Sonic qualities of SET output tubes?

@blindjim 

How well do they play stuff like Brian Setzer, Gordon Goodwin, Basie, big band blues, multi piece bands in general?

This begs the point then, exactly what musical genre is not served well by SET amps and High Eff speakers, if any?

Amps and loudspeakers don't have taste in musical genres! I wouldn't worry about that. Amps and speakers treat all music as electrical signals. There literally is no amp and no speaker that is particularly better at a specific musical genre. Seriously.

@blindjim 

I’m simply hard pressed to accept the fact a Triode amp even with nice horns will kick out the jams quite like amps I have become accustomed to hearing with far more horsepower and point source speakers of standard varieties, within the confines of my home.

I play our M-60s at home and they are triode amps. The speakers I use employ horns as well. As far as kicking out the jams: the speakers go down to 20 Hz and are 98 db and 16 ohms. I can literally shake the walls with bass that can be felt. That last bit of the bottom octave is one way you separate the men from the boys.

I played in orchestras from junior high all the way post-college. These days I play in a space rock band. That experience has made me a bit hard to please and I like to kick out the jams very much!

I think you will find if others respond, that horns and triodes are quite capable of bringing home the bacon! When I go to friend's houses and also when I go to shows, most systems come off to me as 'small and just little' because the images are tiny and the systems can't play bass, dynamic range or impact in the way that I have become accustomed.

@charles1dad FWIW, the Coincidents we had here at the shop were Total Eclipses- they had five drivers- two side firing woofers, two midranges and a single ribbon tweeter.

The 6C33 can sound excellent and despite bing indirectly heated can easily keep up with any DHT. It is as charles1dad says immediately above. The disadvantage with the 6C33 is that the tube eats tube sockets and there is nothing for it; the socket(s) simply have to be replaced after a time.

Life is so simple. Only money complicates it.

**That** certainly seems to be true.

The Coincident loudspeaker concept (tending to higher impedance) that Israel has had for the last 10-15 years or so was developed using an OTL, not an SET. Unless you get a fairly powerful SET, I think you will find that in most rooms the Coincidents really are not efficient enough to bring out the best of such amps, although they are otherwise easy to drive- I find most of them need about 60 watts as a good minimum in most rooms (we've showed with Coincident a bit in the past at CES and had them in our shop so this is based on direct experience). We had one customer (also a reviewer) that ran one of the smaller Coincidents with our S-30, and while it sounded fine, she found that she was wanting more power. 

We had a set of Coincident Total Victory loudspeakers here at our shop for several years; they sounded just fine but if I had to rate the efficiency I would say they were around 93 db.

We had a set of Audiokinesis loudspeakers here too; Duke of Audiokinesis rated them at 92 or 93 db 1 watt/1meter but they seemed more efficient than the Coincidents (Duke tends to be very conservative with his ratings; we've yet to see any of his speakers that didn't play well with our S-30).

We also had ZU Druids which were obviously more efficient (rated 101db which I think was a bit optimistic as well; 97db seems more accurate- otherwise also about 10 ohms) and also Classic Audio Loudspeakers which are 98db 1 watt/1meter (and 16 ohms).

So I have good reason to think the 93 db number I suggested above is pretty close.

In a nutshell, while Israel's amps sound just fine, on any of his speakers (unless you only play at lower volumes) I would think his amps won't make enough power although they do sound good together.

You get used to the increased performance that is available! Going back to ’lesser goods’ can be a bit depressing. But I find that if you are careful, good sound can be had on almost any budget.

Number of tubes would not be a problem for me unless each cost $1k and lasted only for a year. It's a bit more difficult to do amp first approach than speakers first approach. But with Atma-Sphere and LAMM I would still do exactly that.

We warrant the tubes in our gear for a year and they are relatively inexpensive.

The only real worry is as said, clipping. Its when a driver is starved for power predominately. Usually it’s a tweeter. Ordinarily clipping occurs when running the spkrs at higher levels and the power a driver asks for can’t be provided appropriately or comes intermittedly. That’s when actual . damage to the loudspeaker can take place. It happened in a pr I had bought preowned. His INT did not have the power to satisfy those speaker’s demands and one of the four blew out. Quit working. The factory fixed it eventually. Once I paid for shipping and so forth.

The mechanism for this is when the amplifier clips, the distortion generated is high enough in frequency that the crossover allows it through to the tweeter(s), which usually only handle maybe 2 watts or so. With solid state amps in particular, overloading the amp is a good way to toast the tweeters.

This is considerably more rare with tube amps because they generate less higher ordered harmonics when overloaded (which is usually caused by bass notes) but still can happen.

If you see tweeters blown and nothing else, its a sure sign that the speaker was exposed to an amplifier that was clipping. Most manufacturers would not consider that as a warranty claim as it is clearly abuse of the speaker.

Ralph, besides Classic Audio, what would you recommend to use with your amps? Also, do you believe that short length Mogami RCA cables are just as good as XLR? I mean 2497 wire.

If you keep the cables short you can do pretty well with single-ended connections. I can't comment about the single-ended Mogami- not tried it.

Regarding speakers, Audiokinesis, older Quads (newer ones need the ZEROs), Sound Lab (with our bigger amps), Coincident, Devore, Lowther, PHY (will need a speaker protection fuse as they only handle 20 watts), most horn systems and single-driver systems, AvantGarde, Merlin, Magnaplanars (use ZEROs, Eminent Technology (may need to use ZEROs) Vandersteen, Certain Avalons such as the Eidelon, JM Labs, Rogers LS35a, certain Sonus Fabors, Pipedreams, and many more.

The trick is similar to what you look for with an SET or any tube amp using no loop feedback, which is to say that the woofers should be the same impedance generally speaking as the midrange. Dual woofer systems that are 4 ohms in the bass and 8 ohms otherwise don't work so well. The difference between our amps and SETs is that we can work with speakers of much less efficiency, but personally I do prefer efficient speakers as I like to crank the hell out of the system once in a while and don't see the point of struggling to do that.

I guess in the golden days of audio weren't most speakers 16 ohm to be more tube friendly in that respect.

Of course!

1 4 ohm load impedance speakers are easier to design/build.
2  An assumption that the majority of buyers plan to use solid state amplification.
3  The designers/builders simply believe that the speakers sound better with the lower 4 ohm impedance.

My SET and 2 push pull amplifiers all mate beautifully with my 14  ohm speakers. I do believe Ralph's assertion that higher speaker impedance is beneficial to all amplifier topologies.  It seems many speaker builders would disagree with him given the predominance of 4 ohm speakers in the marketplace.

The vast majority of amplifiers are solid state. These days most of them are safe with 4 ohms, and further many of them can double power into 4 ohms. In this regard, many speaker manufacturers are seeking to make (under a false impression) to make their speakers more compatible with solid state and thus secure a greater market. Its a good example of dollars being the goal rather than the ultimate in sound reproduction.

If they were to simply raise the speaker's impedance and keep all other things the same (which in many cases would be a bit of a trick) they would find that their speaker is sounding smoother and more transparent, regardless of the amp they use.

There really isn't any point in making a speaker that is 'difficult to drive' as the result will be more distortion, of a type that will cause the amp to be harsher and less detailed. You can see it in the specs.

What am I missing?

Grannyring got it in one- well done!

Apparently that this applies specifically to SETs and not to other kinds of amps. For example, despite being zero feedback (like SETs), our amps make a fraction of the distortion at full output (about 0.5% if set up correctly) so you have a much greater *percentage of usable output power*.

I like to use that idea of ’percentage of usable output power’ as its a number that can vary widely from amp to amp. Generally speaking, tubes have greater usable output power than solid state amps seem to have but SETs are the exception.

This is all about distortion, which all amps make so the ’percentage of usable output power’ has to do with the **kinds** of distortion that are particularly objectionable to the human ear. These are the higher ordered harmonics (5th and above, particularly the 7th), IM distortion, and in the case of class D, inharmonic distortion (caused by intermodulations with the scanning frequency; similar to aliasing in digital audio).

Because the ear converts all forms of distortion into tonality (including aliasing), the distortions above all manifest as some form of brightness and hardness. By contrast the lower ordered harmonics are perceived as ’warmth’ and ’bloom’.

Its quite worthy of note that these objectionable distortions always occur in very small amounts as opposed to the lower ordered harmonics- the issue is that the ear is far less sensitive to the lower orders. In this regard I propose a ’weighting’ scheme so that trace amounts of higher ordered harmonics can be seen for what they really are- which is to say: audible!

In SETs, the first onset of higher ordered harmonics is heard as ’dynamics’ by the ear as musical energy often occurs on transients where greater power is required (in a way fundamentally different from solid state, which tends to make the the higher ordered harmonics all the time). What happens is the higher ordered harmonics (to which very importantly the ear uses to calculate sound pressure) are thus only showing up initially on the leading edges of transients and no where else, so then it is perceived as ’dynamics’. The problem is, once you are aware of this fact its easier to hear that what is really happening is the amp is making distortion. So the simple act of reading this paragraph may have ruined it for some people. To avoid this you simply need more power or more efficiency. More power in an SET means it won’t sound as good and therein lies the dilemma!

Of course all forms of distortion are to be avoided, but avoiding the the objectionable forms is far more important overall for a pleasant listening experience. This, in a nutshell, is why tubes are still around nearly 60 years after being declared ’obsolete’.

One thing that has really become evident in the last 15 years or so is that if the loudspeaker is a higher impedance that avoiding the more objectionable distortions is easier to do **regardless** of the amplifier technology. I think speaker manufacturers are slowly figuring this out as there are more 16 ohms speakers now than there were 15 years ago. Put another way, if you are tying to make a lower powered tube amp sound like music, a four ohm load is absolute anathema!

Its my opinion and also my experience that four ohms has no place in high end audio (IOW if high quality audio production is in fact the goal). The cables are more critical and all amps sound harsher and less detailed on four ohms as opposed to eight or sixteen all other things being equal. Its not only audible, you can see it in the specs of any amplifier made.

there is usually a lot of distortion in the recording already. How do you try to correct that? Or you simply live with it and do your best not to make it worse?

There may or may not depending entirely on the recording.

You certainly don't want to **add** to it! So yes, you do your best not to make it worse.

@blindjim , all the comments regarding speaker efficiency are spot on.

And you are quite correct that different brands of power tubes are going to sound different. Once you settle on a particular power tube type, then you sort out which brand of that tube to get. The choice of power tube type is heavily dependent on which speaker you get!!

The one thing you want to pay attention to is the way SETs make distortion. In that regard, they are very very good at low power; distortion falls to the point of being unmeasurable (which is the source of the ’amazing inner detail’ for which so many SETs are known; the amps we make are some of the very few push-pull amps that share this low power lack of distortion character). But at full power, the typical distortion is in the neighborhood of 10%.

SETs are known for being quite lush and very dynamic. Both of these properties are due to distortion- the 2nd harmonic brings the lushness, and at powers over about 20% of full power, the upper harmonics cause the dynamic properties on account of the way the human ear perceives sound pressure.

You really want to avoid the higher ordered harmonics! Not only do they cause the amp to sound dynamic, but they will also cause the system to sound loud and shouty. You will see a lot of people talk about the fact that their system plays as loud as they want; this is **often** caused by the distortion and not the actual sound pressure. A good system should **never** sound loud!

To avoid this, the speaker simply has to have the efficiency such that you don’t exceed about 20% of full power of the amp you have in mind. This will have the amp playing its most transparent and with its smoothest qualities. Generally speaking, this means that in most cases, a speaker of less than 100db is not going to be suitable. I am full aware that there will be many that read this and will dispute these words; my recommendation is to reread this post! For these people, if your system is not set up by this rule of thumb, you really aren’t hearing the finer qualities of your amp, no matter how enjoyable it might be.

@inna 

Ralph, would you like to tell us about disadvantages of OTL amp design? 0.5% distortion at full power is quite an achievement.
One thing is certain - more powerful SET amps done right are very expensive.

We get that figure without using feedback. With feedback the THD can be as low as 0.05% but we don't like feedback for the same reason that SET designers don't: it adds brightness.

OTLs have several disadvantages- speaker choice is an issue, as the smaller OTLs can't do 4 ohms very well. Generally speaking, if the speaker works with a low power SET it will work with a low power OTL too- the difference being that the 'low power' OTL will likely have several times the power of the low power SET.

The next issue is the number of power tubes, for example we use the 6AS7G which is good for about 10 watts. So you need a few of them if you want to make any power. Some OTLs use the EL509 or EL519, which are pentodes but the most common alternate is the 6C33. Four 6C33s will make about 60 watt in an OTL.

Finally, the smaller you make an OTL, the less efficient it becomes. So while our M-60 is perfectly comfortable on 8 ohm speakers, our S-30 is a lot more finicky about that- it can do well on some 8 ohm speakers but if there is a wild phase angle or the amp has to work hard it may not like it. Conversely, the bigger the OTL the more efficient it becomes so even though the class of operation has not changed, the bigger amps tend to have lower temperatures on each tube because more of the power is dissipated in the speaker rather than in the output section. 

The difference between tubes themselves is not that profound. They are just tubes and they have very similar linearity and bandwidth.

This is not true of output transformers!  Here's the problem- the more power there is, the harder it is to get bandwidth out of the transformer (and in this regard, much moreso than is encountered with push-pull; we sidestepped the whole thing by eliminating the transformer altogether).

So as the output power goes down, the sound of the amp improves. 25 years ago the 211 was king, 20 years ago the 300b was king, by sometime in the early 2000s the 2A3 was king, now we're seeing the type 45, 10 and 71 vying for the title. A traditional 45 is good for about 0.75 watts.

Now as transformer designs have improved (some taking advantage of technology not available in the 1920s and 1930s) its been possible to extend bandwidth in higher power designs. So the result is you're going to get a lot of conflicting anecdotal comments.

The other issue I've not mentioned is how well the tube and the output transformer work together. Some tubes want a 5K load and some output transformers present 4.9K or 5.1 K instead of exactly 5K (which will vary from tube to tube...) so you're going to see variance on that account as well.

So don't expect consistency in the comments to follow.