A-B Switch Box For Amplifier Comparisons

I have finally been able to add 'Project X', thanks to Phil Allison.  Somehow, 'X' just seemed appropriate   Have fun.

Main Index

Projects Index

Articles Index

This is probably going to be a very controversial device.  Its purpose is to prove people wrong and that is very confronting.  If you don't wish to have your cherished beliefs about amplifiers and audio generally challenged then do not build or use this unit.

Many of you will know about the ABX system for doing audio comparisons.  No doubt it is a very fine piece of design but out of reach for the average person.  Some years ago I felt that a much simpler device would at least allow me to do comparisons on power amplifiers while the music played in a similar way to ABX.  This device was the outcome.  After using it for a few seconds my attitude to audio listening tests changed forever.  If you are game for a challenge then try it yourself.  It will cost you under $50 to build.  It may be the best or worst fifty bucks you ever spent depending on your attitude.

The idea is that the listener, you, sits in the 'hot seat' and concentrates on familiar music on your favourite speakers in your own lounge room with the ability swap power amps over without moving more than one finger.  If there is even a small change in timbre or definition it should be instantly audible.  The stereo image and any other factor can be checked precisely since you can sit totally still during the switchover.  Well at least that was what was expected to happen.

The design is very simple and I lashed mine up in a couple of hours and put it to use immediately.  A couple of good quality relays, some hefty terminals and banana plugs and a long wire finishing in a hand held push button are the ingredients.  The circuit is as shown and is self-explanatory.  What is does however, is staggering.

Providing the two amps to be compared are of high quality (why would you be interested in anything else?) and of course fault free, the gains are carefully matched and have similar bandwidth the device permits instant and seamless switching of the amplifier outputs to the speakers at the push of the button in the listeners hand.

Because the relays switch in a couple of milliseconds there is no audible interruption.  This surprised me at first and I tried my sine wave generator set to 100 Hz and pushed the switch.  About half the time I could hear a faint 'tick' sound but this was not audible on musical programme.

Figure 1 - The A-B Switch Box Schematic

A description of the circuit is hardly needed (but I'll give a brief one anyway).  The DC voltage must be matched to the relays, and 12V is suggested as the most practical.  The relays should be high quality, high current types, and DPDT relays can be used with the contacts paralleled for lowest resistance.  Gold flashed contacts are desirable, but standard silver contacts should be quite adequate unless there is severe atmospheric pollution in your area.

The speaker grounds for the two amps were connected together at the box and the speaker grounds were coupled as well.  Use short thick leads for connection to the amps.  It is possible this might cause an earth loop hum with some combinations.

Warning!   Never attempt to use this switching device with amplifiers that have a BTL (bridge-tied-load) output arrangement.  If there is a warning that neither speaker terminal may be earthed then this describes your amplifier.  If you connect it to the switching unit you will almost certainly cause severe damage to the amplifier.  Both amplifiers under test must be conventional amps that have the -ve speaker terminal connected to earth (ground).

The switch (marked 'Push-On / Push-Off') is the remote switch, and needs a lead that is long enough to reach the listening position.  There is no real limit to the length, even with light duty figure-8 'speaker' lead, but in excess of 10 metres or so may cause some voltage loss.  The LED is optional, and may be omitted.  If fitted, be very careful that the LED cannot be seen from the hot seat, as it may dim slightly when the relays are energised, thus giving a visual clue - this you don't need!

1. Build the box as shown in Figure 1 and place the two amplifiers close together.  If one is on top of the other, then place some cardboard in between to prevent metallic contact and possible hum loops.  (Make sure that you don't obstruct any air vents.)
   
   
2. Make 'Y' leads to connect the left and right input signals to both amplifiers.
   
   
3. Use an audio generator or test CD to set gains to read equally ±1% on a digital multimeter at about 400 Hz.  Leave the speakers unconnected while you do this.
   
   
4. Place the pushbutton on the 'hot seat' or at the place where you would normally sit to listen to stereo.
   
   
5. Check operation of the box by switching with no music and listen for the click of the relays as they changeover.  There may be a faint click from the speakers too, due to small DC offsets.  Trim these out if possible - the idea is to get a seamless change with no audible cues.
   
   
6. Put your favourite track on the CD player or turntable if you prefer vinyl.
   
   
7. Listen, and gently push the switch whenever you want to change amps.
   
   
8. Go and check why the box is not working.
   
   
9. If step 8 is not needed check your wiring for correct phase and levels again.
   
   
10. When you have everything right try it on your friends and family.  They will likely think you have gone mad and/ or are playing tricks on them.  The most likely comment is simply "The switch is not working."
    
    
11. Sit quietly and contemplate what you have just found, and all its implications.
    
    
12. Don't blame me, I did warn you!

Notes:
When comparing amps of different power ratings, stay within the capacity of the smaller unit.  Valve (vacuum tube) and transistor amps may be compared as long as the valve model has a damping factor greater than 25.  Valve amps with a low damping factor (output impedance of 2Ω or more) will sound different (note: different does not mean 'better'!).  Any test involving a valve amp is at your own risk!  Be careful, as many valve amps don't like an open-circuit output.  Amps with subsonic filters may be distinguishable from those without on some material.

Matching the gain may be difficult if the amps do not have level controls.  Solder a 10kΩ multiturn trim pot to the back of each RCA plug on the amp with more level to set the gain.

You must use a push-on, push-off hand held switch.  Never use one that needs to be held down to make the relays change over.  The switch must be one that does not change 'feel' from one position to the other (some feel slightly different depending on the latching mechanism).

After the initial surprise (shock) wears off, try the old 'stop and restart A-B test' method again.  See what happens!

This is a contributed article, and the author (Phil Allison) has made this information available for the sake of audio.

I do suggest that if you want to test this technique that Phil's instructions must be followed to the letter - even the smallest variation in level can invalidate any A-B test.  Comparisons between valve and transistorised amplifiers are likely to show differences, only partly due to the higher output impedance of a valve amp, and extra care is needed to balance the levels with this combination.

It is also important that there are no visual cues that might alert you that one amp or the other is in operation.  To be safe, place a screen of some sort between you and the amplifiers and switch box.  As Phil has stated, if amplifiers are of different power ratings, make sure that neither amp clips (distorts) at any point during the tests.  This will be immediately audible, and is not a valid test for an amplifier's sound quality (since it has none when clipping!)

You might find it hard to get a push-on/ push-off switch that has no difference in feel between states.  If this is the case, you can use the circuit shown in Project 166.This allows a momentary switch to be used, and there will be absolutely no difference in feel either way.  If you decide to include an indicator LED (which is a good idea so you can verify the switching action), it must have a switch in series so it can be turned off for testing.  Before you start, you might want to get someone else to press the button a (random) number of times to make sure that there's no in-built bias.  This is recommended regardless of the switch used.

If you don't want to be confronted by this switch box, build one anyway.  It will allow you to make comparisons between amplifiers that are otherwise impossible to do with accuracy or repeatability.  You might find that there are audible differences, or you might not.  Either way, it gives you the ability to know (rather than assume or imagine) that one amplifier is different from another.

This tester can also be used to change speakers for comparisons.  These are the most inaccurate of all electronic components, and there are some interesting traps that can affect the result, leading to a very wrong conclusion.  I have discussed this elsewhere, but most people will be unaware that some things are decidedly counter-intuitive.  If one speaker (A) has a notch (aka 'suck-out') at some frequency, if you listen to it for 30 seconds or so, then switch over to another speaker (B) that has (comparatively) flat response, speaker B will sound wrong!

It will seem to the listener that speaker B has a peak at the frequency where the notch was situated in speaker A.  This can be verified either with hardware (a notch filter that you can switch in and out of circuit), or you can use Audacity or similar to insert a notch.  This is human hearing (ear-brain interface) at work, and it's surprisingly easy to be fooled by the way our auditory systems work.  This is always active, and (perhaps surprisingly) it doesn't matter much if the test is blind or sighted.  The test should be blind to prevent visual cues that can lead to other issues, but it will still work even when you know which is which.  Avoiding the experimenter-expectancy effect is still important though.

There's a discussion of this issue at Harbeth, with four videos.  The first two discuss this issue in some depth.  Never underestimate the apparent 'problems' you might hear that are caused by our hearing mechanism.  Despite the (fallacious) claims that only listening can reveal the 'truly best' audio, it becomes very obvious that test equipment is your friend.  Measurements have no inbuilt prejudices - provided the measurement is set up properly of course.

If comparing speakers, the amplifier goes to the 'speaker' terminals, and the speakers are wired to the 'amp 1' and 'amp 2' terminals.  Unless the speaker systems have identical sensitivity (dB/W/m) they will sound different, with the louder speaker almost invariably sounding better (even if it's not).

Main Index

Projects Index

Articles Index