Sound Investment?

[Gordon]

When used correctly they either attenuate the signal by 10 dB or they don't. If they don't, either:
(a) They're not fit for purpose which seems unlikely, or
(b) The circuit into which they're being inserted doesn't have the expected "line" properties.

Typically the amp should have an input impedance around 10Kohms or more and the source very low. The attenuator will be two resistors of the right values to get the signal down by around 3 times - that is what 10dB means. These resistors' values will assume something about the electrical environment.

Can't you just turn down the volume control on the pre -amp?

Bear in mind that 10 dB really isn't very much, a 1 dB change in speaker output power is on the threshold of human discernibility in carefully controlled experiments. I can adjust the input gain on my amplifier and just tried reducing it by 6 dB as an experiment. I can tell, but the effect is pretty subtle.

If your system has an amplifier with far too much gain (do you have exceptionally sensitive speakers or something?) and really is only operable at "very low" settings you probably need a lot more than 10 dB of attenuation.

Simon B is quite right. 10dB isn't a lot, about a third. Your amp must be very sensitive!!

It would help to know what the source driving the amp is - it should be around 200-500 mVolts perhaps more but not usually. Yours sounds as if it's a lot less - what make and model is it? If you could look in the amp's handbook and get out the sensitivity figure - so many millivolts for full power output - and the input impedance we can begin to understand the cause of this problem. Similarly get out the hand book of the driving/source device - is it a pre-amp without a volume control? If a pre amp what happens for other inputs?

If it is a CD player connected directly this will likely be giving out 2 Volts and this could be part of the problem! With a sensitive amp and 2 Volts you need to attenuate by at least 20dB [ten times] down to 200 mVolts and even then it's possible the amp needs less than that for full power output so maybe as much as 30dB to get the volume control in a sensible place.

[Gordon]

Scratch some of my last post - I've just looked up your amp on the web and found its handbook. It's integrated so it has various inputs the output power for which are all set by the volume control. So why are you getting overload?? -you must have a very high level source connected to one of the low level inputs!! Which input are you using that gets the problem or do they all do it? I can't believe that they all do! Which of the several inputs are you using the attenuator on?

I'm begining to see what your problem is: if you have normally sensitive speakers then they will need only a few tens of watts to get a decent sound level so the 270 mVolts line input level needed for full 150 watts out is getting on for more than 10 times too much and so you only need about 30-50 mVolts input!!! If one of your inputs is a CD player without an output level control - as I said before this is probably giving out 2 volts!! I'm not surprised you have level problems on that input. As I said before what happens on the other inputs eg tuner where the input signal may be lower? Even there the sensitivity is such that even 100 mVolts might cause problems.

The amp has an overload margin of 26dB on the line inputs which means that the signal input can be 20 times higher than rated before the input stages [not the power amp part] limit and they distort. You must be near that limit or just below it otherwise the sound you do get would be distorted.

[umslopogaas]

Phew! Many thanks Gordon, I will digest these comments and reply later, I have to go out this evening. I should emphasise that this is not really a problem, in that the system actually works fine. Its simply that the difference between "not loud enough" and "too loud" is rather small, and there is an awful lot of adjustment left in the "dangerously too loud!" sector. I just have to be careful with the volume control, but so far I've had the setup for several years and have managed to avoid blowing anything up.

I use an FM tuner in the daytime for R3/4 and a turntable in the evening to play my vinyl. I also occasionally play CDs. All three inputs have a similar "too soft" - "too loud" range on the output volume control. At the moment I only have the attenuators on the FM tuner, I havent confirmed if their effect is equally small on phono and CD.

[PhilipT]

Anyone else got any other low price audio solutions which have audibly improved the performance of their systems?

In my experience, which has mostly been of helping out people with ordinary rather than audiophile hi-fi, the biggest single improvement comes from decently-heavy speaker cables. I leave spending hundreds to other people (my own system runs on 18-year-old Linn K20, and I see no reason to change it); the trouble comes when people think they only need 'a piece of wire', and they have that flex from the old standard lamp, and ..

[johnb]

You mean connecting the standard lamp to your hifi doesn't work?

But seriously, I was amazed a few years ago when I had a pair of Chord Chorus interconnects on loan. There was a real difference in the sound. Mind you, it was a little too bright, to my ears at least.

[Gordon]

If this thing is happening on all the inputs it looks like the power scaling of the input sensitivity to the output power is the root of the problem and you may have to attentuate all of them by at least 30dB if not more - and that's about 30 times in voltage. The amp will want the volume control to be at maximum for the sensitivity figures quoted in the specification. All the inputs are scaled to provide 150 watts which is clearly too much for the speakers [you need really insensitive ones for this amp but appreciate your argument for a big 'un]. One solution could be provided by a professional audio engineer who can get inside the amp and reduce the power section input level whilst leaving the pre-amp alone. All you need is to shift the volume control point much further around its rotation.

I'm sure I don't need me to remind you that you need to be very careful when pulling and inserting audio plugs that the transient doesn't cause speaker damage. Don't take any plugs etc out or in without at least turning the volume down to zero.

[Gordon]

Quite right PhilipT: given the amount of power needed to drive speakers and the fact that they are nominally only 8 ohms impedance then Ohm's law tells us what current is needed. An Amplifier is a voltage source and is required to deliver whatever current the load [the speaker] demands. While the amp may be able to source it, can the cable deliver it withour introducing its own resistance into the effective load the amp sees?

A 50 watt amp loaded with 8 ohms needs to deliver 2.5 amps [assuming the load is purely resistive which it never is and so the current can be much greater still in practice]. The voltage needed is 20 volts. To make the speaker cable itself an insignificant part of the load [and thereby reducing the power reaching the real load] we need a cable with [at least] less than 0.8 resistance. A 10 metre lead would need to be less than 0.08 ohms per metre. Even this reduces the power reaching the speaker by 10% or about 1dB - only 18 volts get to the speaker. In the process the source impedance seen by the speaker is now the lead's resistance, not the amplifier source which is much lower so the damping factor of the amp is diluted thus also reducing its control over the speaker. Some speakers are around 4 ohms or less and so need twice the current.

Transient peaks are the main reason why some poorly fed speakers sound dull and lacking sparkling dynamics. These peaks demand short bursts of high current which develop voltages in the cables rather than getting it all to the speaker. You can never have too little resistance in the speaker leads so make them as fat as possible!! Biwiring helps and 2.5mm square ring main cable isn't bad and is much cheaper than so called speaker leads but may not be so pretty. I don't believe all this oxygen free copper stuff.

Ideally the speaker lead should be an 8 ohm transmission line with conjugate match to the speaker impedance so that it presents the amp with a pure 8 ohms resistive load but which of the HiFi cable or speaker companies supplies one? It has to be matched to the individual speaker and would have to be the right length too and kept that way. Too difficult.

[PhilipT]

Ideally the speaker lead should be an 8 ohm transmission line ...

I've never understood that line of argument. I don't see how a matched impedance transmission line can help when the length isn't at least a meaningful fraction of the wavelength. Even at frequencies as low as 30Hz the wavelength is still a few thousand miles ...

[umslopogaas]

For info, for those who have kindly been advising me about amplifier settings:

The speakers are bookshelf Elac BS244, Nominal Power/Handling 80/120 Watts, Nominal Impedance 4 Ohms, Suitable For Amplifiers 4-8 Ohms (according to the label on the back). So yes, if I got some 15 Ohm speakers that would open up more of the volume control range and lessen the danger of speaker overload. However, I've only just spent quite a lot of money on those Elacs, so I think I'll stay with them and take extra care not to turn up the volume too far.

[Gordon]

I've never understood that line of argument. I don't see how a matched impedance transmission line can help when the length isn't at least a meaningful fraction of the wavelength. Even at frequencies as low as 30Hz the wavelength is still a few thousand miles ...

You are right of course which is why I said "ideally"!! The posting was long enough already so I didn't expand further on the "too difficult". The speaker lead and load has to be considered as lumped components [even electrostatics] but in principle a voltage source amplifier shouldn't mind as long as it can supply the current in the right phase - therein lies a stability issue when feedback is used [esp. with electrostatics]. My point was that the leads are part of the source as far as the speaker is concerned and dilutes the amp's control. It is a compromise for a sinlge amp - multi-amping inside the speaker will help of course but now we are getting a bit expensive as well as drifting off topic.

Now we know that U's speakers are nominally 4 ohm then current demand doubles [no problem at all for the MF amp he has] and so lead resistance is more important.