Tube amps VS. Solid State amps (stereo)

[Rick Denney]

So, I have been thinking about picking up a pre-amp on "that auction site." As part of this I have been considering selling the Carver and trying to pick up a Marantz or McIntosh integrated tube amplifier. So, what do you guys think? Should I-- sell it and get a tube amp, just try to get a decent pre-amp? Or does it even matter, since most of my music is on CD?

Personally, I don't think you'll be able to tell the difference.

Now, for power amps used for a public address system, tubes have an advantage. They can produce more power per dollar than can amps with transistors. And for musical instrument amps that are looking for a particular vintage sound, tube amps still have a devoted following.

But think about this: Tubes require very high voltages, typically in the range of 1500-2500 volts. That puts a lot of stress on components, especially if something goes pop. And it means there are things you can touch and be instantly killed for doing so, even after the amp has been powered down. Making them really linear is not that easy, and their appeal in guitar amps is more about their smooth punch than about their realistic linearity. And vintage electronics have a lot of old electrolytic capacitors in them that steadily degrade over their lives.

There is a middle ground. When Carvers first came out, they championed the use of switching power supplies instead of linear power supplies, and that's why they are relatively small and light for their power output. The switching power supplies require a lot of electronics to make them clean for analog amplification. I had an old 100-watt Spectro-Acoustics solid-state power amp with a linear power supply, and the transformer had the power rating to provide that punch. It was about as heavy as a car battery, too. I replaced it with a 100-watt Carver with its switching power supply, which though rated for the same power output didn't have the same ability to reproduce thumping transients.

But the Spectro Acoustics amp died, and the Carver is still going, 20 years later.

If you want a solid-state amp with the required punch, go to Guitar Center and look for a power amp for a PA system. It won't be as clean, but it will have more oomph. I doubt you'll notice the difference in total harmonic distortion, but the improvement in transient response and slew rate will be more noticeable.

I am currently repairing a 600-watt radio amplifier which is similar to an audio power amp except for the frequency involved and the range of frequencies being amplified (a radio amp only needs a linear pass band of about 3 kHz, instead of the 20 kHz needed for good audio). It uses the lowest-power electron tubes still available, which is an 811A triode. These tubes are popular for guitar amps, and they are relatively cheap at 20 bucks each. But I still have to wait an hour before removing the lid, and I still have to use a chicken stick to make sure the power-supply capacitors for the plate voltage are fully discharged. Replacing the tubas takes 30 seconds, but doing it safely requires an hour or two. The 1800 volts plate voltage produced by the power supply can be instantly deadly.

Transistors generally run on low voltage. Even in high-power applications, they rarely need more than 50 volts. And they are much more broadbanded, making it possible to make them linear over a wider frequency response without a big loss of efficiency.

We are actually approaching a time when discrete power transistors might be as hard to buy as electron tubes.

Rick "noting that solid-state RF amplifiers generally cost twice as much per watt as amps with vacuum electron tubes" Denney

[SplatterTone]

I am currently repairing a 600-watt radio amplifier which is similar to an audio power amp except for the frequency involved and the range of frequencies being amplified

... and the lack of the approximately 100 pound impedance matching transformer depending on fidelity sought.

[iiipopes]

Uh, no on the bias voltage. The standard audio power tubes most often used have anywhere from 250 to 600 volts plate when operated in push-pull mode.

And as far as tonality, so long as the components are operated in their linear range, there should be no audible difference. But in reality, everything peaks. When components peak, solid state components, with the possible exception of some MOS-FET components, tend to clip right off in more of a square wave, generating odd order distortion components, which the ear usually perceives as harsh, while tubes try to pass the signal then fall off as the bias current can't keep up with the signal coming off the cathode, which has the effect of generating more of a compressed sawtooth wave, which generates even order distortion components, which the ear usually perceives as warmth. In a different context, that's why most guitar players who play with some aspect of distortion, from mild warmth to moderate crunch to even high gain saturation, prefer tube amplifiers, including me.

I hate to quote Wiki, but it actually has a pretty good summary on this subject: http://en.wikipedia.org/wiki/Tube_sound

[SplatterTone]

Many moons ago, in the golden age of the Hi-Fi shop (we're talking 1970's) when just about every square mile of every city had a Hi-Fi shop, and working at a Hi-Fi shop was cool, there was fierce debate between the pragmatists and the esoteric crowd. Then Stereo Review did a big series of double-blind tests with the subject audience consisting equally of esoteric and pragmatic members. When the results were published, a plain old $200 Pioneer receiver fared just as well as the big bucks esoteric stuff.

Now that Stereo Review is long dead, one sees a resurgence in the esoteric goblety gook.

If all that is wrong with your pre-amp is a scratchy volume control, that should be cheap to fix.

By the way, if you tried to pump 600 watts of class AB audio through four 811's (like the Collins 600-watt radio (intermittent radio, that is) amplifier), my guess is they would last about 5 seconds ... maybe. And the required impedance matching transformer alone would exceed the cost of a solid state amp capable of 600 watts. The biggest tube audio amp I ever saw was the 350 watt monaural monster made by McIntosh back around 1970. The thing weighed around 120 pounds (big power transformer, big impedance matching transformer) and cost around $1000 (that's in 1970 dollars).

[iiipopes]

What he said. Carver was and is one of the top lines. A Carver power amp is even used to drive the 32' Subbourdon electronic emulation stop on a local church's large Casavant pipe organ.

[Rick Denney]

Uh, no on the bias voltage. The standard audio power tubes most often used have anywhere from 250 to 600 volts plate when operated in push-pull mode.

I dunno. The idle plate voltage of an 811A triode is 1800 volts, and 1500 volts under load, and that's the smallest tube now used in RF amplification. (Everything at about 400 watts or less is done with field-effect transistors.) For the size envelope needed for audio, I would expect you could get at most between 100 and 200 watts in a Class A or AB push-pull arrangement using two tubes. The 811A's would be good tubes for very large audio amps, it seems to me--amps large enough to compete with big FET solid-state amps. But, yup, the output impedance-matching transformer would be a big sucker.

I looked at the specs for several classic vacuum tubes used in audio, like the EL84 and the KT88. Yes, they had plate voltages in use of hundreds rather than thousands of volts, but they also had power outputs when used in Class A or in push-pull arrangements of maybe 10-25 watts. If that's all the power you have, you might indeed hear that amp driven to clipping on the peaks, especially with modern low-efficiency (but high linearity) loudspeakers. A bigger amp of whatever technology would not be driven to clipping nearly as easily.

And size does matter when it comes to realistically high listening levels with modern speakers. My own standard for power output is that I can play an orchestral recording at similar listening levels to sitting in the orchestra, so I can play the tuba with the recording and normal playing levels. You won't meet that standard with a 25-watt amp.

My Spectro Acoustics FET amp with its linear power supply was conservatively rated at 100 watts per channel. My Carver is rated at 75 watts a side, and it provides noticeably lower listening levels, and it only marginally meets my standard. The lower power alone does not explain that, it seems to me. My relatively cheap Samson PA power amp is rated at 250 watts a side, and it fairly blows the Carver away. The Carver has better linearity at close to its power limits, and it's no doubt cleaner, but the Samson is rarely driven to a large fraction of its rated power and so it is never asked to demonstrate how well or poorly it clips. Dynamic range means more than total harmonic distortion, when it comes to realism.

Musical instrument amps are a wholly different application. They are dealing with the dynamic range of the instrument itself, not the dynamic range of an instrument that is recorded, processed, mixed down, and reproduced through several generations before being put in a CD player that applies further limits on dynamic range. The peaks in musical instrument amps are much sharper and have a much fatter envelope, and test the limits of an amp to a much greater extent. Of course, tube-type guitar amps are anything but linear, and that's very much on purpose. The sound it makes while clipping is a distortion that is part of its characteristic sound. My 75-watt bass amp makes a biggish sound, but only with a frequency response at peak power appropriate for an electric bass. It's definitely driven into non-linearity, especially with respect to a 20-20K frequency response.

All that said, I agree with others that it's a lot easier and cheaper to replace the scratchy pot. On the other hand, I don't recall ever buying stereo equipment solely because I needed it.

Rick "amused by the description in that Wiki article of radio transmitting tubes as 'monster' tubes" Denney

[SplatterTone]

The only tube amps I ever owned are long gone. They were a Scott 20 W/ch (with matching FM tuner) and a Heath 30 W/ch. I wish I still had the Scott. The Heath was a bit mundane. Another thing about tubes is they poop out (grids start to lose their shape) over time with ever increasing distortion and decreasing power. But they can be pretty to look at.

I found this picture of a pair of 3-500Z radio tubes. And yes, the plates really do glow orange when in use, and thats with a fairly hefty fan blowing on them. My piddly amp only uses one 3-500Z. I've never seen an example of anyone trying to use these in an audio amp. I have no idea what kind of fidelity or signal to noise one would get. I did run across a site with schematics of old Altec amps one of which used a pair of 813 tubes for 270 watts out. The 813 used to be a rather popular tube for homebrew radio amplifier builders.

[iiipopes]

I guessed I missed that initial comment about the 811A triode being an RF tube instead of an audio tube.

Yes, when a couple of friends and I worked at a small town local AM radio station for a couple of summers in high school, we didn't go near the power section, or at least not without the chief engineer being there with us.

I have forgotten how many thousands of dollars each he said the Westinghouse output tubes cost, and of course they had to be bought in matched pairs for efficiency.

Great pix of the orange glow. Yes, I can also vouch that RF power tubes really do glow. It has been over 30 years since I last thought of that.

The "fun" part was that they got an FM transmitter while I was there. Something wasn't installed correctly on it, and for awhile they had problems with a nitrogen tank hooked up to keep the power conduit to the tower purged as something about it wouldn't hold the slight pressure necessary to keep out water vapor. I left shortly thereafter to go to college and I never found out what the cause or solution to the problem was.

Stereo Review is still with us, under the name Sound & Vision:
http://www.roger-russell.com/magrevsr.htm

I did purchase my Yamaha 960II integrated amplifier after it was listed as amp of the year. At the time we had a great stereo store with everything from Panasonic to Macintosh and a good room layout for auditioning, so I could take my time to choose which one I wanted. The light bulbs have burned out of the switches, but it still soldiers on. The only thing about it is occasionally on some of the loudest dynamics there is some noticable compression (not distortion), probably from the power supply. I believe it has a different power supply as well, because Carver claimed this particular amplifier infringed on some of his patents and Yamaha had to stop making it.

The best sounding amplifier I ever heard was an old Dynakit that a friend of mine had in the early 80's, and it was old then. When we hooked a turntable up to it (pre-CD era, of course!) and played a few of the now classic albums, like Sgt. Peppers and Dark Side of the Moon, the stereo image had much greater depth and breadth than my Yamaha had, albeit with a tad more noticable distortion. You could pick out the exact placement of each alarm clock, for example. We used the same speakers, same placements, same turntable, etc., changing out only the integrated amp for the comparison.

[Rick Denney]

I found this picture of a pair of 3-500Z radio tubes. And yes, the plates really do glow orange when in use, and thats with a fairly hefty fan blowing on them. My piddly amp only uses one 3-500Z....

I'm as impressed by that beautifully wound plate choke as by the tubes.

The 3-500Z is a big tube, and and expensive sucker. That pair would cost about $350 to replace.

The red glow is really required for them, too. The plates are covered in zirconium which gathers the remaining air and water molecules in the tube during operation (called "gettering"). At the voltages involved, those molecules can cause arcing and that's why tubes have things in them that glow and gather them up.

I just fried a triplet of 811a's, trying to tune it up at about 28 MHz. The plates on those 811's glowed, too. I can tell, because they had melted spots where the gettering material was burned off. I replace them with 572b tubes that are quite similar but have graphite plates that won't melt at gettering temperatures.

Of course, by broadcast standards, these are not big tubes at all. Broadcast stations use tubes that cost thousands, and they have really specialized ways of taking care of them. The bigger ones, of course, are mounted in metal rather than glass envelopes. I don't think I even want to be in the same (large) room as the transformer it would take to drive plate voltage required for the 500KW border-blaster AM radio stations of yesteryear.

Rick "whose 600-watt amp is considered small, even for amateur service" Denney

[Rick Denney]

I guessed I missed that initial comment about the 811A triode being an RF tube instead of an audio tube.

In reality, though, the difference isn't that much, and 811a's were used in audio applications frequently.

A tube has a characteristic capacitance and that controls the range of frequencies over which it will operate efficiently. Of course, that isn't the same thing as the frequency response, which is a matter of how big an envelope of power the circuits around the tube are designed to provide. The 811a's will wiggle right fine in the audio spectrum. After all, audio frequencies extend to at least 20 kHz for hi-fidelity stuff, and that's a higher frequency than the AM broadcast band and right in the amateur 160-meter band.

The difference is that instead of sending those electromagnetic waves into an antenna, through a tuned output circuit, the audio amp has to convert them to mechanical energy in a loudspeaker. That takes current in those speaker coils. And current requires fat wire, and the right number of turns, which results in low impedance at audio frequencies. That's why you need that monster impedance matching transformer that I forgot about earlier.

But when it's still back at the high-Z electromagnetic stage, it's basically the same stuff.

In RF, the solid-state amps are much more expensive. But they have appeal because they are so broadbanded that you don't have to tune the signals going into and coming out of them to nearly the same extent. You can just switch bands and go. With tubes, I have raise the drive power in steps, adjusting the capacitors on the output tank to keep the tubes tuned to the exciter frequency. If I don't, the grid or the plates end up eating that power instead of sending it to the antenna in a process known as "tuning for maximum smoke".

Rick "still replacing parts resulting from tube death" Denney

[iiipopes]

Indeed. The urban legends I have heard include that the "X" pumped so much power that the electrical wiring in nearby bars or other buildings was enough of an inductance coil to dimly light up light bulbs even with the building's power off.