REGISTER NEW USERLOST PASSWORD? WELCOME, Logout

Volume controls

Paul's Posts — 04 August 2012

By

Volume controls

Well, this is a fine hornet’s nest I’ve gotten into.  Half of you seem to think I live in a cave and haven’t any clue what high-end audio is all about and the other half think I have it right – and then yet another half think I am opening Pandora’s box and letting all the digital nasties out to seek and destroy all of analog.  Wait a minute, how many halves are there?

For the record (and not to make a pun) I love analog, have been designing analog circuits for over 40 years, continue to do so and near the end of this series will be advocating a hybrid approach as part of the new paradigm I am advocating that plays nice with both analog and digital.  So, deep breath my friends, I am not out to kick your dog or put a damper on your analog dreams and I hope we can simply enjoy these posts in the spirit they are written, fun and informative – forward looking.  I really am on your side and the side of great music and audio.  Really.  :)

One of the core elements of any preamplifier is the volume control – without it you cannot adjust the level of what you’re playing and the system becomes rather useless.  Let’s talk analog volume controls today.

There are 50 ways to Sunday to build an analog volume control and I think I may have tried and listened to every single one.

In its simplest classic form most analog preamplifier volume controls are but a simple variable resistor or potentiometer (called a “pot” for short).  How do these work?  They are simple mechanical devices that have a long strip of resistive material with a connection on each end of the strip.  One of these two connections (A) is attached to the audio source and the other side (B) is connected to ground (zero audio).

A third contact (W) is pressed against this strip and connected to a knob or a lever so that the user can move the contact up and down the length of the resistive strip.  This contact becomes the output of the volume control.  By placing the contact on a different area of the strip you get different volume levels out – louder when you move towards A and softer as you move towards B.  Put two strips together with two contacts moving together and you have a dual potentiometer or – a stereo pot.  Move either one of the two contacts differently and you adjust the left and right channels differently – thus making a balance control.

resistor Volume controls

Remembering that all types of resistors sound different when used in the signal path, it should be no surprise that the quality of the music running through one of these pots is very much dependent on the quality of the resistive element, the contacts and the way it’s implemented.  In fact, there is no such thing as a neutral or transparent sounding pot or volume control.

The fact that any form of volume control has a negative effect on the sound is a really important concept to grasp.  Switched attenuators, Gain Cells, pots, electronic attenuators, CMOS switches, photo cells and every scheme imaginable to change the volume has a negative affect on the sound.  The challenge is to figure out how to do as little damage to the sound as possible and there are many schemes out there – ranging from the absurd to the lame and cheap and everything in between.

The point I want to make in today’s post is that the volume and balance controls in an analog preamp all suck to some degree.  We have to accept that we cannot change the volume without affecting the sound.

Lastly I want to give you a piece of advice when it comes to volume controls.  Less is more.  By that I mean the higher the volume control level the better it sounds because less of the resistor is in the signal path.  Look at the drawing I included – if the W connection is tied right to the audio input A (full volume), then essentially there is nothing in the signal path to damage the music.  The farther away W gets from A (turning the level down), the more sonic damage is done.  Less is more.

I wrote about this nearly a year ago in a post labeled Step on the gas and I think it’s not only worth repeating but it also leads us to tomorrow’s post about digital volume controls.

Interesting how in both cases, less is more when it comes to volume controls.

email Volume controls Forward to a friend and help us engage more readers

Get new and fresh stories like this each morning by joining the folks reading Paul's Posts. Click here

Related Articles

Get that mole! Treat the cause or the symptoms? Jelly rolls If you have to ask …..

Share

About Author

Paul McGowan

Paul McGowan is the CEO and co-founder of PS Audio Inc. a Boulder Colorado design and manufacturing company of high-end audio products and services. McGowan has been designing and building high-end products for nearly 40 years. Hobbies include skiing, music, hiking, artisan bread baking, kick boxing and cooking. He lives in Boulder Colorado with his wife Terri and his 4 sons.

(9) Readers Comments

  1. petewilson

    August 4, 2012 at 8:45 am

    Paul

    I think that another aspect of the traditional volume control which can cause evil is that you’ve often got the possibility that the pot wiper will develop some minor oxidation, whereupon you run the risk of having a non-linear (that is, its resistance varies with signal level) element in the audio path. Subtle (or not so!) damage to the microcontrast of the music is likely to ensue…

    Same happens with ordinary selector switches.

    (You can fix the switches by replacing them with sealed relays, of course)

    – P

  2. hahax

    August 4, 2012 at 12:44 pm

    Paul

    I thought I should tell you. I just logged in to this site and Kaspersky Internet Security(which I just installed today replacing McAfee) warned me there was a trojan on the file. You might check your computer.

    By the way I recall a review of a German amp in Stereophile that claimed the volume control worked by varying the gain of the voltage stage rather than by the variable resistor of a conventional volume control. But I still think you have to be able to cut the input voltage some way. Otherwise you could only add volume with the control, not reduce volume. Thoughts?

    • August 5, 2012 at 7:39 am

      Yes, we built such a product ourselves called the Gain Cell which basically used a voltage controlled amplifier to change the gain and while it was the most neutral I have yet heard there simply isn’t anything you can do to change a parameter in the amplification circuit and not have it change the sound – since the sound itself is a function of all that made it – changing those elements has the unfortunate affect of changing the result.

      • Soundminded

        August 5, 2012 at 8:17 am

        Aren’t you forgetting something? The perceived quality of sound is a function of loudness itself? Take a look at the Fletcher-Munson curves or whichever other comparable curves you believe in. They all tell you the same thing. Beyond the obvious conclusion that the human ear is relatively insensitive to low and high frequency tones at low sound pressure levels notice the spacing of the lines. This graph shows you equal perceived loudness as a function of frequency compared to equal measured loudness as function of frequency. Notice that at the frequency extremes the lines are close together while in the middle, say around 4 khz they are relatively spread apart. This means that in the bass and treble, small changes in gain will have a far greater perceived change in loudness than at middle frequencies. Therefore if you do not play the recording at the right loudness it will not sound flat. What this demonstrates is that the concept of flat measured frequency reponse of the playback system by itself for this reason alone (there are so many others) will not assure anything like realistic timbre of recorded music. Therefore, the inability to control the system to meet this and other variables that affect perceived tone is a fatal flaw of omission in the system design.

        One more point. When you connect a potentiometer to two stages to adjust system gain, you are creating a series impedance between the source and load on the potentiometer as well as a parallel load. As the volume is lowered, the series impedance increases. The source impedance of the previous gain stage had better be low and the load impedance of the subsequent stage had better be high or the potentiometer itself will load the source when viewed from the point of view of the load. Diagram the Thevenin’s equvalent circuit and you’ll see what I mean. Best to always choose a low impedance source, in this case an emitter follower or cathode follower followed by a high input impedance load like a Mosfet and carefully choose the impedance of the potentiometer to diminish this problem.

  3. straylight

    August 4, 2012 at 9:40 pm

    “there is no such thing as a neutral or transparent sounding pot or volume control”.

    I built my own headphone amp a few years back (based on a Sheldon Stokes design) and used a stepped resisitor ladder – attenuator approach. Each 3db step of attentuation for L/R are seperate resistors on the knob/switch…works fine and is the most noise-free and precise volume control I’ve ever heard. Expensive…yes…and requires some military-grade resistors and some good ol’ fashion soldering technique…but sounds wonderful. Many of the classic CJ and Audio Research preamps used the same solution with excellent and long lasting results.

  4. Soundminded

    August 5, 2012 at 7:15 am

    How do you know? How do you know it’s the control element and not merely the act of changing the gain itself? What is there about the control that could cause it to change sound quality? Parasitic inductance and capacitance in real potentiometers not accounted for in the mathematical model? Chanigng the load resistance on the stage before the potentiometer or the source resistance after it? If that’s the reason, it’s the circuit design itself that’s flawed because the potentiometer wasn’t taken into account. Electrical engineers are familiar with the concept of gain bandwidth product. As the gain increases, bandwidth decreases. Is that what’s happening? (This is the real story behind so called slew rate distortion or Transient Intermodulation distortion, lack of sufficient bandwidth at higher power levels. Perhaps there is a similar concept that should be developed of power output bandwidth product or voltage output bandwidth product.) Then there’s the problem of “nested feedback loops. This is where not only individual gain stages have negative feedback but there are one or more additional feedback loops encompassing more than one stage. If there is a potentiometer in such a loop between stages that’s a potential source inviting disaster. BTW, how does the introduction of a variable resistor affect those impossibly complex feedback loop equations? Are they a factor either changing the forward or reverse gain or loading down the output of one stage to a varying degree? This could be power supply dependent. Many audio components IMO could be substantially improved simply by providing a better power supply.

    The notion that it is the potentiometer itself that’s the source of the problem raises a lot of red flags. If the problem is blamed on parasitic reactance, then the question of what happens to other types of circuits where trimmer pots are used to adjust gain or “sensitivity” such as in calibrating analog control circuits that operate at very low or very high frequencies? Wouldn’t that be disasterous causing unacceptable degradation of performance? Yet these types of systems were used successfully for decades throughout industry. They are being replaced by DDC, direct digital control which also require sensitivity calibration and would suffer the same problems.

    I always hear a lot of talk from people who manufacture in this industry and their customers who parrot their words back that are subjective with hypothetical qualitative explanations but with no backup of measured data correlating what they say they hear with what they claim is the correct explanation. This is a general characteristic which destroys credibility these people have in my mind at least. One example is “digital jitter” in wires. Lots of talk but no proof or measured thresholds and demonstrations that one product versus another crosses that threshold. Nothing personal Paul but in virtually every ABX test where these theories actually are challenged one way or the other, they always seem to come up short. Where I was trained, theoretical explanations were backed up with mathematical equations and lab test data that demonstrated that the theory and equations were valid. For this reason, I have to discount practicaly everything I read and hear about high end audio. The term itself is insulting and degrading when less “exotic” and expensive designs prove to be of equal or comparable performance when fairly tested in use.

    • August 5, 2012 at 9:06 am

      Well, good points – so let’s clarify. First, when I suggest volume controls harm the sound I do a straight wire bypass to determine this – and that basically means both are tested at the same loudness level – just one has the element inserted and the other doesn’t and you compare the two to see if there are any differences.

      As to your comments about proper source impedances – I could not agree more.

      • acuvox

        August 5, 2012 at 11:48 am

        I am curious about your “bypass” experiment. A volume control can’t have the same gain as a straight wire, so your circuit has to have extra makeup gain to match levels. How can you separate this affect from the bypass?

        Also, have you tried bulk metal pots?

    • acuvox

      August 5, 2012 at 11:38 am

      We know it’s the ears. How? We have had for fifty years balanced, stepped, constant input and output impedance attenuators using precision ultra-low-noise, non-inductive resistors in shielded cans using wiping gold-on-gold contacts with no ferromagnetic materials and only low dielectric absorbtion materials near the signal path. I am not saying this perfect, but it is orders of magnitude better than the measurable effects of volume changes on hearing.

      Charlie Hansen’s FET switched ladder networks are pretty damn good, his silver contact ladder networks are outstanding and I built an “ultimate” volume control using weighted series and parallel Dale resistors and ultra-low noise/thermal EMF OMRON latching instrumentation relays. AFAIK these are the best non-wiping mechanical contacts in history, they are used on the front end of my HP 34401A voltmeter which has a stable resolution of 1uV and 100nA DC. Typical Gold relay contacts are unstable below 10uA.

Leave a Reply