Sanders paper on Tubes vs Transistors for ESLs

[Garth]

Sanders Good Read

Posted by Jonfo over on the Logan forum

http://www.sanderssoundsystems.com/technical-white-papers/172-tubes-vs-transistors

Nice read

A few years ago I was at RMAF I arrived a bit early my room was not ready the sign in was in less than ready mode. So I walked the halls most rooms had locked doors as they were still setting up. A few were open with music playing and a come in and listen mode one was Sanders room. Sanders stated much as he has in this paper you need a lot of power to get good sound . I walked a few doors down another room was open I do not remember the brand names but they had a big horn speaker and a 4 watt amp and the guy said their is no way to get good sound if you go over 4 watts. I hope he and Sanders do not meet in the bar.

I think Sanders got a lot more right than he got wrong in what he wrote as he was writing about his speakers. I do find it odd that anyone building a big solid state amp good or bad is fair game. People feels free to say solid state is harsh , they have no music in them , cost too much but if a guy builds a 4 watt amp very little is said against them. You never see a reviewer say about a 4 watt amp the only way you are getting Bass out of this amp is to drop it on a Kettle Drum. I have heard great sounding big amps and if you are willing to limit the bass range great sounding set amps.

But why are small amps and small speakers given such a pass on sound? A large amp or Speaker is a lot more likely to get tarred and feathered. If a amp or speaker is both big and expensive well their is going to be a hanging for sure.

I wonder if the Absolute Sound the idea not the Mag is a myth and most people are looking for a sound they like. If they find it and the guy selling it has a bigger price on it than they can work with. Well then blame is laid be it over priced sound or what ever makes you feel good. Just IMO any thoughts

 

[MylesBAstor]

A few years ago I was at RMAF I arrived a bit early my room was not ready the sign in was in less than ready mode. So I walked the halls most rooms had locked doors as they were still setting up. A few were open with music playing and a come in and listen mode one was Sanders room. Sanders stated much as he has in this paper you need a lot of power to get good sound . I walked a few doors down another room was open I do not remember the brand names but they had a big horn speaker and a 4 watt amp and the guy said their is no way to get good sound if you go over 4 watts. I hope he and Sanders do not meet in the bar.

I think Sanders got a lot more right than he got wrong in what he wrote as he was writing about his speakers. I do find it odd that anyone building a big solid state amp good or bad is fair game. People feels free to say solid state is harsh , they have no music in them , cost too much but if a guy builds a 4 watt amp very little is said against them. [/B] You never see a reviewer say about a 4 watt amp the only way you are getting Bass out of this amp is to drop it on a Kettle Drum. I have heard great sounding big amps and if you are willing to limit the bass range great sounding set amps.

But why are small amps and small speakers given such a pass on sound? A large amp or Speaker is a lot more likely to get tarred and feathered. If a amp or speaker is both big and expensive well their is going to be a hanging for sure.

I wonder if the Absolute Sound the idea not the Mag is a myth and most people are looking for a sound they like. If they find it and the guy selling it has a bigger price on it than they can work with. Well then blame is laid be it over priced sound or what ever makes you feel good. Just IMO any thoughts

I think the only people who would say that in this day and age about the best solid-state amplifiers out there today only sees the world in black and white. Tubes and solid-state do different things but they do overlap in a few areas nowadays. And that comes from a tube owner.

 

[Garth]

I think the only people who would say that in this day and age about the best solid-state amplifiers out there today only sees the world in black and white. Tubes and solid-state do different things but they do overlap in a few areas nowadays. And that comes from a tube owner.

Agreed but do you think small amps and speakers get more of a pass than larger gear . I have read reviews on small speakers where the word BASS was never even hinted at.

 

[bonzo75]

A few years ago I was at RMAF I arrived a bit early my room was not ready the sign in was in less than ready mode. So I walked the halls most rooms had locked doors as they were still setting up. A few were open with music playing and a come in and listen mode one was Sanders room. Sanders stated much as he has in this paper you need a lot of power to get good sound . I walked a few doors down another room was open I do not remember the brand names but they had a big horn speaker and a 4 watt amp and the guy said their is no way to get good sound if you go over 4 watts. I hope he and Sanders do not meet in the bar.

Sanders does in the above paper mention that his discussion makes an exception to horn-loaded speakers. Horns and small amps are different, such systems are so fast and transient those qualities are unmatched. Yes they have bass integration issues then. Depends what you want to listen to. Panels are more inefficient, but because of the nature of the panel (moving a large volume over a small area rather than a small volume over a large area like in cones) they are fast as well, though nowhere like horns with small amps. I find cones sssssslow and bloated but that's me. I wish I was as open minded as those who can appreciate all 3 - panels, horns, and cones

 

[Atmasphere]

Interesting read, nothing new really; this caught my attention:

The worst tube amp I measured was an Atma-sphere OTL amp. It was rated at 30 watts. It only produced 14 watts into an 8 ohms load and had 13% distortion. It was a new amp and both channels measured virtually identically. So I don't think it was defective.

The following is of course also my experience - I just prefer amps with parallel thermal monitor circuits that simply cut off all power until powered back on:

Sure sounds like something was up with that S-30 to me! Its telling that this is the first I have heard of this. The S-30 makes 30 watts into 8 ohms, not 14 watts... I know Sanders didn't get the amp from us so he must have obtained it from a customer. If I were such a customer with a new amp I would have called the manufacturer to see what was up with the measurements but we have never received any such call.

If Sanders did make that measurement, it is likely that he made a simple mistake that several have made before him- if you don't take special steps to make sure that your input is ground isolated and that the test equipment at the output is also ground-isolated, you can wind up shorting one speaker terminal to ground through the test equipment, which results in uneven drive to the output section. This results in higher distortion and lower output power. There is a review in a French magazine where they made precisely this mistake.

But right now I don't think this actually happened- we would have heard about it.

I noticed also that his numbers were off on the output impedance of an OTL. I found this comment:

Note that OTL tube amps don't solve this problem. They have no transformer, so must relay on putting many output tubes in parallel to lower the impedance. This quickly results in having an absurd number of tubes with all their heat and power requirements. So OTL amps do not get down to very low impedances.

Most get to just around 10 ohms and the best only get a bit lower. As a result, they have severe impedance mismatch issues and are really quite a poor choice for driving ESLs. They also measure really poorly on a spectrum analyzer compared to transformer-coupled tube amps.

The first paragraph has a conclusion not supported by the rest of the paragraph. It is also incorrect. It is apparent that Sanders operates on the Voltage Paradigm (http://www.atma-sphere.com/Resources/Paradigms_in_Amplifier_Design.php); when using the Voltage Rules and if feedback is applied to our amps we can get the output impedance well below one ohm. Thing is, we don't use negative feedback and we are one of the very few OTL manufacturers where that is the case- most use 20 db or more and can get voltage source response out of their amps, as such behavior is often a function of feedback rather than the actual impedance of the output section. The comment amounts to a logical fallacy known as a Hasty Generalization (all logical fallacies are inherently false BTW...). It would appear that we were targeted for this article.

The comment in bold is simply false. OTLs will perform as well or better than a transformer-coupled amplifier. This is because they have wider bandwidth and typically lower distortion; it should not come as any surprise.

As I have stated elsewhere ad nauseum (I have Myles to thank here ) a lot of the difference we hear between tubes and transistors does not have anything to do with how they clip (although at clipping the differences between the two technologies are magnified) and this is easy to prove- just compare both amps on an efficient loudspeaker. The differences we hear have to do with how the two technologies make distortion when they are **not** clipping. The human ear is very sensitive to the odd ordered harmonics that transistors make in trace amounts; this results in the ear perceiving them as brighter and harsher than tubes, even if not by very much.

Now I have stated in the past that negative feedback is responsible for some of the odd-ordered harmonic distortion and that is true, but it is not the whole picture. Semiconductors often have a non-linear capacitive element that is inherent in their junctions (a certain diode known as a variactor takes advantage of this capacitance to tune most modern FM radios) and is magnified by current. This characteristic also contributes to odd ordered harmonic distortion.

If you rely solely on test equipment to suss things out you will think you are in the clear. This is because usually the ear is insensitive to a lot of stuff that shows up easily on test equipment. The thing that is poorly understood is how the ear/brain system perceives odd ordered harmonics; in a nutshell it uses them to calculate how loud a sound is. Additionally, the human ear/brain system is tuned to be the most sensitive at birdsong and insect frequencies (Fletcher-Munson); combine the two and you have an organic system that can detect odd ordered harmonic distortion that is difficult to see on test equipment. In addition, the ear translates distortion into tonality- thus not only does amplification with trace amounts of odd ordered harmonic distortion sound louder than it should, it will also sound brighter even though no variation in frequency response is measurable!

At any rate we have a lot of customers with ESLs; about 80% of our MA-2 production has been for Sound Labs. The MA-1 and M-60 are often used on ESL 63s. Sander's ESLs don't show up in our customer list- likely because he has kept his impedances lower to the point that driving them with tubes is impractical- not due to efficiency, but due to impedance.

 

[RogerD]

I would have to think in this day and age it would be pretty hard to build a brittle and brite sounding solid state amplifier. Maybe right out of the box you could hear something unpleasent,but with todays transistor devices, resistors and electrolytics I would bet the majority sound good and I actually prefer a SS amp to a tube amp. Most high powered SS amps usually have large power supply capacitors.I know mine have 200Kuf and I think that makes a difference and that's a mono block so for 2 channel I'm talking 400,000uf and depending on the source material like opera I have never heard them clip and there's something to be said for effortless sound.YMMV

 

[Atmasphere]

^^ I'm a big fan of not clipping the amp!!

The flip side of it though is that most push pull amps (tube or solid state) have a distortion character such that as power is decreased, the distortion decreases to a point, then at some nominal power level, usually about 5-7% of full power, the distortion starts going back up. To avoid that (if you want good low level detail, since distortion can mask detail via the ear's masking rule) the speaker has to be matched with the amp such that it stays out of that power level yet does not clip.

Plan B: use an amplifier where the distortion linearly decreases to unmeasurable as power is decreased to zero.

 

[Mosin]

I would have to think in this day and age it would be pretty hard to build a brittle and brite sounding solid state amplifier...<snip>

I would like to think that, too. Unfortunately, a lot of them sound bright and edgy to my ears.

 

[DonH50]

Not sure I have seen distortion go up at very low power. Noise rises relative to output at low levels, however, and is why you see the rising low-power response in most "distortion" plots.

 

[Atmasphere]

^^ Take a look- there are plenty of examples, although this is not universal by any means.

Here is a typical example (and not picking on Parasound, IMO a very competently designed amplifier lineup):

http://www.stereophile.com/content/parasound-halo-jc-1-monoblock-power-amplifier-measurements

 

[DonH50]

Not sure that is valid (not sure it is not, not enough info). The problem is that most "distortion" analyzers actually measure SINAD, not just THD/IMD/whateverD. They do an FFT and anything not the fundamental is considered distortion. Even if they just look at e.g. the harmonic spur bins, rising noise in the bins is treated as rising distortion, neglecting the fact that all the other bins are also rising. That forces the reduction in SNR at low levels to be interpreted as rising distortion when it is really just rising noise. Same thing happens if there is clock feedthrough, power supply ripple, etc. Wish they'd show an FFT at low levels, then it should be pretty clear. Distortion rising with reduced signal is counterintuitive for most amplifier topologies. That said, I am sure you have lots more practical experience measuring audio amplifiers than I do!