Sauermann Amplifier

[manisandher]

Hey, great to 'see' you here Gerd!

I know you're busy, but if you have a few spare minutes could you share your thoughts as to why my mono amps sound so much 'fuller and extended' in the bass (and yet utterly taut, fast and tuneful) than even my 900W amp?

I'll be posting more of my purely subjective thoughts on the sound of the amplifier this evening...

Mani.

[CoenP]

Hello Gerd,

Thanks for taking the time to clarify on your design!

Hi Coen, You are pretty close but not totally right.
Since there is constant current flowing through the voltage gain stage this stage has an output impedance of some kilo ohms. This stage can not deliver any power or AC current at all.
The current source that is keeping the current through the voltage gain stage constant has a control loop circuit at a current gain of more then 10 million. This control circuit delivers any current the speacer draws in addition to the constant current flowing through the voltage gain stage. Due to the extremly high current gain factor of that controller loop the output impedance of a single current source drops down to as little as 6 to 7 milli ohms. So the symmetric amplifiers has an output impedance of about 15 milli ohms (which no other single stage single transistor amp can deliver).
One big advantage of this circuit design is that the distortion produced is independent of frequency and load impedance. It is just related to the output voltage. The distortion spectrum remains homogenious at nearly any load and frequency change.

Very desirable for a true voltage source. I remotely associate it with a similar strategy that was done in the Halcro amplifiers that included a high gain feedbackloop in the outputbuffer. Their basic design was a pimped up version of a standard Douglas Self bipolar though (containing a multiple bootstrapped currentsourced gainstage).

But harmonic distortion is not the only aspect that matters. Group delay, phase shift, bandwidth etc. the entire behaviour within the time domain maters much more. And since my amps are very "fast" also at low impedance loads (1 ohm or less) - the current control loop operates at a bandwith of about 1 mega Herz, the amplifiers bandwith is above 300 kilo Herz - the behaviour within the time domain on complex (low) impedence is best.  

The excellent complex impedance performance should be very favourable to electrostatic loudspeakers!

But all these electronic technical aspects are not really the main thing. We talk about reproducing the illusion of music. That is what matters. In the end you have to have the feeling that someone real is making music in your room just for you. You have to be able to feel the "puls" of the music the attitude of the artist, his "message".  
I tried to fine tune my amp by various means to give that illusion best. The circuit is only the basis. The next thing is the quality of the parts used and their exact values within the circuit (determined in listening sessions). Then the thermal stability and at last but not least the mechanical attributes. The cabinets of my amps and power supplies are assambled the way to be not sensitive to vibration. Every screw for instance is treated at least twice bevor being mounted. Plates are damped so that no vibration can cause any microphonic effect ....

All of this is responsible for the entire sonic result.

Gerd

I cannot agree more on this philosophy. Attention to all details seperates the good from the excellent products. You have got my attention!

Regards, Coen

[GerdS]

Reading this, I am happy not having bought a new amp yet. Peter, have you meassured the high power distortion level of the amp or is there any information? I would need the whole 50W for sure.

Hi,
I have put some plots into my white paper (just in german, but for the plots that does not matter):
http://www.sauermann-audio.de/dokumente/sauermann_white_paper.pdf
But remember music usually has a dynamic range of about 10.
If the amp is fully driven, so that the pulses are at 1% distortion (45 Watt) the RMS level is at about 5 Watt where the distortion is at less than 0,1%.
The amp can be easily overdriven by at least 3dB without audible distortion or compression!
This is due to the circuit design and its' dynamic behaviour.
I would be pleased to give a demonstration.

Gerd

[manisandher]

I would be please to give a demonstration.

WARNING: If anyone takes Gerd up on this, you're probably going to buy the amp. Just saying...

Mani.

[CoenP]

Hello Gerd,

As I look again at the diagram, I start to worry about DC over the speakers and failure of one of the legs. How do you adress a safe operation?

Regards, Coen

[GerdS]

Hey, great to 'see' you here Gerd!

I know you're busy, but if you have a few spare minutes could you share your thoughts as to why my mono amps sound so much 'fuller and extended' in the bass (and yet utterly taut, fast and tuneful) than even my 900W amp?

Mani.

Hi Mani,

there are (at least) 2 aspects about bass performance:
1. Low output impedence of the amp to control the energy of the fundamental. Even though your Quad speakers do not have large (cone and coil) mass to move there is energy stored as magnetic energy in the speakers transformer which has to be kept under tight control. This is the same as for dynamic speakers where a lot of people say that the discussion about damping factor is senseless since there is a coil of 0.3 Ohm resistance or more within the dividing network for the bass. And compared to this 0.3 Ohm resistance a difference in cable resistance or output impedance of some 50 milli Ohm or less does not matter. It does, because the magnetic energy within this coil has to be kept under control as well. Its about energy and not only about resistance.  
2. The illusion of a bass as an instrument is the sum of the fundamental and all the other frequencies that makes the natural sound of the acoustic bass or electronic bass guitar or whatever instrument. As long as we are not talking about a single tone generated by a synthesizer the sum of all these frequencies have to be radiated by the speaker accuratly in phase at the right amount of energy. If the amp does not do this job well and is not keeping the speaker well under control in the mid- and high frequency domain, the "bass" woun't be really good. I can demonstate to you the improvement of the bass performance of a small 2-way speaker system in exchanging the capacitor in the dividing network of the tweeter!

I the case of your setup I think it is a combination of both aspects.

As I wrote in an earlier post I stated that the damping of mechanical vibrations improved the bass performance of my amp. This has something to do with aspect 2. The energy of all frequencies of a tone played on a bass a consistend "at the point" at the right time.

This has nothing to do with rated output power, not even with the frequency response plot of an amp. It is again a subject of energy handling in the time domain at every energy level. The amp has to deal with milliwatt or microwatt as accurate as with one, ten or more watt.
And I'm sorry to say but most of the amps can't.

Gerd

[GerdS]

Hello Gerd,

As I look again at the diagram, I start to worry about DC over the speakers and failure of one of the legs. How do you adress a safe operation?

Regards, Coen

Hi Coen,

As with all symmetric (bridged) amplifiers with unipolar power supply there is a DC offset on the output. Between the output terminals the DC is less than 200mV when fully warmed up 40mV. This setting is long term stable since resistors of 0,01% tolerance are used which will have a long term drift of 0,01% in 10 years (Vishay S102C). Even the potentiometer to adjust the DC offset on the output is a military grade thick film metal poti (Vishay 1280G) which I buy at more than 15,00€ excl. VAT per piece.
The amp is long term short circuit proof by design. If one leg fails there will be always a huge overload in this leg and the power supply DC fuse will blow (special fast blow fuse). So in this case there will be only a short low frequency square wave puls of 1/4 of the maximum peak output rating on the output (about 25 Watt).
That should not harm any type of speaker.

Regards, Gerd

[CoenP]

Hello Gerd

Again thanks for the elaborate answer. I find your design intriguing, since it resembles familiar designs (totem poles) so much but is actually quite different in operation. In the end it is all about how it performs, and it seems it does very well on this critical forum.

Selecting and matching components to keep DC offset under control is a much more elegant way than a adding a servo. It keeps circuits simple but is also much more costly.

Furthermore I wondered why you chose for a balanced/bridged design. Save for the impressive reduction of even harmonics, it is a way to be able to use a single supply. Are the outputs floating or are they in some way referenced to the power "ground"?

Regards, Coen

[manisandher]

I'll be posting more of my purely subjective thoughts on the sound of the amplifier this evening...

Well it's midnight here and I've just listened to 3 albums straight through. I think I'm getting used to the sound of the Sauermann with the Quads. I wouldn't say the sound is totally transparent, but I'm pretty certain this has more to do with the way I currently have things set up than anything the amplifier is doing. What's amazing though (for me at least) is the total elimination of a 'metalic glaze' that I think I've pretty much always had, and something I've always mistaken for 'clarity' and 'hyper-resolution'. This metalic glaze has now gone and been replaced by a much, much more pleasant sort of 'woody' sound. Again, I don't think this is a characteristic of the amp - perhaps this woody sound has always been there but just been obscured by the metalic glaze... who knows?

I've had loads of amps in the past, and I can see now that pretty much all of them have, to varying degrees, had this metalic glaze. It seems remarkable to me that they could all have sounded wrong. But maybe that really is the case. I mean, I now know that all the DACs I've had in the past sounded wrong.

I really am beginning to think that the Sauermann will prove as big a revelation in my system as the NOS1 did.

Mani.

[PeterSt]

Mani,

What you should explicitly look for, is whether the detail is still there; you might need to play the stuff I use for that though (you know). So, glaze gone, but the very high transient "ticks" etc. still there. *Now* you won't have false detail.

... And I didn't notice any of that lacking ...

Btw FYI: I measured the impulse response as I do it for myself to test the filtering and which I normally do behind the DAC. All was followed as expected (this would come down to 29KHz square wave response, but now impulses (only one side of the voltage rails). You can't have it better when THD is OK (I didn't look at that).
What amazed me most is the perfectly straight noise line; This is indication #1 for something good - at least how I look at it. Remember the other pictures you came up with yourself ...

Maybe even more important is that I didn't expect anything else than "just very good", in advance of meeting Gerd. Of course, this "very good" is (to) my standards only, but since I have them anyway ... Some people can let know that their product is good, just by talking about it;
I use to have no conficence at all in whatever comes by. This time it was the complete other way around. And without further thinking about it, it just worked.

Whatever it is you further find, I hope you can map it onto real merits of things. It should (IMO) be similar to the NOS1; nothing is perfect, but at some stage you can see that whatever it is what's not good, is encouraged by something else. In this case, *if* you come to the conclusion not all is 100% transparant (difficult to achieve to begin with), maybe try your other speakers. Reason further from there.
What I like to see (for you, for maybe many more) is that this amplifier *is* the most transparent (because I judged it so), and thus when it doesn't show that, something else might be going on. And oh, not that you so much talked about this, but aren't we hunting for the best playback ever ?
That's why. The more "sources" can be taken as close to 100% OK, the more easy it will be to blame cables etc. by just exchanging what's left to be unsure.

Peter


PS: Might it help someone ... Gerd Sauermann uses the exact same approach for his designs as I do. This is up to the software, although Gerd didn't go that route.
Since this approach is not the most common one, it was really strange (for me) to find this out, which was only afterwards ...

[manisandher]

What you should explicitly look for, is whether the detail is still there... So, glaze gone, but the very high transient "ticks" etc. still there. *Now* you won't have false detail.

Oh there's plenty of detail... but no glaze now. And what's interesting is that this detail is throughout the whole bandwidth, not just in the treble.

Btw FYI: I measured the impulse response as I do it for myself to test the filtering and which I normally do behind the DAC. All was followed as expected (this would come down to 29KHz square wave response, but now impulses (only one side of the voltage rails). You can't have it better when THD is OK (I didn't look at that).

Probably the best description of the sound I can give is of it being 'totally coherent'. Nothing, absolutely nothing 'jumps out'. Everything seems in it's place and is a pleasure to listen to. A great impulse response no doubt contributes towards this.

What amazed me most is the perfectly straight noise line; This is indication #1 for something good - at least how I look at it.

All I can say is that this is easily the 'quietest' amp I've ever heard. And I mean quiet through the speakers, not mechanically (although you can put your ears right next to the outboard power supplies and you won't hear a thing). The background really is totally black. This is best heard on classical music during really, really quiet pieces. The level of detail coming through even at these low volumes is very impressive.

...*if* you come to the conclusion not all is 100% transparant (difficult to achieve to begin with), maybe try your other speakers.

Not sure which 'other speakers' you're talking about. My smaller Quad 2805s? Well they have a similar sonic character to my bigger 2905s, though it would be easy to set the amps up in my study/office and take a listen. The only other speakers I have here are my old Celestion SL600s - great speakers, but not really in the same league as my Quads, I'd say.

What I like to see (for you, for maybe many more) is that this amplifier *is* the most transparent (because I judged it so), and thus when it doesn't show that, something else might be going on.

Yep, I'm sure there are a number of other things going on. Once I have some more time, I'll do some experimenting. Right now, I can certainly live with the sound I have.

What's interesting is that if anyone were to come and visit to take a listen to my system, I would be almost embarassed. There is nothing 'amazing' about the sound at all. But that's why I like it. The system gets out of the way and allows you to focus on the music.

Mani.

[GerdS]

Furthermore I wondered why you chose for a balanced/bridged design. Save for the impressive reduction of even harmonics, it is a way to be able to use a single supply. Are the outputs floating or are they in some way referenced to the power "ground"?

Regards, Coen

Hello Coen,

if you drive the amps with a balanced input signal the outputs are floating.
Correction: The output is allways floating, no matter what the input is. There is no relation to the internal ground (- pole of the power supply) nor to earth ground. So you could even hook one output pin to earth ground without problems (what some messurement equipment do when testing on complex load).   

This are the two most important reasons for a balanced design:
1. Symmetry as such to ensure the best protection against any influence from outside.
Think of asymmetric circuit designs with a direct relation to ground. The signal on interconnection cables is directly related to ground since one wire is the grounded shield as well as the ground pin on the mains plug. So the signal might use (and is to a certain extend) the ground wire of the mains as the interconnection between the two connected pieces of equipment. To get best results most expensive mains cables have to be used.
The reduction of even harmonics is a nice side effect that helps a lot to get the best puls response (in my opinion).
2. The circuit I developed in this simplest form is only working on an unipolar power supply. Therefor the output has DC offset. To get rid of the DC a balanced design is the best in terms of sonic result.

Regards, Gerd

[PeterSt]

Hey Gerd,

So you could even hook one output pin to earth ground without problems (what some messurement equipment do when testing on complex load).

I would never do this !
This assumes earth ground to be equal to the (created) potential of the amplifier ... which it will not be.
And there you will have DC offset ...

Or ?

Regards,
Peter

[manisandher]

A quick update on how I'm getting on with my Sauermann monoblock amps.

I wouldn't say the sound is totally transparent, but I'm pretty certain this has more to do with the way I currently have things set up than anything the amplifier is doing.

Well, I was wrong and right at the same time!

These are the most transparent amps I've ever heard. This is most obvious with complex music, especially large orchestral pieces. It's as if each instrument has it's own amplifier/speaker channel. Nothing gets muddled up - each instrument remains totally pure irrespective of what's happening around it. I've never heard this before. Actually, with the Sauermann, I think it's a case of 'you don't know what you don't know' - I didn't know that all the other amps I'd heard were 'distorting'... until now.

But I was right about something: the way I had things set up wasn't optimum. And I'm talking mainly about my PC(s) now. I've been a long-time advocate of totally silent PCs. And for good reason I think - I just abhor even the slightest background noise when I'm listening to music (usually late evening once the kids are asleep). So my music PCs have always used fanless cases, fanless PSs and also SSDs. But I did an experiment a short while back and although it pained me no end, I had to concur with Peter that HDDs sound better than SSDs. The music sounds more 'open' and full of life. All the SSDs I have really sound 'filtered' in comparison. So a few weeks ago I decided to build a 'monster' PC and stick it in my basement, as close under my listening room as possible. The PC ('Le Monster') was completed yesterday. It has a total of eight 120mm fans and although it is much, much quieter than I had expected, it certainly isn't something I would want in any domestic enviroment. Nevertheless, I temporarily placed it on the floor next to the NOS1 and connected everything up, just to take a listen. And my God, what a difference to my 'totally silent' PC. Actually, the sound was a bit too forward for my liking so I played around with the SFS and settled on 475 in the end. I'm only now hearing the gorgeous 'sweetness' of the Sauermann amps. They sound truly outstanding.

I suppose I should also mention the quality of the Quad 2905 speakers. They just seem to spit out what they were given. Oh the sound won't be to everyone's liking - the LF dynamics are still somewhat lacking - but hell, I can't imagine a faster and more cohesive sound from a pair of speakers. And now there's a beautiful 'sweetness' thrown into the mix which is intoxicating.

I remember writing a post here immediately after I'd heard the original NOS1, saying something like, "beg, borrow or steal the money for this DAC" - the first time I felt so sure about any component. Well here it is again. If you're in the market for an amplifier [and can cope with its class A operation], beg, borrow or steal the money for the Sauermann. It's on-par with the NOS1... and sorry Peter/Ciska, but much better looking and better built. But just like the NOS1, it's been born out of sheer passion for the perfect sound.

Mani.

(PS. Yes, I know this sounds like advertising. But just for the record, I'd like to say that I paid the full retail price for these amps. It's not because I have a vested interest that I started this thread... rather, it's because I share Gerd's passion for reproducing music faithfully, but just don't have the capability to create something like this myself.)

[christoffe]

So a few weeks ago I decided to build a 'monster' PC

Hi Mani,

interesting to read. I was just on the road to buy a silent PC.
So I owe you some buckets of beer.

Joachim