Buffy the volume slayer

January 13, 2013
 by Paul McGowan

Number 5 on our list of 9 important items to take into consideration when designing a DAC output stage or, for that matter, a preamp output stage is buffering the volume control.

On many DACS, including our own, a digital volume control is used rather than an analog one. From a purists standpoint this is a much better solution than that of an analog volume control at any volume level useful for critical listening. Properly implemented a digital volume control loses no resolution above a reasonable point, while an analog volume control is always in the signal path and adds a sonic signature that isn't always welcome.

Many DAC designers insist on keeping an analog volume control in the signal path for two reasons: some DACS allow the user to have an extra set of analog inputs and secondly, the general perception amongst users is that analog volume controls are better than digital. They are not - but that's another post.

Should the DAC or preamp designer decide to use an analog volume control it is pretty important to make sure it is properly buffered. Buffering refers to impedance matching and maintaining a consistent output impedance. This is important if the output of the volume control is going to drive the interconnect cable between two devices, but more likely it's being used before the output stage and here you want to make sure the loading on it is much higher than the attenuator you're using.

A volume control, or attenuator, potentiometer, stepped ladder etc. are all variable resistors forming a divider that either lets all the music pass or eliminates some of it.

If you turn a volume control all the way up you are letting all the music through without any hinderance - save for the contacts and resistive elements it must still pass through. When you turn that same control down halfway, half of the musical signal is passing through the control and the other half is being thrown away to ground. All the way down and 100% of the signal is sent off to ground and nothing reaches the output stage.

The simplest of these devices is known as a potentiometer or "pot" for short. A pot is a resistor that has three contact points on it:

  1. Input (A)
  2. Output (W)
  3. Ground (B)

On the input we place the full signal from the DAC output. On the output we connect the input to our analog output stage. The ground is going to ground or zero (a black hole for musical signals). The output is actually a piece of smooth metal that is in contact with the resistor material and configured such that it slides up and down the length of the resistor - in fact, this output is often times referred to as the "slider" or "wiper" because it does both along the resistor.

At one extreme of its movement, the slider is closest to the input of the pot - this is where it's loudest - and at the other extreme it's physically touching the ground side - where there is no sound. In the middle you get half the volume and along the way you get everything else. You can imagine that with such a mechanical device, the quality of the metal contact and the resistive material has everything to do with the sound. A cheap pot can cost a designer $0.50 and an expensive Penny and Giles can be several hundreds of dollars - then there's everything in between.

As this slider moves up and down, changing the volume, it is at multiple different resistance settings. So, let's imagine that this pot of ours has a 50k Ohm resistive element. That means that as far as the internal DAC output is concerned, it sees a 50K load - which is very nice and easy to manage. The output of the pot, however, is seeing something very different. It sees everything from 0 Ohms (all the way off) to 25k Ohms (halfway up) to whatever the output impedance of the internal DAC is (all the way up). In other words, the output of the pot is all over the map and, perhaps most important, our analog output stage's input impedance must be significantly higher than the highest output impedance level of the pot - preferably 10 times higher so we don't have any unexpected results.

So the challenge for the designer is selecting a volume control with a high enough input impedance to make the internal DAC or source happy, yet not so high that our output stage isn't unhappy as well. It's a fine balance and one completely irrelevant should we decide to get rid of this old analog beast and stick with a modern math based digital volume control.

Tomorrow I'll cover yet another way to handle volume that doesn't involve either, known as a Gilbert Cell or in PS terms, the Gain Cell.

Subscribe to Paul's Posts

One comment on “Buffy the volume slayer”

  1. The Circuit analysis is wrong. When the pot is at full output, full clockwise rotation, the source (DAC or whatever is driving it) sees 50 Kohms in parallel with the load. If the load is low impedance, say 5 or 10 Kohms input of a power amplifier and the source can't drive it, the sound will suffer, the source of the DAC will be loaded down. At zero output, full counterclockwise rotation it sees just 50 Kohms. In between it sees the resistance between the source and the wiper in series with two impedances, the remainder of the 50 Kohm potentiometer and the load. The problem is with the way the load sees the source looking back at it. The load for the next gain stage or buffer input looks back and sees the source impedance in series with the impedance between the source and the wiper and a parallel impedance which is the remainder of the impedance between the wiper and ground. If this is inches away and the load impedance is selected by the designer then it can be dealt with having minimal effect. But if the output is not buffered meaning the potentiometer is what is called a "passive preamplifier" and the distance to the load is considerable, then the impedance looking backward can be drastically affected by the interconnecting cable because the characteristic parameters of the cable in the telegrapher's equation describing the wire are of the same order of magnitude or greater compared to the rest of the network. Not only is the so called "passive preamplfier" a very bad idea especially when it, the source, and the power amplifier are not in immediate proximity, but the price usually charged for this minimal device is usually far greater than the cost of the parts. The design is virtually non existant and the parts are few. Several jacks, a box, a switch or two and a potentiomenter. I went into the wrong business. Had I only known it would be so easy. Lincoln said you can fool some of the people all of the time 🙂

Leave a Reply

© 2022 PS Audio, Inc.

linkedin facebook pinterest youtube rss twitter instagram facebook-blank rss-blank linkedin-blank pinterest youtube twitter instagram