2. amps-preamps
  3. 1420484322

Best tube amp for electrostatic speakers

For over 35 years I've almost exclusively used either ribbons or electrostats with solid state amplification and have been generally happy with the sound. Over the last several years, though, my hearing has become somewhat degraded and more sensitive to certain frequencies. The frequencies which seem to cause the most discomfort occur in the 1400 to 1900Hz range and come across as "bright" to my hearing. After researching this matter and having been given some expert advice, I've decided to pursue the idea of replacing my present amplification equipment with tube based gear.

The purpose of this post, then, is to solicit advice regarding the best approach to making this decision based on the following information: the current basic equipment is Shanling Solid state CD player, Peachtree Audio Nova used as preamp, two DBX 1531 EQ's to help compensate for age related hearing loss, Peachtree 220 amp, Silversonic T114 cable and Martin Logan Ethos speakers.

The listening area is our living room measuring 15 by 22 feet with my listening position 16 feet from the plane of the 2 speakers which are positioned 11 feet apart measured center to center. Located between the 2 speakers is an entertainment center which is about 9 feet wide. My listening interests are varied from solo guitar and light jazz to occasional orchestral music. I don't generally listen at high volumes and am not particularly interested in strong bass except for the rare action movie background.

Unless not advisable for some reason, I would like to keep the Peachtree Nova as a preamp because of the significant latitude for source connection and what seems to have a decent internal DAC. If this option would substantially defeat the purpose of the intended modification I would work around it. I can no longer deal with sounds that are "bright" which I now find uncomfortable but detailed sound is very important.

So, the questions are: is the move to tubes the best option and, if so, what might be some reasonably priced amps that could accomplish the goal. This, of course, would take into consideration room size, etc. for determining power requirements. If there are other more practical and less expensive options to consider, I would appreciate that advice as well.
128x128broadstone

Showing 13 responses by atmasphere

atmasphere's avatar
atmasphere
11,566 posts
 
ESLs traditionally are a tricky problem for solid state amps. The reason is that solid state amps are often able to act like a voltage source, which is to say that they can often double power as the load impedance is cut in half.

This causes troubles with ESLs, as unlike cone-based speakers with a box or baffle, the impedance curve of the speaker is not an efficiency curve as well. IOW, the speaker has the same efficiency at all frequencies. Additionally, the typical ESL has an impedance curve that varies about 10:1 if it is a full-range system. For example, Sound Labs vary from a high of about 30 ohms in the bass down to about 1.5-3 ohms depending on the position of the Brilliance control. The Quad ESL57 varies about 45 ohms in the bass to about 4 ohms in the treble.

So the ability to double power as impedance is cut in half does not help so much- you get too much high frequency energy. This is why tube amps are usually the preferred amp if you run ESLs.

However Martin Logan has understood for a long time that solid state amps dominate the market, so they have kept their impedances low in order to limit the amp's ability to double power. As a result they are often only 0.5 ohms at 20KHz! To drive this with a tube amp a set of ZEROs is a good idea
http://www.zeroimpedance.com

The more feedback the tube amp has, its likely it will also sound brighter for two reasons. First, most applications of negative feedback contribute to brightness as it causes the amp to make higher ordered harmonics (while overall dropping the THD) which the ear/brain system uses to detect loudness (which is another way of saying that our ears are very sensitive to higher ordered harmonic content; much more so than human vocal frequencies). The second reason is that the more feedback is used, the more the tube amp will try to behave as a voltage source, which does not work so well with ESLs as previously explained.

More:
http://www.atma-sphere.com/Resources/Paradigms_in_Amplifier_Design.php

So- I would get the biggest tube amp you can that does not run negative feedback, and get a set of ZEROs so the amp can handle the otherwise difficult load. This will get you the speed and delicacy ESLs are known for, but without excess brightness.
atmasphere's avatar
atmasphere
11,566 posts
 
One has to be careful with the ZEROs. Like anything else they can be mis-used. Generally I go for the least amount of multiplication that gets the job done.

With ESLs this generally means you are trying to restore lost highs. What is nice about the ZERO is if you have an 8 ohm cone woofer you might just use the ZEROs on the ESL panel and let the woofer run direct. This gives you a little flexibility when doing the setup.

Also, be aware that they need about 45 minutes of warm-up before you take them seriously! This befuddles A/B experiments but it seems to be very real.
atmasphere's avatar
atmasphere
11,566 posts
 
Running tones like that is not a good way to tell what's going on- you are not the only one that would find tones like that unpleasant. In fact all the audiophiles I know don't listen to raw tones.

If you find a particular recording that seems unpleasant on your system, it might be worth it to play that recording on another system and see if you still think the same thing of it.
atmasphere's avatar
atmasphere
11,566 posts
 
Broadstone, If that same CD track is what irritates you on other systems, one strong possibility is the recording itself, unless you used the same CD player in those systems as well.

How about this- are there any recordings that sound fine on your system and are not irritating?
atmasphere's avatar
atmasphere
11,566 posts
 
As I have pointed out many times in the past, the impedance curve of an ESL is not the same as an efficiency curve.

With many box speakers the two are the same- higher impedances usually representing resonance, usually a driver in a box.

(If one were to apply this idea to an ESL, it is easy to see how one would think that you need lots of current to drive those low impedances. In a nutshell, it does not work that way with ESLs; they simply are not the same technology!)

ESLs are generally not in a box; their impedance curve is a result of a capacitive function unrelated to resonance.

The result is that the speaker generally needs a constant power characteristic out of the amp rather than a constant voltage characteristic in order to obtain flat frequency response. Constant voltage is that quality that allows an amplifier to double power as impedance is halved (and will generally result in brightness in ESLs). Constant power is that quality where power remains constant regardless of impedance.

There is more at this link:
http://www.atma-sphere.com/Resources/Paradigms_in_Amplifier_Design.php

While the Power Paradigm pretty well went the way decades ago in mid- and lo-fi situations, it is still very much alive and well in high end. Horns, ESLs, magnetic planars, single-driver full range speakers and various conventional box speakers represent examples of speakers made to conform to Power Paradigm rules.
atmasphere's avatar
atmasphere
11,566 posts
 
Uh, George, please read the thread through (beginning with the title of the thread) rather than just jumping in in the end.

First, this thread is about what tube amps can be used and is not about solid state.

As explained earlier:

Martin-Logans have a very low impedance at high frequencies as you know- we went through that early on. To use any tube amp with them, a set of ZEROs is recommended
http://www.zeroimpedance.com

This eliminates the rather ridiculously low impedance (0.5 ohm) high frequency problem.

Once this issue is dealt with your comments are rendered moot as the ZEROs level the playing field.
atmasphere's avatar
atmasphere
11,566 posts
 
The ZERO is not a band-aid. It is a problem solver, and very effective- usually removing coloration rather than adding it. There are many accolades on the zeroimpedance.com website and the product has been around now for nearly 20 years.

The issue we are dealing with here is that Martin-Logan recognized that solid state amps dominate the market, and they wanted to be part of that- because it sells more speakers. So they made their impedances extremely low.

The problem is that ESLs have the same efficiency at all frequencies. If you combine that with an amplifier that can double power as George suggests, the result is brightness. If the amp can otherwise drive the load but not double power, the tonality will be correct.

The ZERO allows this with many tube amps which otherwise would not be able to drive the half-ohm load which is what a Martin-Logan is at 20KHz. Sure, there is not much energy up there, but if you have too much it gets painful.
atmasphere's avatar
atmasphere
11,566 posts
 
George, Its obvious you've not actually heard the ZEROs work.

Too bright does not work any better than too dull- both are colorations.

The ZEROs are an excellent solution for tube amps that are being asked to drive speakers that they simply would not work with otherwise. I think more have been sold to ARC owners than our amps, FWIW. A lot of people want to use tubes with Martin-Logans to calm them down.
atmasphere's avatar
atmasphere
11,566 posts
 
George, you are simply not credible. I think you are simply trolling this thread and me as you typically do.
atmasphere's avatar
atmasphere
11,566 posts
 
Chuckie, do you have the new backpanels for your Sound Labs? They make the speaker easier to drive and better sounding with all amplifiers.
atmasphere's avatar
atmasphere
11,566 posts
 
Chuckie, There are some simple mods that can be done to the older backplates that make the speaker a lot easier to drive and sound better. PM me if interested.
atmasphere's avatar
atmasphere
11,566 posts
 
For some time now I've suspected that there is some sort of antagonistic relationship between timbre and brightness and, because the timbre issue is one of the most important issues to me and brightness is my least acceptable, I seem always to be walking a tightrope between the two.

This is a common problem! The thing to understand here is that the brightness caused by distortion does not contribute to timbre at all- instead simply makes the presentation more irritating, as the ear finds the higher ordered harmonics to be less pleasant.

To get timbre right, you have to have low distortion and pretty good bandwidth, and the distortion has to utterly lack any of the higher orders (5th harmonic and beyond).

In many amplifiers the feedback is poorly applied, the result being that those harmonics are injected into the resulting output. This has been known for a very long time as Norman Crowhurst was writing about this problem over 50 years ago. This is not to say that feedback in an amplifier won't work, but until the advent of simulation, its been almost impossible to do the design correctly using the traditional formulae.

So as a result many designers got it wrong- and the amps they made tend to sound bright. I find that many audiophiles tend to prefer sins of omission rather than commission...
atmasphere's avatar
atmasphere
11,566 posts
 
George, your comments fit the classic definition of trolling:
https://en.wikipedia.org/wiki/Troll_%28Internet%29

The fact is you have not used a set of ZEROs at your house. If you had, your comments would be in line with that of the experience of hundreds of people that use them. One can only conclude that you have never used them at all or are coincidentally the lone voice in the wilderness- Occam's Razor suggests otherwise.

Your use of imitation of my posts as a means of belittling is another example of trolling. The intent behind the text is the issue. As best I can make out, you just want to make me wrong and it gets tiresome, which also seems to be your intent.

If you don't want me to accuse you of trolling you will have to stop doing it.

So far you have yet to actually show the math that supports that the ZERO does not work or that ESLs have an efficiency curve that is the same as their impedance curve making them an ideal match with transistors. If you can do that then we have a discussion.
More to discover