Using a speaker switch

[Martin Bush]

Hi all


I have a speaker switch box coming soon which allows you to connect two set of speakers to your amp and switch between them, and even use both at once.

Is it safe to switch between speakers while the amp is on and playing? or should I switch off before switching? Also, are there any problems with using both speakers at once?

[Analogue man]

The key here is the speaker impedance which must not fall below the minimum specified for the amplifier, two 8 ohm speakers wired in parallel equals 4 ohms which may be ok but 2 sets of 4 ohm speakers will equal 2 ohms which is too low for most amplifiers. Some amplifiers won't mind switching between speakers but others will so switching off before you change over is a wise move.

[Ambientnoise]

Without knowing a lot more, do not switch “live” or use both sets of speakers at once. In general, transistor amps will not like too low an impedance (both speakers connected) and valve amps will not like no speakers connected even momentarily ( as will happen whilst switching).

[simpsons]

Martin

Many receivers do have this option available on the front via a dedicated switch and very useful it can be.

However, the amplifiers are designed for the interruption and there are specifications for the loudspeaker impedance that can be handled.

It maybe your amplifier is fine but better those "who know" have the make and model of the amplifier in question.

Chris

[Martin Bush]

The switch in question is a COMMTEL A098J although there are identical units online with different branding. I don't know what the circuit is inside the box though.

I don't need two sets of speakers on at once really, but it would be nice to be able to use the two pairs I have - one pair is currently being used as a stand for the other pair I expect I'd only have one set on at once.

[Station X]

What amp are you using?

[Martin Bush]

It's an Arcam, so a transistor amp.

I don't plan to do much more than have one pair of speakers on or the other. But something tells me I need to find out whether switching between the two while the amp is running is safe.

I am thinking of a situation where I may have a piece of music playing and want to see out of curiosity how it sounds on the other set of speakers - temptation would be to flick over and not pause the music, turn the amp off etc.

[Ted Kendall]

QED switch boxes, and others for all I know, connect the speakers in series when both are selected.

[Martin Bush]

Thanks all.

I think the best approach is to switch speakers with the amp off so will go with that. I will see if I can work out whether the two sets are in parallel or series when I get the box.

[Tractorfan]

Hi,
I have an old Leak Delta transistor amp that has a button to mute the speakers. If your amp has a similar control, you could maybe use that when changing from one set of speakers to t'other?
I assume that the Leak switches dummy loads in when the speakers are switched out.
Hope this helps.
Cheers! Pete

[Trigon.]

Originally Posted by Tractorfan

Hi,
I have an old Leak Delta transistor amp that has a button to mute the speakers....
I assume that the Leak switches dummy loads in when the speakers are switched out.

The Delta 30 and 70 both mute by disconnection, no dummy loads. The 70 has the additional facility to switch in 'remote' speakers, simply in parallel with the main ones.



Cheers

[Station X]

Originally Posted by Martin Bush

Thanks all.

I think the best approach is to switch speakers with the amp off so will go with that. I will see if I can work out whether the two sets are in parallel or series when I get the box.

Just use your meter on Ohms range to check the resistance at the input of the switching box. I know that resistance isn't the same as impedance, but it won't matter in this case.

[GrimJosef]

Originally Posted by Martin Bush

... I am thinking of a situation where I may have a piece of music playing and want to see out of curiosity how it sounds on the other set of speakers - temptation would be to flick over and not pause the music ...

Doing this will tell you something, but maybe not the thing you really want to know.

The problem is the speaker sensitivity. If one pair is more sensitive than the other then the music volume will vary when you switch between them. Our impression of music is surprisingly sensitive to its volume. Consciously we might stuggle to identify a volume change of 1dB, but subconsciously it can make a difference.

In general listeners judge louder music to sound 'better', at least for the short periods typical of many listening tests. This gives rise to two quite well-known behaviours.

The first is of hi-fi salesmen who are keen to sell you item A over item B. They will demonstrate them both to you but will make sure that the music is slightly louder for item A than for item B. You will find yourself choosing item A.

The second is of people who have a new hi-fi setup but somehow aren't quite happy with it. They will turn the volume up, trying to make it sound 'better'. This will work to begin with, but the effect will gradually wear off as whatever is actually wrong makes its presence felt again. So they will turn the volume up again. And again. The first time you visit them you will wonder why they're playing their music so very loud, but you will be too polite to mention it.

Long story short - if you want your speaker switch to let you really compare the quality of two sets of speakers then you also need to make sure you match the speaker volumes every time you switch between them. It isn't trivial to do this - it really requires a stable microphone and a decision to be made about the frequency at which you will equalise the volumes.

Cheers,

GJ

[Martin Bush]

Cheers all for the help. I will follow the guidance above when the switch arrives.

For anyone interested, the amp is an Arcam Alpha 3. No mute function that I can see.

thanks also to @GrimJosef - I only saw your post after I published mine. Very interesting.

[Lucien Nunes]

With a transistor amp, at moderate drive levels it's really not going to matter what the switch does to the circuit as you switch it, you shouldn't hurt the amp. I've plugged and unplugged an awful lot of speakers (hundreds or thousands) while running without damaging anything (admittedly professional amps, not domestic.) I would be wary of glitching the speaker connections when running at full welly as some amps might not be as robustly protected against impulses of voice-coil back-EMF from a switch arc as they should.

Obviously this does not apply to a valve amp unless some permanent load is applied to the secondary, as open-circuiting the output with signal present can seriously stress the output transformer primary, valves and holders with overvoltage.

Looking at that switch, and agreeing with the above comment that many speaker switches put the two sets in series on A+B, I would not be surprised to find that outputs A and B and the amplifier input are all hard-wired in series. The switch would be wired to short out whichever speakers are not in use, i.e. the A+B position is actually the centre-off position of the switch.

The switch would be wired like that to prevent the load impedance applied to the amplifier falling below the impedance of either set of speakers alone. This would avoid anything getting cooked but will reduce the maximum power available. More importantly, connecting dissimilar speakers in series can upset the response of both of them, because the variation of impedance with frequency of one speaker will affect the signal seen by the other. E.g. the impedance peak at the main LF resonance of speaker A will correspondingly attenuate that frequency for speaker B, and vice versa. And GJ's point about matching amplitudes for A/B comparisons is well made.

[wd40addict]

I can't remember who it was, but some Hi-Fi company was caught doing demos at a show where they had a CD with the same track repeated each time at slightly higher level. That way it helped the good>better>best demonstration of their various products...

[kalee20]

Originally Posted by Lucien Nunes

Obviously this does not apply to a valve amp unless some permanent load is applied to the secondary, as open-circuiting the output with signal present can seriously stress the output transformer primary, valves and holders with overvoltage.

I've been playing with a small push-pull valve amp (<0.5W) using Russian rod pentodes which a friend made and for which I've recently made a custom output transformer - the amplifier is seriously good BTW - and as one of the tests I wound up the input with no load connected while 'scoping the output anodes.

The amplifier runs from 100V, and I didn't see any voltage excursions beyond 220V.

With negative feedback applied, the circuit was very tame - disconnecting the load made under 5% difference in anode AC voltages.

Though, I'd concede that with a single-ended amplifier, high voltages could indeed be produced, the output transformer would act like a car ignition coil, or a line output transformer, as the valve goes rapidly into cutoff with a strong signal to its grid.

[Lucien Nunes]

I agree that a refined amplifier working under normal conditions is not likely to exhibit excess voltage excursions due to a disconnected load. I was playing safe for the benefit of those many users of valve amps who routinely apply extreme working conditions to their output stages. E.g. some kinds of 'music' could drive what is supposed to be a low-bias AB1 into grid current, charging the coupling capacitors enough for both valves to get momentarily cut off.

[Radio Wrangler]

In general, do not disconnect speakers, whether by pulling plugs, external switches or even by switches built into amplifiers while the amplifier is on.

As Lucien says, the break of a circuit with current flowing into an inductive load will create a spark at the switch and send a voltage transient backwards into the amplifier.

Many amplifiers are built with some degree of protection against this, many amplifiers are not. Those catching diodes cost pennies at the prices manufacturers pay, so in consumer amps, there is an incentive to not bother. Some boutique amplifiers eschew them on grounds that they will do imagined damage to the sound and reduce the niceness of the adjectives which reviewers will use. Even in amplifiers with some protection, they often haven't used fast enough diodes to catch all such transients reliably.

So there is a risk. Even with a fairly well-protected amplifier, the risk isn't quite driven down to zero. Lucien works with amplifiers designed for a stage environment where speaker connections get mucked about with regularly. Their protection has to be excellent they need things to be reliable.

So, even with a domestic hifi amp with a speaker switch on the front panel, don't change the switch with the amp on. Assume that the protection from speaker circuit transients won't always save you. It's statistical. You'll probably get away with it, but that one exception in a large number of occasions will eventually leave you with a dead amplifier.

Have a look through the forum and you'll find lots of transistor amps with failed output transistors and some damage further back into them. You'll also find that these repairs are difficult and don't always go well. You'll find that getting the right transistors can be difficult.

So, those blown amplifiers.... did they fall, or were they pushed?

This area is one of the major risk factors. A lot of people get away with it, but then not every time. Voltage transient overstress can do cumulative damage leading to eventual failure.

"How lucky do you feel?" as Mr Eastwood asked.

David

[Silicon]

Some designers would have a switching arrangement that shorted the unused speakers to reduce the possibility of them vibrating in sympathy with the powered speakers.

Some switches are 'make before break' while others are 'break before make'.

Check out your new speaker switch carefully.