Amplifier Circuitry

[rollo]

   There are many ways to skin the proverbial cat. Single Ended design is the oldest and trusted circuit. Over the years Push Pull and now class "D" operation.
    Is the SE circuit still King ? Or have newer designs bettered or equaled SE.
    For me it is hard to understand when not using a SET design that the signal is split using a phase splitter. Kinda like the Humpty Dumpty effect. Meaning how can it be put back together properly.
    Why do newer designs need to be so complicated ? Then there is the issue of even and odd distortion.
     Now apply amp design to speaker choice. A high efficient speaker with SET or say low efficient speaker and push pull or other design ?
     Is one option actually better or just different ?


charles
   

[Werd]

Depends on what floats your boat mixed with convenience. If one likes the sound of music scaled down to the least amount of electronics. A 3 watt SET with no preamp into a fullrange speaker. You  have illuminated a host of electronics but at the expense of only a single volume and a single driver.  Bass and high frequency is nerfed but who cares you get the impact of resolution from a source driving fewer electronics into midrange frequncies. If you want bass now you need a preamp to drive the low end and maybe you might use an A/B because the low frequency air from your woofers makes it fun.

[steve]

    Is the SE circuit still King ? Or have newer designs bettered or equaled SE.
    For me it is hard to understand when not using a SET design that the signal is split using a phase splitter. Kinda like the Humpty Dumpty effect. Meaning how can it be put back together properly.
    Why do newer designs need to be so complicated ? Then there is the issue of even and odd distortion.  

charles
    

I will try to answer your questions in order.

1) I designed SE for sometime but now will take PP any day of the week. Pt. 2 will answer why.

2) To answer your second question, the signal is not "humpty dumpty effect", or taken apart. However, quality does depend upon the type of phase splitter.

By the way, SE is often too simple for maximum fidelity. For instance separate power supplies is a plus when designing. Using a common power supply is a weakness for sure. So the best SE is not that simple.

Current through the plate resistor is the same as current in the cathode circuit, so the same signal current (Kirchoff's laws). What occurs in the plate circuit occurs in the cathode circuit, except at the very high frequencies does this deviate. That frequency depends upon the plate resistor value. For other types of phase splitters, there may be little if any difference at high frequencies.

3) It does not have to be so complicated. For instance my amps are only two total stages.


SE amps have their own inherent weaknesses, so not really the cats meow either.

A. Chokes or some sort of regulation needed to limit power supply ripple, noise etc. However, it must also be superior from the signal/output side of the power supply/regulator.

However, the choke's reactance varies widely vs frequency, so the separaton between filter capacitors on either side of the choke varies wildly, causing problems. It is almost like capacitors changing values except phase and feedback between stages occurs.

A chokes reactance is given by XL = 6.28 X Frequency X Inductance. As the frequency changes, the inductive reactance changes, wildly, by 1000 times between 20hz and 20khz.

A 1 henry choke has 125,000 ohms reactance at 20khz.
At 20hz, the reactance is only 125 ohms. So the filter capacitors almost appear to be added together.
For 5 henries, 625,000 ohms at 20khz, but only 625 ohms at 20hz.

If some are hyping how heavy or large chokes are, you are being misled. However, SEs need chokes or regulators by definition, a weakness.

B. Damping factor is weak in almost all cases. Besides that, the damping factor varies between signal halves and the amplitude of the signal. What a mess.

C. Distortion is rather high, even at 1 watt output.

D. Higher order distortion products are usually present in SE designs because of the large drive signal required for the output tubes.

300bs require 100 volts peak. Larger tubes require 200  even higher drive voltages. Driver tubes create distortion which combines with output tube distortions. So higher orders of distortions are created.

Besides that, the intermodulation distortion products are some 3 times that of harmonic distortion products. So at 5% HD, the IMD is some 15%. That is why many say that  orchestras sound congested vs single instruments.

E. The high frequency response is generally quite poor, often several tenths down at 15khz or lower. Low frequencies are also often down. This gives the impression of dominant and clean midrange. Problem is that the midrange can be just as open, sound more natural, and provide clean, accurate highs and lows.

Common power supplies, which is cheaper at the expense of higher quality reproduction, in both SE and PP amps.

Finally, if one actually tests for artificial flavors, one will find the SE amp not that good either. So not really that honest to the music. The more complex the music, the bigger the problems.

SE amplifiers may look simple, but they are far from it, and/or they have real problems. Worst of all is that the vast majority of all "designers" of both PP and SE amps simply copy another's work, or use a few equations, nothing more. So we do not get to hear the maximum quality from each type. We audition compromises.

Lastly, I am retired and taking care of Mom.

Cheers
Steve

[rollo]

  Thanks Steve nice to see you posting here. Please explain again the phase splitter. Does it not split the signal and reconstruct it ?
  I would imagine there are better designs of splitters available.

charles

[KLH007]

   There are many ways to skin the proverbial cat. Single Ended design is the oldest and trusted circuit. Over the years Push Pull and now class "D" operation.
    Is the SE circuit still King ? Or have newer designs bettered or equaled SE.
    For me it is hard to understand when not using a SET design that the signal is split using a phase splitter. Kinda like the Humpty Dumpty effect. Meaning how can it be put back together properly.
    Why do newer designs need to be so complicated ? Then there is the issue of even and odd distortion.
     Now apply amp design to speaker choice. A high efficient speaker with SET or say low efficient speaker and push pull or other design ?
     Is one option actually better or just different ?


charles
   

Charlie, Allow me to add another "new" architecture, Direct Digital: The pristine signal from your source is sent via SPDIF or USB into the amplifier. The SPDIF signal is a PCM (Pulse Code Modulation) signal, which is a type of digital signal transmission. This PCM signal is converted to a Pulse Width Modulation, which is another form of square wave containing amplitude and timing information. As a PWM, the signal is reclocked, amplified in the digital domain, reclocked again, and then demodulated to drive a speaker directly. There is virtually no analog in the signal path all the way up to the binding posts.
Lyngdorf, NAD, and Core Audio are using this method with success. Any analog sources must undergo A/D conversion first. The Lyngdorf TDAI-2170 also includes Room Perfect DSP correction as well, since the room is part of your system it seems like a good addition.

[steve]

 Thanks Steve nice to see you posting here. Please explain again the phase splitter. Does it not split the signal and reconstruct it ?
  I would imagine there are better designs of splitters available.

charles

Thanks Charles. These days I am retired and taking care of Mom, so time is somewhat limited, but nice to post from time to time.

The signal is not split and reconstructed in the sense of a signal falling apart and then put back together. In the most simple phase splitter, the signal is actually created like a typical common cathode circuit but divided in two, with two halves but 180 degrees out of phase. Half of the plate resistance is placed in the cathode side of the tube so half the signal is there, but again out of phase.

In more complex circuits, some take a portion of, say gainstage T's output and is introduce it into another gainstage U, where the output is close to the same amplitude as T, but the phase is approximately 180 different. Some designs are automatic balancing phase splitters. (I use the word phase instead of polarity because at the frequency extremes, the polarity is not exactly 180 degrees different.)

Another type of circuit drives gainstage T and the cathode of gainstage T is connected to the cathode of gainstage U. The output signal's amplitude is approximately the same as T but 180 degrees out of phase from T's output signal. The cathode signal is extremely close (nanoseconds) to the input signal.

These are just a few of the major types of circuits out there that produce two signals of near equal amplitude but 180 degrees out of phase over most all, if not all the audio band plus.

Cheers and hope this helps Charles.
Steve

[rollo]

  Steve helped big time. Thanks for taking the time. Hope Mom is doing well.
  KLH always a pleasure to hear your input. Good info thanks.


charles

[steve]

  Steve helped big time. Thanks for taking the time. Hope Mom is doing well.
  KLH always a pleasure to hear your input. Good info thanks.

charles

Mom is doing about as well as can be expected at 88. It is getting hard to walk plus she does not drive anymore.

I think I can go deeper, even though the general public is also reading, and still keep things simple enough for them.

The simplist has the advantage of few parts, but that is not necessarily a great idea. The disadvantages are the output impedances of both outputs (out of phase) are usually quite different, so the very high frequency response is somewhat different depending upon the tube used.
 
Another disadvantage is there is no gain, so a wasted stage so to speak, compared to other types of phase splitters.

The middle example has one advantage of equal output impedances from both outputs, even though out of phase. I also mentioned self balancing, depending upon the circuit, which is always good.
A disadvantage is more complication, has time delay between the two outputs, dependent upon the tube(s) used. Low frequencies between outputs would also probably be different.

The last example has the disadvantage of one output tending to have a little more output than the other output.
The output impedances are basically the same, and only one tube/stage is needed for both gain and phase splitting.

So one has to overcome at least one obstacle in each type of design.

Cheers
Steve

[tmazz]

A) Nothing in electronics is perfect. Every type of circuit no matter what it is has its own set of design challenges.

B) By the very nature of engineering itself, every design is the result of a number of cost benefit decisions. You can always get better by spending more money and you can always save money by sacrificing quality. But at the end of the day, you have to be able to sell whatever it is you are designing So the engineer (in conjunction with others in the company) needs to determine what is the minimum quality that the buyers in your target market are willing to accept and how much are those same consumers willing to spend before the price puts them off. And then the engineer must design something that can be sold at a price somewhere between those two points. (And keep in mind that the minimum acceptable quality does mean something that puts out poor SQ. If your target market is to sell a $10K amp for example, the minimum acceptable quality to somebody willing to shell out 10 grand is pretty darn high.)

Everything we do in this hobby, whether from the viewpoint of a designer and manufacturer or of a consumer looking to build a system is the result of a series of compromises. We can never get it all, so the best we can do is decide what is most important to us in a music system and make our compromises in such a way that  they support those preferences while giving up things that don't matter as much to us.

Along those lines I have always felt that the biggest drawback to SET amps has not as much do do with the amps themselves as it does with the systems that are built around them. Because of the low output wattages  that go along with SET designs, the amps must be used with very high efficiency speakers in order to get any kind of reasonable SPLs out of the system. The problem this causes is that because of the ultra high efficiency inthe speakers, the tiniest little bit of noise can become very loud acoustically. So great care (and often great expense) must be taken in the design of not only the system , but also every other component in the system  to make sure that the noise level is low enough not to become annoying. So great care must be taken in the selection and setup of all components used with a SET based system for good results to occur. While great care should be taken with any system setup SET based system are much less forgiving of less than perfect situations.