DCAudioDIY.com

Don't worry, Charlie: no one listens to AM radio anymore. Just watch your harmonics (both up and down)

_________________
- Guy

I wouldn't call the soundstage that Charlie's system produces anything but enormous in all dimensions, and the ability of David's 845s to improve the stage is impressive and unexpected. We heard improved detail and transparency, deeper, tighter bass, better top end extension and delicacy, even a more holographic sense where you can touch and walk around the musicians, without losing tone richness or image density, they are superb. I'll bet that having no output transformer is a major factor in the soundstage improvement. I fondly recall my Futterman OTL monos driving Dayton-Wright, Quad, and Martin Logan stats resulting in a sound I wish I had now. I wish Charlie all the best in his new 845 project, try to go OTL if possible.

Ok. Here I go into the lions’ den. I am going to offer some opinions as to why you heard what you did at Charlie's which I am sure are going to piss off a number of you. I am also going to get David into the middle of this because his great skills produced an amp that you all heard as very revealing and eminently listenable. Now for the motherhood and apple pie section.

In order for any amp to sound good it should have relatively low distortion, it should either have enough power to not overload or if it does overload it should do it with the least amount of damage to the signal as possible. The frequency response should be flat and the bandwidth at any power level sufficient to cover any audible frequency. Frequencies that may be present outside of audible limits should not disturb signals within the audible range. It should be free from interaction with AC power problems. That is the amplifying section should only see the power supply, not the line. In order for that to be true, if absolute regulation cannot be assured, the time constant in the power supply with any current demand should be long enough that any change to power supply voltages should be below the audible frequency range. Noise should be sufficiently low that given the sensitivity of the speakers in use, noise does not mask fine details.

How can you design an amplifier that will meet these requirements? See David. Seriously, he does an excellent job of identifying those areas that have direct effects on sonic behavior. David, if my recollection of our many conversations is in error, please provide corrections.

If you look at any of his power supplies you will see that they are very robust and truly isolate the amplifying sections from the foibles of the AC line. What does he feel about using power conditioners vs designing a proper supply? Ask him.

Does he have an opinion about “golden ear” power cords? Ask him.

Does he use a ZOTL circuit because it is simpler than an output transformer? NOT! Check out his web site.

Does he feel there is only one type of component that will produce the ultimate sound quality in all cases? Ask him.

Does he feel that any amplifier design and implementation that produces the same signal at the speaker terminals will sound the same? Ask him.

Now for my opinions. There are many ways to design a good amp. It does not have to be tube or transistor or single ended or push pull or any of the other sacred cows that people cling to. With each design there are the inevitable tradeoffs. A good designer will know what compromises to make. Here are a few of my takes on the process. BTW, I am not an amplifier designer in any stretch of the imagination, but I know a good one when I meet him/her.

If an amp is not going to use feedback it must be very linear in all stages. That is why many tube amps that use transformers and have no feedback run into problems. Transformer non-linearities are a real PITA(Pain In The A..). If you use feedback with a transformer, phase problems in the transformer give rise to very difficult problems and compromises. Keeping the process linear requires extensive knowledge of topology, devices, and the many other “minor” details that can derail a design. As we previously heard at David’s house, even the values of damping resistors in a tube amp can affect the sound.

When feedback is used, there are some giant holes into which you can fall. At the very minimum, if the gain bandwidth product of the circuit is insufficient, feedback will cause more problems than you can imagine. Dealing with phase problems can also be tricky. However, when it is done properly, feedback can result in a very pure signal. I should not be a crutch to fix a poor design as that will never work properly.

Here comes the stuff now where you start throwing bricks at me. You can take SET amplifier, a 845 PP amplifier and an all solid state commercial high powered amp and have them all sound the same. David and I did some long term listening tests between his 300B SET and my Crown Macro Reference amp driving my BG RD75s. Granted they are a fairly pure 6 ohm load which negated differences in the output impedances of the amps. We also tested another smaller amp that David had and his 845s. When we balanced the drive levels and listened, we were both shocked. David was shocked because regardless of the source material, as long as the 300B was even close to its maximum output, neither of us could hear a difference between the two amps. I was totally shocked at how loud the 6 wpc SET could drive my 87dB sensitivity speakers. In fact the average level had to get to round 96 dB before the 300B had problems. Amazing! That was because the SET was so benign even when driven to clipping levels that if it were a solid state design it would have you running screaming from the room.

We also did the same thing with his 845s and the only difference is that they could play louder than the 300B before they ran out of steam. His other small amp did sound deficient compared to the others when first auditioned. He went back and made a change that should have only affected frequencies well above the audible range but it then sounded like the rest with the problems gone. Go figure.

What does this have to do with what you heard? Simply you heard what a really good amp can do with a good set of speakers. I will venture to say that you can achieve the same results with a good solid state amp. OUCH! One of the bricks just smacked me in the head.

What I would expect between David’s 845 and my Crown on Charlie’s system would be if anything a change in timber because of the difference in output impedances reacting with the variable impedance vs frequency with his speakers. David and I have spoken about that and a good test would be to match the output impedance of the Crown to the 845s with resistors and see what happens.

I’ll make a suggestion. We should at some point go back to Charlie’s with David’s 845s and my Crown. We should first listen with the amps straight out of the box. Then we should change the impedance of the Crown with the appropriate resistors and listen again. I think it would be a fascinating experience. Any comments?

Tom,

Sometimes listening comes first and engineering comes later. Here is a blurb about Stewart Hegeman who famously commented something to the effect that if it sounds good and measures bad, you're measuring the wrong thing. I think we are still to a large degree measuring the wrong thing. I was a professional musician for 10 years and spent a lot of time listening to little nuances in music. They're there.

-----

''I always wanted to be a concert violinist,'' Mr. Hegeman told an interviewer not long before his death in 1986. ''But I was never any good on the fiddle. For me, engineering turned out to be the easiest way to get close to music.''

This remark points to his method. Stewart Hegeman always put musical judgment ahead of established engineering practice - an approach that at the time struck most professional engineers as exotic, to put it mildly. Despite his impeccable credentials, he could muster little sympathy for a music-oriented approach in an industry that still regarded sound equipment as plain electronic hardwaree. Working alone, he was a holdover from the days when individual inventors - relying on little more than creative imagination - laid the technical groundwork for entire industries.

While serving as a consultant to Harman-Kardon, a pioneering audio company, Hegeman developed a theory of amplifier design based on ultra-wide frequency response. He argued that the frequency response of an amplifier should not be limited to the range of human hearing (20 to 20,000 Hertz) but should extend far above and below it.

This notion was promptly shrugged off. Why bother with frequencies you can't hear?

Hegeman countered that the narrow response limits of amplifiers at that time actually reflected distorting waves back into the audible range - somewhat as a sea wall throws back the surf. But at the time there was no way to confirm this through measurement. The argument could only be stated in complex mathematical terms understood by few practical engineers. Besides, few were musically astute enough to appreciate the difference in sound (as demonstrated in several amplifiers built by Harman-Kardon on the Hegeman principle), and so the issue was largely ignored.

It took nearly two decades for Hegeman's wide-band principle to be recognized as essential for low-distortion amplifiers. Today, nearly all makers of high-quality audio amplifiers adhere to it, at least to some degree. Harman-Kardon, the company for which the original research was done, has refined this principle in the current models of its exemplary Citation Series.

Hegeman was also far ahead of his time in recognizing that negative feedback - a widely used method of reducing measurable distortion - was itself the source of different kinds of distortion that went undetected in the laboratory but gave the sound a strange touch of brittleness. Again, Hegeman could confirm his hunch only with his ears and his slide rule, still lacking appropriate tools to obtain supportive laboratory data.

---------

It may just be that reason Dave's amps sound so good is that they are incredibly linear without feedback.

Charlie

Charlie:

I didn't say anything about measurement except for setting levels It's all about listening. I think it woule be a very informative listening test.

Tom

Watching for bricks coming my way, I recommend the test not involve an A/B switch, which could be the weakest link in the chain, overlaying its sound over the two units being compared to the extent that they both sound the same.

(Sometimes it's fun to poke sticks at lions.)

_________________
- Guy

Actually, I think we've hijacked this thread from its original intent to simply thank some people into arranging an equipment dual. We should move the discussion to another new topic and leave this thread.

_________________
- Guy

Good idea guy. I'll move it.

Tom

A few years ago when I was breadboarding different SE amps, I used a fix bias so I could easily swap various 6L6(pin out) output tubes and re-bias them as necessary. Then I decided to try something a little different and used a cathode follower output, so my output transformer primary was placed in the cathode circuit of my output tube. The good news was that my Chinese 6L6GC sounded very close to my Phillips 6CA7 The bad news was that I noticed my sound-stage had collapsed to about 1 foot in front of the speakers, and the overall presentation was "relaxed". So my theory is that 2nd harmonic has an effect on the perceived size of the sound stage. Of course, the down side is that too much 2nd and you'll lose the presentation when the music gets complex. If I had a PP amp that I wanted to alter the size of the soundstage I would look at a driver or gain stage that has slightly higher 2nd harmonic than whatever I am currently using.

My experiment is easy to duplicate, so it's worth trying out to see for yourself. Just be prepared to build a robust driver, with enough swing to make any power at all.

Yikes! 2 years since I started this thread, and 5 YEARS since Stuart invited me to his house (see post #1) to hear his stereo. Was finally able to get together with Stuart at my house to listen to music. Stuart, that means we're on pace for 4-6 more meets max, in this lifetime!

Just wanted to reiterate how much I appreciated his initial invite, and how the smallest decisions we make from day to day effect where we go, and who we know. I would never have become friends with Charlie, Dave B., Dave P., Tom, or Bob if not for Stuart. I consider all those guys friends and have no doubt they've made my life more interesting.