Simple yet versatile relay-controlled A/B switch box

Author: Dmitry Nizhegorodov ( My other projects and articles

1.   Introduction, or why an article on A/B Switch Box

An A/B switch box (a.k.a. comparator) with a remote control, a simple device that can instantly flip input and output signals seems like a very handy listening/evaluation tool, yet one can not expect to see it at a typical listening session. Why? Casual switching between two compared channels can be very helpful yet it is free from complexity and controversy of double-blind (ABX) setups. When done correctly, and anyone can learn to do it correctly, simple A/B comparison can reveal subtlety that can never be caught without a switch box. This is so because switching is practically instantaneous.

This meant to be a proof-of-concept prototype version, yet it survived many years of use. The switch box can be assembled easily and can be a pleasant weekend project.

On the outside, 6 RCA jacks are mounted in the low-power section of the box, and two rows of 3-way binding posts occupy the rest. The switch button is on a very long cord. I used a spare phone cord.

2.   Under the hood

The bottom view reveals 2 relays, the blue one is low-current miniature DPDT and the acrylic one is high-current 4PDT. Four shunt potentiometers (1Mom each) are also shown. These are optional.


The blue cylinder is an electrolytic capacitor, also optional, wired in prallel to the relays.

The parts list includes 6 RCA jacks, 12 binding posts, one low-signal DPDT relay (that is, having a coil and 2 groups of switching contacts), one 4PDT relay (that is, having a coil and 4 groups of switching contacts) a suitable toggle switch, long cord and wire.

The budget is from zero (assuming you can find all the parts in your "junkbox"), to ~ $30 (budget or ptototype version, includes two good-quality relays and average quality sockets) to .. who knows? Sky is the limit, as there are very expensive RCA sockets and binding posts out ther.

Relays are a crucial part of this project, yet it is very iasy to find excellent relays.

The signal relay must be small and preferably one designed to switch audio/video signals. Sealed relays with gold-clad contacts are preferred ("Ag contacts, Au clad"). Typical Omron 12V relays suitable for this project: Mouser 653-G5A-234P-DC12, or DigiKey Z106-ND. These are just examples. Even, RadioShack or (Bay Area local supplier) have sufficient selection of suitable relays.

Another very good option is reed relays. Reed relay contacts are completely sealed and often made of rhodium. Switching speed is very fast (~ 1 mS). Mouser 528-172-7 is a typical quality examples. Because it is a SPDT unit, you'll need two of them. Another example - DigiKey 306-1050-ND. Solid state relays should not be used for this project.

Coil power supply. Most relays have DC coil and are powered from 6, 9, 12, 18, 24V. Among these, 12VDC relays seem cheaper and are plentiful. A suitable wall-pluggable (aka "wall-wart") power DC supply providing 80..100mA should be sufficient to power both.

The output relay must hold significant current. It should be able to switch 2-3 times more AC than what goes into your speakers. As an example, a 10W amplifier sends 1.5A into 4 ohms, so a 3..5A relay is OK. A 400W amp sends 10A into 4 ohm, hence 20A is a minimum for that.

Typical Omron 12V relays suitable as high-current relay: Digikey Z2279-ND.

Life expectancy of relays - how many switches the relays can withstand: expect around 1000000 switches... That is 2000 switches per week for 10 years...

3.   Schematic

The schematic may not be easy to grasp at first, but keep in mind that it is nothing but 3 pairs of grounded RCA jacks and 6 pairs of floating binding posts switched with two relays.

Note regards diodes: Certain models of relays already contain diodes inside, some not. The diodes are needed to shunt unwanted reverse-polarity spikes that relay coils (being faithful inductors) generate at current disconnect. One diode can serve both relays with no problem.

4.   Rules for succesful comparison

There is one rule of high importance:

The Main Rule of Succesful ABing
Loudness and ballnce in channel A must be exactly replicated in channel B.

Failure to follow this rule precisely leads to false observations.

The best way to make sure the levels match is to use an SPL meter. In most situations, the best signal for that is pink noise; in some rare cases it can be other continious signal.

By the way, after adusting SPL levels with pink noise spend some time switching the channels while listening to pink noise. In many cases it reveals more differences than musical material would.

5.   Basic Applications

The switching box has surprisingly many applications. Here are the most common ones.

5.1   Source to Amp A or Amp B to Load

this is the most frequently used configuration. "source" can be a CD/DVD/record player or a preamp of some kind; "Amplifier" is typically a power amplifier. Amplifiers (stereo) should have built-in volume controls, since it is unrealistic to expect the same gain. Amplifiers without volume controls can be compared using add-on inline attenuators, see the "Advanced Switching" section.

One subtlety is in listening to amplifiers that are in phase. If it is known (from the specs, dscussion forums, etc) that one of the amps preserves phase and another flips, reversal of one amps's output is preferable for best comparison. On the first picture below the amps are in phase, on the second picture amp B is reversed.

ab-box-saal.gif ab-box-saarl.gif

Note that this is possible because the A/B box switches both + an -, and that is why there are 4 groups of contacts (see schematic). It comes very handy indeed in many situations.

Another subtlety is: how to know whether an amp preserves phase or reverses? The best is to use a scope with XY mode. Confect the X channel to the amp's input fed from source, and Y channel to the amp's output feeding the load (speaker). Now, if the trace on the scope crosses the 1st and the 3rd quadrants, the amp preserves the phase. If it crosses the other two quadrants, the amp is reversing.

If a scope with XY mode is not available, then it is not easy to tell anything regards one amplifier, yet it is possible to tell whether two are in phase or out of phase. for that, wire them in-phase and adjust levels. Play mono first on one amp and then on another and observe image between the speakers.

Then reconnect the left speaker, so that it is always on amp A, as shown on the right.

Now, play mono again and switch back and forth. In position A the image remains always in the center. If the image in position B stays in the center, the amps are in phase. If the image disappears, the amps are out of phase and the reversing setup is better be used.

Why it is necessary to phase-align the amps and why and how that matters? First, if a multi-way system is used, it is a very strong requirement; phase must be the same, otherwise switching will ineviably cause an apparent difference in sound due to 180 shift phase interaction between the switched channel and the rest.

In full-range setups in-phase switching is still preferable, as absolute phase may be audible, yet the danger of phase msiamtch is small.

Fine print: One complication arises when single ended, no-NFB amplifiers are compared. These develop significant even harmonics and hence produce non-symmetrical signals. Speakers may react to that with audible differences! Note that phase alignment in multi-way systems is still more dominant, yet when comparing wide-band SETs, aligning the sign of harmonics is more important than absolute phase.

5.2   Source to preamp A or preamp B to Amp to Load

This is the configuration most have in mind, vaguely, when thinking of comparing preamps. Here we show only the basic version; we assume that the gain stages have internal volume controls and that teh source has 2 pairs of RCA output jacks.

Refer to the advanced section in [2] for more flexible setups.





5.3   Source/preamp A or Source/preamp B to Amp to Load

While this configuration is also the one most have in mind, it has limitation in its basic form shown here, in that the source boxes either have the same output level or contain internal volume regulators. Refer to the advanced section for more flexible setups.

Still, in this basic form it allows to comparable modern CD or SACD or DVD players, as most produce the same voltage. This is, perhaps, the most surprising setup especially to those (say, on the digital source asylum) who claim apparent and "dramatic" differences in various CD, DVD or SACD players out there, and a useful reality check for those who mod or upgrade parts of stock players and report immense changes. The mods may indeed change the sound in the right direction, but many such mods result in changes that are inaudible, and the drastic results reported may be nothing but illusion. One can find that differences that were apparent in interrupted playbacks completely disappear. To "hear" them back again, it is often sufficient to listen in position A, then stop, switch to position B, wait as much as it takes to manually switch CD player cables, disks, etc, and hear the same fragment again.

Sources of non-equal loudness, in case of no volume controls, need a setup described in the advanced section of this article. Avoid using digital volume controls, unless, of course, that's the point of experiment.

"Sources" need not be physically separate and can in fact stand for two different stereo preamps playing from the same player; "sources" can also stand for different output channels of the same preamp which differ in some way (say, use different output caps).

"Sources" can also stand for turntables. Two cartridges can be switched and compared this way as long as they are the same make/brand or are very close in voltage (+-1dB). As long as two turntables use the same cartridge models, this opens up possibilities to compare and tweak any of those infinite parameters (VTA, VTF, azimuth etc) that alter the sound in an objective way (the "o" word slipped). And if you have two tables that sound the same in A/B tests, then you can have fun comparing various pressings, releases, reissues, etc of your favorite LP recording you have gathered in your collection.

To finish this section, a tip on how to time-align sources. The best way to sync-start two same-model CD players is by pressing the play buttons simultaneously. Different models may produce different startup delay, and the best is to play one, pause it at specific fragment, play the other one and resume the first one just in time. Same technique can be used to time align a turntable and a CD player. Time-aligning two turntables is the most difficult task. Bear in mind, though, that up to 1 sec of skew can be easily tolerated.

5.4   Source to Amp to Load A or Load B

This is the basic setup for comparing similar loudspeakers or selecting drivers from a batch for some criteria; since no attenuation is shown, this has limited value for comparing dissimilar speakers or drivers; refer to the advanced section for that.

On the left is bi-polar variant, on the right is single-ended (plus-only) switch.

ab-box-sall.gif ab-box-sall-se.gif

It can be pointed out that there is no functional demand to switch both + and - of the loads; it would suffice to switch just + cables and leave all - cables connected together as shown on the right picture. Indeed, it is so, yet when all cables can be plugged into 3-way binding posts with ease (meaning all cables terminate with lugs, pins or bananas, it may be less cumbersome to connect the components the way shown on the left picture (consider what to do if all speaker cables terminate with bananas). Hence the bi-wired setup preferred as the basic setup.

Another advantage of bi-polar switching is in possibilities for reversals, see the next example and various examples in the advanced section.

5.5   Source to Amp to Load or Reversed-polarity Load

This is a variant of the above, yet only one pair of loads is used here. This setup is very interesting and opens up possibilities for interesting experiments. If you amplification path contains single-ended tube stages, especially no-NFB single-ended output stage, you may find that one position sounds better. This is so because many drivers are also single-ended, and this way interact with single-ended gear differently depending on polarity. You may hear no, tiny or considerable difference, depending on physical properties of the equipment and depending on power headroom for the amplifier and the speakers during tests. The effect, if detectable at all, will be more apparent when both the amplifier and the speakers are about to enter overdrive region. Electric guitar players know well about cranked-up character of both amps and driver heads. Same effect, althoug much qless pronounsed, can contribute to the character of a home system.

if the amplification path does not contain single-ended output stages (or otherwise does not produce significant even-order harmonics), the difference is much harder if not impossible to detect even if the load is stressed hard (close to overdrive). This is so because with symmetrical drive the pressure patterns in the qir will be unchanged. Yet this offers a possibility to detect (or not) the absolute phase of the recording. Indeed, some transients may sound more natural (more convincing) one way and not the other way around - if the absolute phase is detectible.

Aside from these, the reversing-polarity setup is very useful for bi-amped and multi-channel systems. It is the best way to hear phase interaction between channels. For instance, flipping woofers or flipping mid-ranges or flipping tweeters independently from the rest of the system. If you hear no difference switching while playing music tracks, try to play pink noise and switch. Often this reveals more than what music does.

6.   Part two has more!

There is more to say - the part two, contains additional wiring arrangements.

Author: Dmitry Nizhegorodov ( My other projects and articles