[SIY]

I don't know how common that was/in in tube amps

Nearly 100% of the amps of the '50s-'70s.

edit: Oops, should have seen Rongon's post. Sorry.

 

[SIY]

Negative feedback from a dedicated secondary winding was how McIntosh did it in those battleship amps like the MC225 and MC250. They also used a separate primary winding for a different feedback loop. There was a LOT of engineering thrown into top quality tube amps back before transistors were readily available. NFB from an additional secondary winding wasn't that uncommon.

The harman-kardon Citation II used 'nested' feedback loops, with local NFB around the driver stage, NFB around the output stage, AND a global NFB loop from the speaker terrminals to the input. I scored one of those when I was very young, but didn't have the where-with-all to refurbish it, so I sold it for a profit. Too bad, it turned out to be one of the most valuable of all vintage amps from the golden age of hi-fi. Audio engineers had gotten quite sophisticated with use of NFB in tube amps by the early 1960s.

The H-K was excellent. Ditto the Mac amps (I'm restoring a pair at the moment). The two other circuits I would tell every serious student of tube amps to study are the Lux 3045 (a Tim deParavicini design) and the 50 watt amp at the back of the RCA tube manual.

 

[375HP2482]

Why? In the klippel distortion tests, the vast majority of people cannot detect distortion below around -45dB. With speakers, I'm struggling at around -25dB

Most people are vastly overestimating the audibility of tiny amounts of distortion such as 0.1% (-60B)

(Putting aside that SNR is about noise, not distortion - I'll assume you meant SINAD)

You will notice noise long before THD. At full output this amplifier exhibits noise at CD levels.

 

[anmpr1]

Tubes are not and never were 'user friendly'.

I don't know. Back in the day when your device was acting up (radio or television) you could easily pull the tubes and take them to the local Rexall. Insert each into the tester (for free), find the bad, and then the clerk would open up the cabinet door on the bottom of the tube tester, and sell you a new RCA at a reasonable cost.

Can't do that with SS. When SS came around you saw stickers on the back of the device telling you not to open it, because there's no 'user serviceable' parts inside. That was the end of 'user friendly'.

Of course you can't do that with tubes either, today-- because the drug store doesn't have a tube tester.

From an engineering standpoint, even bothering to 'test' and report on tube gear approaches silly. At least if you have expectations. It's not like anyone anywhere is going to learn something they didn't already know, or at least suspect. Aczel refused to even get involved in the scene, because he was more interested in SOA (or at least value proposition) engineering. His (paraphrased) quip (actually written when he was more approachable to the tube scene): If I'm going to investigate the fastest land animals, I'm looking at the big cats, maybe a thoroughbred racehorse, some hounds ... I'm not going to waste my time with the sloth and donkey.

However the fact that old Dyna (and some other) machines are still in service and available to update 75 years after the fact? That is the better and really more important question to contemplate.

 

[Petrushka]

I don't know. Back in the day when your device was acting up (radio or television) you could easily pull the tubes and take them to the local Rexall. Insert each into the tester (for free), find the bad, and then the clerk would open up the cabinet door on the bottom of the tube tester, and sell you a new RCA at a reasonable cost.

Can't do that with SS. When SS came around you saw stickers on the back of the device telling you not to open it, because there's no 'user serviceable' parts inside. That was the end of 'user friendly'.

Of course you can't do that with tubes either, today-- because the drug store doesn't have a tube tester.

From an engineering standpoint, even bothering to 'test' and report on tube gear approaches silly. At least if you have expectations. It's not like anyone anywhere is going to learn something they didn't already know, or at least suspect. Aczel refused to even get involved in the scene, because he was more interested in SOA (or at least value proposition) engineering. His (paraphrased) quip (actually written when he was more approachable to the tube scene): If I'm going to investigate the fastest land animals, I'm looking at the big cats, maybe a thoroughbred racehorse, some hounds ... I'm not going to waste my time with the sloth and donkey.

However the fact that old Dyna (and some other) machines are still in service and available to update 75 years after the fact? That is the better and really more important question to contemplate.

Because good amplifiers reached a point of diminishing returns 70 years ago, in terms of transparency. And there has always been a market for nostalgia and antiques.

 

[rongon]

I'm not sure why I feel like writing about this, because I know what I'm thinking is not scientifically based or even verifiable, and is likely to get blasted to shreds here, but...

1. People who like to listen to music through a tube amp typically don't care about their poorer objective performance. As long as the distortion and noise are not subjectively objectionable then the amp in question passes muster.

2. People who listen to music through a tube amp are usually looking for a smoother presentation of the upper midrange and low treble frequencies. When I was young and had saved up my first few hundred dollars to squander on a 'component' hi-fi system, I went about auditioning power amps and found I preferred the sound of a pair of then 20-year old Dyna Mk3 tube amps over an Adcom GFA-35, because the sound of Elvin Jones's cymbals sounded more natural to me through the Dyna tube amps. (I really wanted a McIntosh MC-225, but that was way beyond my budget.) I should have stopped right there, because those were pretty darn good amplifiers, but you know, I got sucked into the hobby and those amps are long gone.

So what makes the sound from a tube amp that ever-so-slightly different that some people find attractive? I'm starting to think that it's not so much the tubes but the 'sound' of the output transformer. The audio transformer is a very flawed device compared to things like capacitors, resistors, opamps, MOSFETs, even vacuum tubes. The transformer will have restricted bandwidth because of limited primary inductance, unavoidable interwinding capacitance, unwanted leakage inductance, insertion loss, etc. An audio transformer will introduce additional harmonic distortion, although high quality ones keep this very low. The output transformer is also a very expensive component, typically retailing for >$100 USD each (and you need two for stereo). With all that, why bother? Well...

If you visit a recording studio you're likely to find sound processing equipment that's loaded with audio transformers. Transformers can convert unbalanced signals to balanced and back again, and they provide isolation from DC (break ground loops, prevent hum). However, all that can be done with opamps too, much more cheaply, and technically better. So why do recording engineers use transformers? Often, it's for the sound. Really good audio transformers give a nice something to the sound. I believe whatever it is that they do is at work when you listen to your DAC through a stereo tube amp like the Dynaco Series 3 discussed here. The performance is not better, but the slight change to the sound brought by vacuum tubes pumping the audio signal through transformers may be to your liking.

If the noise from the amp is at -95dB at full power output (about 20W rms) then it's likely to be inaudible at the listening position through loudspeakers in a typical listening room. We don't listen to the speakers with one ear crammed into the woofer voice coil. Once again, a good Class D amp will perform much better, but I have definitely noticed a different quality to the sound of cymbals, sibilants, etc. coming from my system with a Behringer A800 (class D) in it than with a Hafler P1000 amp (class AB MOSFET) driving the speakers. The A800 definitely lends more of a sizzle to the high frequencies. For $200 I am NOT complaining, but that's what I hear. The Hafler P1000 sounds 'darker', with less sizzle to the high frequencies (although it does sound a bit 'murky', probably due to higher levels of THD and therefore IMD). I suppose the Hafler amp is objectively much worse than the Behringer amp. Yet I enjoy listening through the Hafler amp more. I guess I have poor taste and you can safely dismiss anything I have to say.

Also...

the fact that old Dyna (and some other) machines are still in service and available to update 75 years after the fact? That is the better and really more important question to contemplate.

Wholeheartedly agreed! Turntables and vinyl records are like that too. I have CDs from around the year 1990 that failed and are no longer playable, but I have a wall unit full of LPs that I bought 20 to 50 years ago that I can spin whenever I feel like it, and they play pretty much the same as they did when I got them. Digital files? You know what they say... Your digital file doesn't fully exist until you have three copies: 1) the working file, 2) the backup, 3) the failsafe backup. We've all had hard drives fail on us, and I've had several USB thumb drives suddenly die for no apparent reason (one I dropped and that killed it). I have two vintage turntables that are about as old as I am and they both work fine and sound really good. They've also appreciated in value by 4X over what I paid for them (including refurbishment costs).

Because good amplifiers reached a point of diminishing returns 70 years ago, in terms of transparency.

Wait... The accepted state of the art in audio power amplifiers in 1954 (70 years ago) was the Williamson amplifier, boasting 10 watts class A per channel from a pair of triode-wired KT66 *tubes*. (Transistors would not become widely available for another 10 years.) Are you saying there's been diminishing returns ever since? If so, then why not listen to a tube amp if that's what one prefers?

And there has always been a market for nostalgia and antiques.

Let me put it this way. Since we're talking about luxury/leisure products (nobody actually needs a hi-fi), let's use cars as an analogy. For pure driving fun, would you rather be driving a fully restored and tuned up Porsche 356 or Fiat Spyder convertible, or would you prefer driving a Mazda Miata, or even a Tesla Model Y? The contemporary cars are far more powerful, hold corners as well or better than the old sports cars, and are technically superior in just about every way. But fun to drive? Now we're getting subjective.

If I ran a recording studio, I'd be using Genelec monitors at the mixing console and perhaps JBL M2 for the big monitors. I'd be recording to DSD or perhaps 24/192k PCM, or maybe future-proofing by recording to 32/384k to keep up with the Joneses. Modern stuff, so that the engineers and clients can hear exactly what's going on.

But listening purely for my own pleasure? That's different.

 

[SIY]

So what makes the sound from a tube amp that ever-so-slightly different that some people find attractive?

Assumes facts not in evidence.

let's use cars as an analogy.

NOOOOOOO!!!!!

Are you saying there's been diminishing returns ever since?

If by that you mean "electronics as solved problem," yes. What we do better now is power, source impedance, and efficiency, but in a blind test against a modern amp where frequency response (due to source impedance) and levels are matched, I will bet that you won't be able to tell a difference as long as neither amp is overloaded.

 

[rongon]

Assumes facts not in evidence.

Not in evidence to you.

If by that you mean "electronics as solved problem," yes. What we do better now is power, source impedance, and efficiency, but in a blind test against a modern amp where frequency response (due to source impedance) and levels are matched, I will bet that you won't be able to tell a difference as long as neither amp is overloaded.

I'd gladly submit to a blind test and let myself be proved wrong. Where can I find a blind test to participate in?

I have a modern amp, a Behringer A800. Practically new.
I have a pair of Snell E-III speakers which are easy enough for a tube amp to drive. Those would be a constant in the test.
I don't have a tube amp at the moment. I guess I need to build a decent push-pull UL EL84 or EL34 amp.
Once that's done, I'd need to figure out a way to quickly switch the amps in/out without stressing the speakers or the amps.
Then I'd need to enlist a friend to administer the test. That's the hard part, along with building the tube amp. Maybe I can borrow a friend's tube amp...

 

[SIY]

Not in evidence to you.

Or anyone else. After decades of claims, not one successful demonstration of a "tube sound."

I'd gladly submit to a blind test and let myself be proved wrong. Where can I find a blind test to participate in?

You can do it yourself. It's much easier these days with superb ADCs and comparison software.

 

[rongon]

Or anyone else. After decades of claims, not one successful demonstration of a "tube sound."

I'm not aware of any attempts at a rigorous demonstration of a "tube sound" vs. a "solid state sound". Lots of anecdotal claims, etc. but not anything systematic.
Has a study been done on this that I'm not aware of? If so, can you point to an original source for that, so we all can read it?

You can do it yourself. It's much easier these days with superb ADCs and comparison software.

The ADC is not a problem, but what superb comparison software is recommended for this kind of testing? foobar2000's ABX plugin?

I'm not sure doing any testing on my own would be received well here. If in the end I claimed to be able to hear differences, I'm sure that result would be dismissed as biased, etc.

 

[Petrushka]

I'm not aware of any attempts at a rigorous demonstration of a "tube sound" vs. a "solid state sound". Lots of anecdotal claims, etc. but not anything systematic.
Has a study been done on this that I'm not aware of? If so, can you point to an original source for that, so we all can read it?


The ADC is not a problem, but what superb comparison software is recommended for this kind of testing? foobar2000's ABX plugin?

I'm not sure doing any testing on my own would be received well here. If in the end I claimed to be able to hear differences, I'm sure that result would be dismissed as biased, etc.

The reason private listening tests are discarded is because they have been put to well controlled double blind tests for decades, and do not hold up. Devices that measure the same cannot be distinguished in controlled tests.

For the first half of my life, ESP was taken seriously. I haven’t seen it mentioned in a couple of decades.

 

[SIY]

I'm not aware of any attempts at a rigorous demonstration of a "tube sound" vs. a "solid state sound".

That's because it doesn't exist. People who insist it does, upon whom burden of proof rests, have failed utterly. I was one of them, couldn't tell the difference once levels were matched and I couldn't peek. I am also unable to find rigorous demonstrations of the existence of elves.

 

[Art of sound]

Why? In the klippel distortion tests, the vast majority of people cannot detect distortion below around -45dB. With speakers, I'm struggling at around -25dB

Most people are vastly overestimating the audibility of tiny amounts of distortion such as 0.1% (-60B)

(Putting aside that SNR is about noise, not distortion - I'll assume you meant SINAD)

I can see the harmonics even in the single 1Khz tone at 2k 3k etc and the noise at around 90db or so which i would consider the real distortion of this product if i have to rely on multi-tone and single tone SINAD graph for a product. If the analyzer can be set to another mode and it ignores the harmonics itd be a more reasonable score for tube products IMO