Review and Measurements of Dynaco ST 70

[anmpr1]

I'm with SIY. I like my tube amps but not for distortion. The definition of 'tube sound' seems to generally be mushy and distorted but a well designed amp that is playing within its reasonable power limits can sound clean and detailed.

Back in the day, the late Harvey Rosenberg of NYAL Futterman OTL fame (more salesman than engineer) had a dozen reasons why tubes were the beez kneez. Harmonics were the key, he argued. With this in mind, he was fond of citing a May 1973 JAES paper by Russell Hamm (Sear Sound Studios, NYC): Tubes v Transistors--Is There an Audible Difference?

I have a copy of the paper, which begins by stating how tube consoles typically sounded less harsh than SS, and had more subjective dynamic range, making the final recording more engaging. In order to find reasons, his team began a pretty rigorous preamplifier measurement protocol, comparing triode tubes (12AY7-12AX7-8628/7586 nuvistor), hybrid opamp (709-LM301), and transistors (2N3391A-2N5089-2N3117 silicon NPN). All were measured, and listened for distortion characteristics. Their conclusions were that the presence of harmonic distortion (espeically third) during overload sunk SS preamplifiers, from a sonic perspective. From the paper:

Overloading an operational amplifier produces such steeply rising edge harmonics that they become objectionable within a 5dB range. Transistors extend this overload range to about 10dB, and tubes within 20dB or more.

In addition, tube preamps 'sounded' subjectively louder due to their unique distortion characteristics, when overdriven. All the devices sounded good when not overdriven.

 

[30 Ounce]

I have enjoyed reading these forums for a while now but had to sign up to respond to this thread. I have an Aragon 4004 that has been my main amp for 20+ years and recently decided to build a kit tube amp for fun. The kit I chose was the VTA M-125 from http://www.tubes4hifi.com/bob.htm#M125 . He also has upgrade kits and parts for the venerable Dynaco St-70 and a solution for the no longer produced 7199 tube. I can honestly say I was not prepared for how much better it sounds than my Aragon. The first thing that struck me was how 3 dimensional and deep the soundstage was. It has a more detailed and punchy midrange that gave new meaning to micro dynamics. The treble was also incredibly detailed and had a realism that my Aragon, with it’s excellent measurements, didn’t portray. I use to be one that trusted measurements and thought if it measured well it should sound good but after this experience I’m going to trust my ears for future purchases. I’ll still trust measurements to weed out obviously poor performing products but if it meets a certain threshold and someone else recommends it I’ll give it a listen.

 

[DonH56]

Tube stages are usually single ended and generate higher second harmonics than third, unlike most SS outputs. But the big thing cited in those papers is their soft saturation characteristics. When they are overloaded they do not immediately spray harmonics as high as SS circuits, for a lot of reasons (low feedback and soft clipping due to limited gain among them). If you don't significantly clip them a SS amp wins for performance. Tubes sound great to many folk because they tend to have higher harmonics, particularly second, and it has been shown that people like that. It is distortion and I'd rather my components not add that. Their midrange tends to stand out because the lows and highs tend to be rolled off, and they tend to "fill in" the music due to their higher distortion, a distortion most folk like. Me too, but after years of tubes, I decided I prefer a more accurate representation of the source material. It's a nice warm fuzzy sound but after a while to me it just wasn't as realistic.

 

[SIY]

I've done a lot of clipping and overload testing of tube amps and have not seen that the overload is any "nicer" than modern solid state amps (the Hamm paper is quite old and used opamps that were notorious for bad clipping behavior). Unless a tube amp is specifically designed to avoid it (most aren't), they feature a particularly bad kind of overload called "blocking," which turns a transient clipping event into a longer and very audible choking. The little tube amp I use is designed for essentially instantaneous recovery from overload, but had to incorporate a lot of design features not commonly seen in order to achieve that.

Florid descriptions are common with sighted listening, especially after thousands of dollars are invested.

 

[30 Ounce]

If anything I was surprised at how un-tube like the sound was. It is definitely not warm and fuzzy. Your in Monument...I’m in Milliken about an hour away...come listen for yourself.

 

[DonH56]

I've had tube amplifiers, many of them, and finally sold my old ARC D-79 some years back. I have an old Eico EL34 amp for nostalgic reasons.

Hard clipping is about the same for either type of amplifier (it ends up being a square wave after all) but the curve is much softer as you enter clipping for a tube amp. Recovery from severe overload for tube or transistor amps can be... interesting. Since much of my career focused on high-speed data converters and support circuits (like analog amplifiers, filters, mixers, etc.) I am pretty sensitive to overload recovery and hysteresis,

When the early studies were done a number of op-amps and some complete amplifiers exhibited absolutely horrid behavior when clipped (many of us will remember those that inverted output polarity, such a feature, and a goodly number broke into oscillation). Tube circuits often had long recovery tails (thermionic operation and all that). Those days are essentially gone, or so I hope.

Decades ago I built a differential tube preamp. It had the linearity of a SS preamp with softer overload more characteristic of tubes. It had wide bandwidth, low noise, and very low distortion for the time (~0.001% at 5 Vrms output). But, being differential, the even-order harmonics were suppressed, and it was panned for sounding "too solid-state". Lesson learned...

 

[anmpr1]

"DonH56,
Decades ago I built a differential tube preamp. It had the linearity of a SS preamp with softer overload more characteristic of tubes. It had wide bandwidth, low noise, and very low distortion for the time (~0.001% at 5 Vrms output). But, being differential, the even-order harmonics were suppressed, and it was panned for sounding "too solid-state". Lesson learned...

Reminds me of an anecdote. It was at one of the 'high-end' audio shows. Mark "I never met a preamp that was too expensive" Levinson was showing off his new Red Rose tube amp. I think it was a Chinese sourced product, but I'm not sure about that. Anyhow, Harvey "the T in P.T. Barnum stands for tubes" Rosenberg happened to walk by, listened for a while and quipped something to the effect of, "Congratulations, Mark. You've made your tube amp sound just like your solid state amps." Who knows if any of that story is true, but it sounds like something Harvey would say.

 

[Labjr]

I'll bet that Eico 5ARZ/GZ34 is a Mullard tube. About the very best 5AR4 type ever made. I have a stash of them for my Wavelength Cardinals.

 

[SIY]

What makes a rectifier "the best"? Really, it's all about voltage drop and sag, there's not much else a tube rectifier really does. And no tube rectifier can outperform a cheap diode.

 

[Labjr]

I'm not quite sure what makes the Mullard 5AR4 so good. Maybe the metallurgy. But try to buy one? Far as I can tell Everything Wavelength Audio makes uses tube rectifiers. Some even more exotic than the 5AR4. I've impressed a lot of listeners. BTW all diodes aren't the same. Some are real noisy and hard sounding. There's more to it than just the specs.

 

[SIY]

That's not what I was asking. A rectifier rectifies. What property of a rectifier with the name "Mullard" stamped on it makes it superior to one with a different name stamped on it? How is that determined?

BTW, I looked up that amplifier. Oh my. The rectifier is the least of your worries there, the performance is somewhat lower than dreadful.

 

[Labjr]

You seem like someone who just criticizes everything. Come back and comment after you've listened.

 

[DonH56]

There is more variance in tubes than in transistors IME. Different materials, build quality in the creation and construction (spacing etc.) of the tube's elements, filament current, coating quality of the elements within the tube, etc. cause variations. In addition to noise and large IR drop poor tube construction can also introduce microphonics that can increase ripple (though in the power supply decoupling and filtering should take care of that). Some manufacturers have a reputation for better performance and Mullard is one of those.

Specific to a rectifier, saturation current, switching speed, forward IR drop, and noise are among the factors to consider. Many of which are more variable and have a greater impact on performance for a tube (thermionic device) rectifier than for a typical solid-state diode. One of the big advantages of a tube rectifier is that it controls in-rush current. In a tube design, tube rectifiers also tend to be more reliable IME than SS rectifiers, probably due to the combination of high voltage and fairly large in-rush current. It is also better to wait until the filaments are warmed up before applying B+ (high voltage) to the plates. My tube designs (audio and RF) used SS rectifiers for the most part but with time delay and soft-start circuits to improve reliability and tube longevity.

IME/IMO - Don

 

[SIY]

It is also better to wait until the filaments are warmed up before applying B+ (high voltage) to the plates.

I often see this asserted, but there seems to be no data indicating that it matters much- and given the number of amps like Marantz 9 and 8B and Dyna SCA35 out there that seemed to have fine tube life, I'm skeptical. Where it really matters is capacitor voltage ratings, since the B+ goes high at startup before the heaters have a chance to warm and the tubes draw current. That's why you can't just drop SS rectifiers into something like a Stereo 70.

I've worked with a lot of 5AR4/GZ34, and there's nothing special about any particular brands- there's more variation from tube to tube than brand to brand, and of course "brand" doesn't mean "manufacturer."

 

[SIY]

You seem like someone who just criticizes everything.

I criticize poor designs with bad performance. There's very good tube equipment out there, this amp is clearly not among that group. Here's the measurements, for those interested. They might be the worst I've ever seen.

 

[DonH56]

I often see this asserted, but there seems to be no data indicating that it matters much- and given the number of amps like Marantz 9 and 8B and Dyna SCA35 out there that seemed to have fine tube life, I'm skeptical. Where it really matters is capacitor voltage ratings, since the B+ goes high at startup before the heaters have a chance to warm and the tubes draw current. That's why you can't just drop SS rectifiers into something like a Stereo 70.

I've worked with a lot of 5AR4/GZ34, and there's nothing special about any particular brands- there's more variation from tube to tube than brand to brand, and of course "brand" doesn't mean "manufacturer."

I haven't worked with tubes in years. And I was specific to say they have a "reputation" for being better. Most of the time "better" in a tube means "lower noise" to an audiophile though IME (ancient) better manufacturers (not brands, a worthy distinction you made there and I agree) had better process control of all parameters like gain (mu), matching of both sides of a dual tube, and such, not just noise.

My experience with B+ problems was mostly in RF transmitter tubes; we were always cautioned not to apply B+ until filaments warmed so cathode bias was established as it was applied. Arcing could occur internally otherwise. I don't pretend to know the physics behind it (too long ago, and scary to think that my electronics course texts included tube circuits!) but have very vivid memories of hundred-watt driver tubes and kilowatt final PAs arcing into oblivion when the delay circuits failed. Thinking back, I am not sure I ever saw that in an audio circuit. I don't know why, whether they are more tolerant of it, because the voltages are lower (100's of volts instead of kilovolts or tens of kV), or the arc failures I saw were due to something other than simply applying B+ too early. It may well be that the very high-voltage supplies had extremely poor load regulation ( I can attest to that). That would tie in with your comment that B+ ran high until the load was established. I have seen 15 kV nominal supplies spike to 20 kV unloaded so maybe that was the whole problem leading to arcing (?)

Way too tired to think tonight -- 60+ hours this week and need to work tomorrow. Eventually I can retire and have time to think about stuff. Mainly stuff I've forgotten...

 

[SIY]

Yeah, once you're in the multikilovolt regime, cathode stripping and arcing become a real issue. Ditto flashing in direct coupled circuits unless the designer was aware enough to put in clamping diodes or something similar.

Get some sleep, buddy.

 

[levimax]

That's not what I was asking. A rectifier rectifies. What property of a rectifier with the name "Mullard" stamped on it makes it superior to one with a different name stamped on it? How is that determined?

BTW, I looked up that amplifier. Oh my. The rectifier is the least of your worries there, the performance is somewhat lower than dreadful.

The rectifier tube is a Mullard from Blackburn UK... While I did not notice any changes in the sound I did notice that instead of replacing the rectifier tube every few months (and blown fuses and some times output tubes) the amp has ran trouble free for over 2 years.... There really is a huge quality difference between good old rectifiers and new ones

 

[levimax]

What makes a rectifier "the best"? Really, it's all about voltage drop and sag, there's not much else a tube rectifier really does. And no tube rectifier can outperform a cheap diode.

IMO it is a big mistake to just replace a rectifier tube with a diode in these old tube amps, voltages will be too high and you will get ringing in the power supply. If you make other changes to lower the voltages and add snubber networks etc. it can work but it is not a "drop in" replacement. There is also some "bounce" in a rectifier tube that is different than a diode which can effect the sound.

 

[SIY]

IMO it is a big mistake to just replace a rectifier tube with a diode in these old tube amps, voltages will be too high and you will get ringing in the power supply. If you make other changes to lower the voltages and add snubber networks etc. it can work but it is not a "drop in" replacement. There is also some "bounce" in a rectifier tube that is different than a diode which can effect the sound.

Not in competent designs, no.