Constant power - tubes vs. solid state amps and wild-impedance speakers

ack · Nov 11, 2010

If this was posted before, please go ahead and kill the thread... But here's a thought-provoking blog post by Atma-Sphere on why OTL tubes sound better with wild-impedance speakers (e.g. electrostatics) than solid state, and what solid state has to do to compensate:

Let's say you have a high quality 150/channel transistor amp. 150 watts into 8 ohms, a reasonable amount of power, but if you have a four Ohm speaker its 300 watts. Nice. Into 2 Ohms, if the amp doesn't blow up or current limit, 600 watts. So what does the amp produce driving 16 Ohms? 75 watts. Into 32 Ohms its only 35 watts! This could result in serious problems were the speaker a typical electrostatic, where such impedances are common in the bass frequencies. This explains why transistor amplifiers are usually such a poor match for electrostatic speakers.

This is what the right OTL can do into these impedances: 150 watts into 8 ohms, 145 into four (less than 1/2db difference), about 80 watts into 2 ohms, but into 16 we have 149 watts, into 32 ohms 145 watts- so you see that as long as the speaker load is moderately well behaved, this OTL example produces far more linear power over the same range of impedances, whereas the transistor amp is quite simply incapable of being linear at all! Why?

When a recording is made, it is assumed that a linear system is to be used so that it is capable of recording the same energy at all frequencies. When we play it back, for best results the playback should be the same at all frequencies, too. If there are variations in the speaker impedance, this will not be possible with a transistor amplifier unless it has a lot of negative feedback (which most of them do), which has the additional effect of decreasing bass impact, restricting dynamics, foreshortening soundstage depth and increasing odd-ordered harmonic distortion. Thus there is no way that a transistor amp can be described as linear if it is subject to these problems and that is one of the reasons why transistor amps produce so many amusical colorations. The reason has to do with the vanishingly small output impedance of the transistor amp (here's some myth bashing for you). The result is that the transistor amp has what is called a constant voltage characteristic, not constant power, which is of course what a power amp should do!

So, despite the fact that smaller OTL amplifiers don't like four ohm speakers, they are quite capable of giving you a more even power characteristic (read: flatter frequency response, all other things being equal), especially on a speaker with a wild impedance curve.

Would love to see reaction to this... For example, exactly how does negative feedback relate to constant power, i.e. how does it make an amplifier not sensitive to wild impedance fluctuations?

On a related note, I had posted on a similar thread that Soulution presumably discovered that a lot of negative feedback is not a problem as long as you can build ultra-fast circuits to handle it, because they claim that the smearing of negative feedback is otherwise very evident. So is this perhaps why such ultra-fast designs (Soulution, Spectral, FM Acoustics) presumably sound so good?

microstrip · Nov 11, 2010

Ralph Karsteen designs excellent sounding amplifiers - I have owned MA50's and MA2's, that coupled with adequate speakers can result in excellent sounding systems.

The comment on the OTL should be regarded in the proper context - used with very low efficiency electrostatic speakers having high impedance. In these conditions, because the OTLs have a power supply with higher voltage than typical transistors amplifiers (+/- 135V for Atmasphere), they can deliver higher power than the transistor amplifier, for the same quoted power at the standard 8 ohms. Not a miracle, just physics.

The second part of the message is just a little more of the religious discussion on feedback properties that will never end. All the techniques have proved that they can produce excellent results - so good that same people claim they are not distinguishable.

BTW, Ralph is one of the few high-end designers who states that between his preamplifier and amplifier (he uses the professional 600 ohm standards) no fancy cables are needed, for optimum performance use just a good quality audio balanced cable. And yes, Steve, after 20 years fighting it, has now implemented a mechanical remote system that does not interfere with the sound quality (very expensive, OK)!

amirm · Nov 11, 2010

We can put his ideas to test. What speakers do we want to look at? Here is the Quad ESL

microstrip · Nov 11, 2010

amirm said:
We can put his ideas to test. What speakers do we want to look at? Here is the Quad ESL

Unhappily the Quad is a poor example - it has a protective system that will trigger the protection much before the limits of typical amplifiers to be used with it.

But my Soundlabs have a similar impedance shape with more than 50 ohms in the bass and are less sensitive - sometimes they need "equivalent peak powers" of 300W at 8 ohms - e.g. ~70V peak.
The Atmasphere MA1 OTL is quoted as having only 120W at 8 ohms, but as music usually needs much higher power at the bottom octaves and very little current is needed at these frequencies it will sound louder without distortion than a comparable power transistor amplifier of 120W that will clip at around 43 V peak.

BTW, how do we post a picture from the hard disk in this forum?

amirm · Nov 11, 2010

microstrip said:
BTW, how do we post a picture from the hard disk in this forum?

See this: http://www.whatsbestforum.com/showt...FULL-SIZE-images&p=13490&viewfull=1#post13490

amirm · Nov 11, 2010

Here is then what we know:

1. Most power is required at low frequencies. In the example I had, below 80 Hz, the impedance is 16 ohms or lower. So in this situation, the SS amp wins.

2. Highest frequencies require the least power. So even though the impedance of that example rose up 16 ohms at 20 Khz, the power advantage is not as material there (assume tube wins at 16 ohms).

3. The peak is at 90 to 100 Hz. So in mid base tubes would do better.

4. In the critical mid-bad frequencies of 2 to 3 Khz where the ear is the most sensitive, impedance is actually quite low at 6 ohms. So again, looks like SS does better which is strange because most people like tubes for their midrange performance.

So overall, I like the point he is making but having trouble realizing it with the one graph I found.

microstrip · Nov 12, 2010

amirm said:
Here is then what we know:
1. Most power is required at low frequencies. In the example I had, below 80 Hz, the impedance is 16 ohms or lower. So in this situation, the SS amp wins.

The graph from Stereophile forgets that the impedance of the ESL63 in the low frequencies depends strongly on the amplitude of the test signal, something that the 1981 test in HFNRR showed clearly, and for signals over 1V will be much higher than the shown graph.

The Soundlab impedance is greater than 30 ohm from 20 to 80 Hz.
I will look for the graphs and try to post them soon.

ack · Nov 12, 2010

amirm said:
Here is then what we know:

1. Most power is required at low frequencies. In the example I had, below 80 Hz, the impedance is 16 ohms or lower. So in this situation, the SS amp wins.

During the Magico Q5 demo at Goodwins that I attended, Wolf said that the soprano's voice can be extremely demanding of power, and that during demos of the Q5's in Europe driven by the large DartZeel monos he would witness the amps' meters report peaks of ~1600W (per channel) when playing back content with sopranos. Moreover, recording studios are very careful when recording sopranos as well, as they can throw their peaks way off the charts - this is why in many recordings sopranos seem to sound distant, or at least not as powerful as they do in concert. But never mind sopranos - a chorus of 100 will easily overpower anything including winds - and I haven't heard a single such recording yet that does justice to a real chorus.

At any rate, Martin Logans are notorious for dropping their impedance below 1 ohm at 20kHz, so perhaps we can look at those too?

RUR · Nov 12, 2010

ack said:
At any rate, Martin Logans are notorious for dropping their impedance below 1 ohm at 20kHz, so perhaps we can look at those too?

e.g. here's the SL3:

http://www.stereophile.com/floorloudspeakers/145/index10.html

I know of no source for this data on the newer models, but it should be similar.

microstrip · Nov 12, 2010

ack said:
At any rate, Martin Logans are notorious for dropping their impedance below 1 ohm at 20kHz, so perhaps we can look at those too?

Only if you want to show a non-synergistic match

.

Martin Logans are known to be incompatible with OTLs, and in my opinion with typical tube amplifiers - but I have no experience with recent models with active woofers.

Both the old Prodigy's and the Monoliths sounded best with high power SS.

But other people can have different opinions!

amirm · Nov 12, 2010

RUR said:
e.g. here's the SL3:

http://www.stereophile.com/floorloudspeakers/145/index10.html

I know of no source for this data on the newer models, but it should be similar.

OK, this one makes a stronger case for him, assume the SS amp didn't have gobs of power to start because at max of 16 ohms, its power only drops by a factor of 2.

Edit: I said the above too strongly

. In every case that the response dips below 8 ohms, the tables get turned. So how would anyone be able to make that case?

amirm · Nov 12, 2010

ack said:
During the Magico Q5 demo at Goodwins that I attended, Wolf said that the soprano's voice can be extremely demanding of power, and that during demos of the Q5's in Europe driven by the large DartZeel monos he would witness the amps' meters report peaks of ~1600W (per channel) when playing back content with sopranos.

I am not into opera so you need to educate me on what that really means

. A quick Internet search shows this link: http://www.listenhear.co.uk/general_acoustics.htm

Which says the range is from 250 Hz to 1 Khz (assuming without harmonics). If so, it makes sense why it needs so much power as I noted. Was the implication above that it consists of very high frequencies? If so, I would be shocked to hear that someone needs 1,600 watts to drive a tweeter.

ack · Nov 12, 2010

Notice, this is midrange territory. Still shocking, perhaps; but most likely true.

FrantzM · Nov 12, 2010

I do have my doubts ... Assuming a passage with the full Orchestra going? maybe .. Soprano solo? 1600 Watts ...

Joao@altheamusica · Nov 12, 2010

microstrip said:
The comment on the OTL should be regarded in the proper context - used with very low efficiency electrostatic speakers having high impedance.

What frequency do you mean for high impedance? An ELS does act like a capacitor and for this reason the impedance decreases with increasing frequency!

When we look at this the ELS is just impedance friendly at low frequency and the matching transformer must correct this - in short you need to equalize the impedance from the ELS for the amplifiers.

Another way is like Beveridge, and others, does it light years ago to use a direct coupled high voltage amplifier to drive the ELS elements.

microstrip · Nov 12, 2010

ack said:
During the Magico Q5 demo at Goodwins that I attended, Wolf said that the soprano's voice can be extremely demanding of power, and that during demos of the Q5's in Europe driven by the large DartZeel monos he would witness the amps' meters report peaks of ~1600W (per channel) when playing back content with sopranos.

I also was a witness at these demos in Audioshow in Estoril, close to Lisbon. It is true, but the Q5s were playing in a very large room - surely over 100 square meters at insane show high levels. No soprano could reach this loudness. The Q5s really have a low impedance, otherwise the Dartzeels could not heve delivered these power levels (the amplifier has a real power meter).

But it is also true that sometimes sopranos can explode a system - a friend of mine who owned Quad ESL63 discovered that Mirella Freni in Puccini "Madama Butterfly" area "Un bel di vedremo" could trigger the crowbar protection of these speakers, shorting his amplifier output with the characteristic smell of electronics smoke... BTW, the peak power frequency of her voice is this demanding track is at 375 Hz, not in the treble zone as suggested sometimes.

If you want to analyze the power of music versus frequency download the excellent demo full version of Spectraplus at http://www.telebyte.com/pioneer/ - its full version can be used freely for 30 days - and analyze a few of your favorite tracks. After using it I quickly understood why very few systems can play a Shostakovitch finale properly - the high frequency power of a full symphonic orchestra playing fortissimo is unusually high.

c1ferrari · Nov 13, 2010

Amir,

Thanks for the plot/graph and interpretation

c1ferrari · Nov 13, 2010

microstrip said:
The graph from Stereophile forgets that the impedance of the ESL63 in the low frequencies depends strongly on the amplitude of the test signal, something that the 1981 test in HFNRR showed clearly, and for signals over 1V will be much higher than the shown graph.

The Soundlab impedance is greater than 30 ohm from 20 to 80 Hz.
I will look for the graphs and try to post them soon.

Micro,

Look forward to viewing this and reading your remarks. Thanks!

microstrip · Nov 15, 2010

Sorry, only today I managed to post the promised ESL 63 impedance graph.
As you see, it is level dependent and will raise with power - 160 mV at 8 ohm is only 3mW .
Low bass impedance is low only at powers that are not significant.

.

Gregadd · Nov 15, 2010

Remember there are different types of OTL's. Atma-Sphere is an all out design. Class A triodes. Bound to be finicky.

Constant power - tubes vs. solid state amps and wild-impedance speakers

VIP/Donor & WBF Founding Member

VIP/Donor

Banned

VIP/Donor

Banned

Banned

VIP/Donor

VIP/Donor & WBF Founding Member

WBF Founding Member

VIP/Donor

Banned

Banned

VIP/Donor & WBF Founding Member

Member Sponsor & WBF Founding Member

Well-Known Member

VIP/Donor

Member Sponsor & WBF Founding Member

Member Sponsor & WBF Founding Member

VIP/Donor

WBF Founding Member

Similar threads