Stereo volume controller with input selector and VFD display

Somehow, despite the large number of reviews I've written, I've hardly ever reviewed devices related to audio equipment in one way or another. Of course, I did review a power supply for a power amplifier, but in my opinion that's only tangentially related. So I've decided to take a look at amplifiers, DACs and other audio devices, starting with a volume control.
I chose this volume control mostly for aesthetic reasons, since functionally it's very simple, so today's review won't be very long.


As you may have gathered from the introduction, I'm going to build something resembling an amplifier, probably with a DAC, though that doesn't really matter here. I used to work with this kind of equipment a lot, but years have passed and I've forgotten some things, while a lot of new things have come along. So I'll be refreshing my memory in some areas and learning as I go in others. There may be mistakes and inaccuracies, so please forgive me in advance.

Audio equipment came up in passing in this review, where I showed a power supply for a power amplifier. This PSU will probably feature again, most likely as a test subject for comparing a switching power supply with a conventional one. But that's a topic for future reviews; for now, let's get back to today's subject: the volume control.

Of course, these days you can adjust the volume not only by changing the electrical signal path, but also in software, right at the source. Personally, though, I'm not too keen on that approach and prefer the "classic" solution: an analog volume control.

First, it's worth noting that volume controls can be linear or logarithmic, and can also have a loudness compensation feature. I don't think there's much point in discussing these in detail, since it's mostly a matter of taste, but here's a very brief explanation:

1. Linear or logarithmic.
A linear control changes the division ratio in direct proportion to the angle of rotation of the shaft.
A logarithmic control (or, more precisely, an inverse-logarithmic one) is better suited to human hearing: adjustment is very gradual at the beginning and becomes more abrupt toward the end. The human ear is better at distinguishing changes in the volume of quiet sounds, so adjustment is gradual at the low end. At high volumes, differences are less noticeable, so the adjustment can be more coarse.

There are three main characteristics:
A (B in the international designation) — linear; resistance changes linearly with the angle of rotation. These resistors are useful, for example, in power supply voltage adjustment circuits.
B (C in the international designation) — logarithmic; resistance changes rapidly at first, then more gradually as it approaches the midpoint.
C (A in the international designation) — inverse-logarithmic; resistance changes gradually at first, then more rapidly as it approaches the midpoint. These resistors are commonly used in volume controls.
An additional type is W, available only in the international designation. It has an S-shaped control characteristic, combining logarithmic and inverse-logarithmic curves. To be honest, I don't know where these are used.
If you're interested, you can read more here.
By the way, I've come across imported potentiometers whose characteristic letter matched the one used in our designation. For example, here's a modern imported potentiometer with a linear characteristic and the letter A in its designation.


2. Loudness compensation.
At low volume levels, the human ear hears the midrange better, but is less sensitive to bass and treble. That's why some controls add a deliberate frequency response correction at the very beginning of the adjustment range. Loudness compensation is usually switchable, since not everyone likes it, and switching it off lets you listen to the original sound. The simplest form of loudness compensation is a small capacitor between the resistor's signal input and wiper. More "advanced" potentiometers have one or more taps, allowing the correction to be adjusted more precisely.


To help explain this, researchers plotted families of curves showing the sensitivity of the human ear — averaged graphs of sensitivity versus frequency for audible sound waves.

The figure below shows these graphs, known as equal-loudness contours, which have been adopted as an international standard.


A circuit using a standard potentiometer to provide loudness compensation.


And a circuit using a special potentiometer.


In my case, I could mostly have used a standard potentiometer. Below is a photo of some basic potentiometers: the one on the left is a bit more expensive, the one on the right is simpler, but they're both the same thing — potentiometers. High-quality potentiometers are made by Alps, and they aren't cheap.


But an even higher-quality option is a stepped attenuator made from a set of switched resistors. In practice, this is a multi-step attenuator. Its advantage is that you can set any adjustment curve you like, but more importantly, it lets you match the channels more precisely.
There are also ordinary variable resistors with a ratchet mechanism, but don't confuse them with this—they're completely different; essentially, they just provide an "emulation".

Stepped attenuators are most often used in high-end equipment. For example, the first one I came across was in the popular Odyssey 010 amplifier. By the way, with a little patience, you can make a similar attenuator yourself using a multi-position switch and carefully selected resistors.


Or even like this: basically just a switch with a bunch of resistors.


Replacing the switch with relays makes for a neater solution, and also allows remote control. To keep things simple, the resistors are controlled using a binary code. Adjusting the resistor values also lets you set a logarithmic response.
Switching the divider ratio using fixed resistors gives you a wide adjustment range with a relatively simple setup: 1 relay gives 2 levels, 2 relays give 4 levels, and 3 relays give 8 levels.
The photo below shows an attenuator with 256 adjustment steps. It's controlled by a dedicated IC, ADC0804, which converts the analog signal from a variable resistor into a binary code. The variable resistor simply changes a DC voltage and isn't connected to the signal path at all.
You need to use special signal relays, not power relays, because power relays can't provide reliable contact at low voltages and currents.
Another advantage of this kind of attenuator is that it's easy to make it multi-channel: just add another relay board in parallel.


You can see pairs of resistors underneath the board, next to each relay. Initially, I was thinking of buying this exact attenuator, but then I changed my mind. I'll explain why later.


The well-known Nikitin attenuator is built on roughly the same principle. Its advantage is that the input and output impedances remain constant, which improves performance and reduces its effect on the rest of the circuit.


As mentioned above, stepped attenuators allow for remote control, but if you prefer, you can also buy a regular attenuator with a "motor" controlled by a dedicated controller. That's essentially how it works: the variable resistor's shaft can be turned by hand or from a remote. In the latter case, a small geared motor does the job, and the adjustment knob turns along with it. Add an LED to indicate its position, and it looks pretty impressive.


Anyway, I spent a while trying to decide which attenuator to use, then stumbled across a rather interesting option. What caught my eye most was the type of display, but more on that later.
The package includes:
1. Attenuator board
2. Control board with display
3. IR remote control
4. Blue filter
5. Power and output wires
6. Ribbon cable for connecting the boards, 280 mm long
7. Adjustment knob.

You can also buy these separately:
1. 12 V, 5 W power transformer - $2.22
2. Load control board - $3.7
3. Upgrade to gold-plated RCA connectors - $1.47

I bought the "basic" version because I already have a transformer, I can make the relay board myself, and I'm not convinced those "gold-plated" connectors are worth an extra buck and a half. I was worried the display might get broken in transit, but it arrived safely.


The bundle of odds and ends is nothing special: a blue filter, a cheap knob, and a couple of wires.
I won't peel the protective paper off the filter yet, since I still need to install it in the case and don't want to scratch it.


The remote looks like it came from some AOC TV. It's reasonably convenient, but has a glossy case. It looks pretty good, though it could have fewer buttons since most of them aren't needed.
You can switch inputs using the Input 1-2-3-4 button, or the Bright buttons in either direction.


The main board, with the relays, attenuator, and power supply for the whole unit.



I'm not sure what they meant by "gold-plated" connectors, which cost extra, but I received the ones shown in the photo. The board can switch signals from four sources, with all the inputs brought out to one large terminal block.


The soldering is a bit rough in places, but I liked the overall build quality: neat, with mounting holes and clear labeling.


The board runs on 12 V AC, although it also worked without any problems from 9 V. Some of the capacitors are marked Elna, though I don't think that matters here. Besides, the Chinese are quite the tricksters, and you can't always trust markings like these.
It also looks like there's a voltage multiplier on the board, since the display needs considerably more than 12–15 V. But there's nothing wrong with a multiplier; it would be worse if the designer had used a switching voltage converter.


There are also four voltage regulators here. Two (78L05 and 79L05) power the control circuit, one 7805 powers the relays, and the other powers the control board.


And here's the regulator with the four-channel switcher.


Signal level is controlled by a dedicated chip, the CS3310 from Cirrus Logic. I didn't list the regulator's specifications at the beginning of the review, but since they actually depend on this chip, it's more accurate to give them here. Though accuracy is relative, since these are the specs for the original chip, and I can't say which one is installed here.


I mentioned stepped signal controls above for a reason. This regulator is stepped too. In the block diagram, the attenuator (i.e. divider) is highlighted in red, and the variable-gain amplifier in green.
Unlike a regular potentiometer, this regulator can operate in two modes: attenuation (-95.5 dB to 0) and gain (0 to 31.5 dB). The attenuator handles attenuation, while the variable-gain amplifier handles gain.


The regulator's circuit is extremely simple, which is why the performance of the unit is determined by the chip's specifications. That said, some parameters can still be ruined by poor PCB routing.
The regulator is originally a two-channel device, but according to the datasheet it can be cascaded and used in multichannel systems; you just need one or more additional chips like this.


The board has a connector for the control panel, as well as a chip with its markings rubbed off that I can't identify.


As mentioned above, the board can control an additional load. To do this, it has terminals for connecting a relay. These terminals provide 5 V when the regulator is switched on and in operating mode; switching is on the negative side.
This output can be used to control power to a power amplifier.


1. CS3310 regulator chip
2. ULN2003 transistor array for controlling the relays; it also controls the additional output.
3. TX2 signal relays TX2, rated for 5 V. I think I have some of the same relays at home, only genuine ones; maybe I'll compare them later.
4. A chip I can't identify. Why the markings were rubbed off is a mystery.


The underside of the board is empty; most of the copper pours are used as shielding against interference.


Since the regulator chip has digital control, the kit includes a control and display board.


It can be controlled either with the encoder or with a remote. For that, there's an IR receiver on the board, and for obvious reasons the light filter needs to cover it too.


And this is one of the reasons I chose this particular regulator model: a VFD display, or, as we call it, a vacuum fluorescent display (VFD).
That's what makes this board a bit of a warm, tube-like piece of gear, since a VFD is a real vacuum tube—though it has nothing to do with audio. The display really is a very basic one; similar displays are used in calculators and other devices where 9 characters are enough.


To be honest, I really like things like this, and I wouldn't mind having displays like these as alternatives to standard 1602, 2004, and similar displays. But they usually cost a lot—though they do look nice.



The control circuitry and other components are on the back of the board, which has the same design as the regulator board. One thing to note: the board isn't quite flat; it bows slightly away from the front panel.


The regulator's controller and display driver.


The board has contacts for connecting an external keypad, as well as jumper pads.
1. Green — keypad: mute, input selection, and volume control. Unlike the encoder, it has a mute function but no power button.
2. Red — full operating mode (attenuator + amplifier) or attenuator only.
3. Yellow — disables the settings memory function.


1. Control microcontroller — 12C5A60S2
2. Display driver — PT6312BLQ
3. EEPROM, presumably for storing settings.
4. Photodetector soldering. At first I thought it was a mess, but later I found out it only looks like this from underneath; the soldering on top is excellent.



To test the regulator, I connected a 9 V power transformer, joined the boards with a ribbon cable, and... that's it, ready to power up.


With the flash on and no filter, it's impossible to make out anything on the display, even though I adjusted the image in Photoshop here.


Without the flash, or with some kind of filter, it's much easier to see; the display itself is quite bright.


The product page shows examples of how this regulator can be used—or, more specifically, how to design a front panel around it. Some examples clearly use a different, noticeably longer filter, though.


For now, I made do with a piece of green filter I found at home. Below, I'll describe the operating modes.
1. Off: only the decimal point in the rightmost digit is lit.
2. A short press on the encoder switches the regulator to its main operating mode. The display shows “Hello,” which then disappears. As I mentioned above, the board has an output for switching an additional load; power appears there as soon as the encoder is pressed. When power is applied to the board, it briefly clicks a relay; all relays are off in standby mode. To put the board into standby, hold down the encoder for about two seconds.
3. The display shows the active channel number and the signal attenuation/amplification level.
4. Briefly shorting the Mute contacts displays dashes in the level field. Shorting them again restores the sound.
5, 6. The minimum is -96 dB and the maximum is +31.5 dB. The datasheet specified a range of -95.5 to +31.5 dB.

And here's the catch in the last item: the full adjustment range has 256 steps, and since the encoder has 20 positions per revolution, it takes almost 13 full turns to go from minimum to maximum. I do like smooth adjustment, but there are limits... In my opinion, 30 steps would be enough; if you want finer adjustment, then 60–65, but 256...


Disabling the built-in amplifier helps a little, with two benefits:
1. The amplifier introduces less distortion into the signal (presumably).
2. Instead of 256 steps, there will be “only” 192, or 9.5 encoder turns.

You can make it even more convenient by replacing the encoder with a 24-position version; then it takes only 8 turns.

Removing jumper P5 disables the built-in amplifier, and the maximum shown on the display becomes 00.0 instead of 31.5. The photo also shows different inputs selected: 1 and 4. Inputs are switched with a short press on the encoder.
The settings are saved, but after power is completely disconnected, the regulator starts in the mode it was in before it was properly switched off. There is no separate memory for each input; the volume level is the same for all inputs. If you solder the memory lock jumper, the first input will be selected and the signal level set to -46.0 dB every time you power it on.


Since the display is always on, power consumption barely changes between operating modes: 187 mA in standby and 236 mA when operating. These figures are for AC input; power is about 1.7 and 2.2, respectively.


Naturally, I did a few basic tests, but for the most part I ran into the limitations of my test equipment, especially the oscilloscope. For a volume control, the key parameters are usually adjustment linearity, distortion, and channel separation, but I honestly don't know how to test all of that with just one signal generator and a basic oscilloscope. With an input voltage of 2.65 V and a level of -70 dB, the voltmeter reads about 1 mV at the output.

For testing, I used a fully analog signal generator covering 10 Hz–100 kHz and a DS203 oscilloscope.
First, I checked the waveform at 10 Hz.
1. Input signal
2. Output signal at 0 dB.
3. Output signal at +8.5 dB
4. Clipping starts at +9.0 dB, but this depends on the input signal's amplitude.
5. Level: -45 dB
6. Level: -30 dB


Frequency: 20 kHz.
1. Input signal
2. Output signal at 0 dB.
3. Output signal at +12 dB
4. Since the input signal's amplitude is lower here, clipping starts at +12.5 dB. As the gain is increased further, the signal gradually turns into a square wave.
5. Level: -45 dB
6. Level: -30 dB


The highest frequency my generator can produce is 100 kHz, so I decided to test it at that frequency too.
1. Input signal
2. Output signal at 0 dB.
3. Output signal at +11.5 dB
4. Output signal at 12.5 dB. At 12.0 dB, clipping was barely noticeable, so I chose 12.5 dB to make it easier to see.


The power amplifiers aren't ready yet, and the DAC hasn't even arrived, so I tried it briefly with this amplifier. It works fine, at least on the only working channel :)
This is the amplifier I'm going to modify. I know it's clearly no Odyssey, but it's what I have. Although, considering that all I'll really be keeping is the case, and possibly the transformer and heatsink, I don't think that matters much. The Odyssey does have a much more substantial look and construction, though.


Video version of the review


For now, I can briefly say that everything works, and I have no complaints on that front. The volume adjusts, the remote works, the display shows all the necessary information, and I haven't noticed any audio distortion. One thing worth noting is the lack of switching converters for powering the display. The display is still multiplexed, but that's a limitation of the display itself.
There is one drawback, though: the signal adjustment is too gradual, so I'll most likely replace the encoder and disable the built-in amplifier.
I'd also like separate volume adjustment for each input, but that's more of a wish-list item, since this feature isn't usually used.

Overall build quality is good, and I haven't found any obvious flaws. Unfortunately, I can't verify whether the volume control chip is genuine.


That's all for now. As usual, I'm looking forward to your questions, suggestions, comments, and so on. I hope you found the review useful.