Why so much gain ?

A place for DIY project discussions.
User avatar
karatestu
Posts: 4790
Joined: Sun Jan 08, 2017 4:40 pm
Location: North Yorkshire
Has thanked: 1048 times
Been thanked: 616 times
Great Britain

Why so much gain ?

Unread post by karatestu »

I put this in diy as it could get technical and bespoke to individuals (manufacturers have to cover all bases).

If like me you have a small room (4.2M x 4M x 2.8M) and never play really loud then what is the point in having so much gain in the system and just throwing it away with a volume control? This could also apply to those with very sensitive speakers. Mine aren't very sensitive but still I have way more gain than I need :roll:

With modern source voltages of approx 2V and sometimes even more there is no need for all this gain. Many of us don't use active preamps (I don't) and use NVA power amps and still have too much gain. Why reduce the signal to noise ratio of the power amp by having so much attenuation before it ?

High gain brings with it more noise and distortion so why bother having more than you need when you could actually be enjoying less noise ? One way to play around with gain is to alter the amount of negative feedback by changing the feedback resistor. More negative feedback by reducing the resistor value means less gain.

There is always a catch though is there not :roll: Changing the amount of feedback can have an effect on phase shift leading to instability and oscillation. Some say you can't change gain this way by more than 10dB without redesigning the whole amp. To help it has been advised to adjust the dominant pole via the miller cap across the VAS.

Is it worth playing around with this or would the results be inaudible except the screeching as my amplifier self destructs ? I don't have a sillyscope or anything which can help set phase margins etc, it would be a shot in the dark.
DIY inspired by Richard "The Doc" Dunn RIP

User avatar
karatestu
Posts: 4790
Joined: Sun Jan 08, 2017 4:40 pm
Location: North Yorkshire
Has thanked: 1048 times
Been thanked: 616 times
Great Britain

Re: Why so much gain ?

Unread post by karatestu »

Ooops, forgot to mention input sensitivity :roll: . Obviously with lower power amp gain via increased feedback we need less attenuation from the volume control for the same volume at our ears. But then there may be an issue with having too much voltage going in to the power amp resulting in it being overdriven. So the input sensitivity of the power amp MAY need to be altered depending on the design.
DIY inspired by Richard "The Doc" Dunn RIP

Geoff.R.G
Posts: 1238
Joined: Thu Aug 11, 2016 2:46 pm
Location: Denham UK
Has thanked: 64 times
Been thanked: 248 times
Great Britain

Re: Why so much gain ?

Unread post by Geoff.R.G »

The level of gain is determined by the source output. A phono cartridge can very low, Moving Coil so obviously needs more gain than does a CD player at 2V. The amplifier needs to be able to handle any input you may wish to use. With a passive pre-amp all the gain is in the power amp so it can't be optimised for each input individually.

A couple of examples:
A tape recorder may have an output ranging from 300mV to 1.4V the former needs more gain than the latter.
A tuner may have an output of 300mV but some are lower, my Leak Troughline has a low output.

Gain and input sensitivity are very closely related; a low sensitivity, say 2V, implies less gain than a high sensitivity, say 250mV.
Yes, noise is also related to gain. The noisiest component is the first transistor in the circuit because any noise it produces is amplified by all the other gain stages.

Now the bit many people overlook; the output from any device depends on the internal gain. A tape head produces a pretty low output so to get the 300mV from the output sockets requires amplification between the tape head and the sockets. Likewise a turntable needs a phono stage. Guess what, the 2V from a CD player requires gain stages in the player. The gain stages have to be somewhere, if they are in the amplifier you can decide which manufacturer you trust to get the noise and distortion down.

The volume control is a potential divider so you won't get too much voltage to the input, at least not before the volume becomes a problem.

User avatar
karatestu
Posts: 4790
Joined: Sun Jan 08, 2017 4:40 pm
Location: North Yorkshire
Has thanked: 1048 times
Been thanked: 616 times
Great Britain

Re: Why so much gain ?

Unread post by karatestu »

Hello Geoff
Geoff.R.G wrote: Mon Aug 30, 2021 11:35 am The level of gain is determined by the source output. A phono cartridge can very low, Moving Coil so obviously needs more gain than does a CD player at 2V. The amplifier needs to be able to handle any input you may wish to use. With a passive pre-amp all the gain is in the power amp so it can't be optimised for each input individually.


A couple of examples:
A tape recorder may have an output ranging from 300mV to 1.4V the former needs more gain than the latter.
A tuner may have an output of 300mV but some are lower, my Leak Troughline has a low output.

Gain and input sensitivity are very closely related; a low sensitivity, say 2V, implies less gain than a high sensitivity, say 250mV.
Yes, noise is also related to gain. The noisiest component is the first transistor in the circuit because any noise it produces is amplified by all the other gain stages.

Now the bit many people overlook; the output from any device depends on the internal gain. A tape head produces a pretty low output so to get the 300mV from the output sockets requires amplification between the tape head and the sockets. Likewise a turntable needs a phono stage. Guess what, the 2V from a CD player requires gain stages in the player. The gain stages have to be somewhere, if they are in the amplifier you can decide which manufacturer you trust to get the noise and distortion down.

The volume control is a potential divider so you won't get too much voltage to the input, at least not before the volume becomes a problem.

Yes I realise different sources have different output voltage level. LP is usually less voltage than CD even after going through a phono amp. I use a NVA phono amp which is quite high output voltage and not too far off that from my CDP. I don't use tape, tuner or anything with really low output voltage. There is quite a difference between output level of different CD's I find.

As it stands I have way too much gain in the power amp than is really required, even allowing for a bit of headroom. The volume control (if I had one) would not be operating right at the bottom of it's range but nowhere near the top either. I think it makes sense to run the signal as "hot" as possible in to the power amp and use as little attenuation via the potential divider volume control as possible without overloading the amp in to clipping of course.

To be able to run the signal in to the power amp as hot as possible then this requires the attenuation provided by the volume control position to be very low - turned almost all the way round to 11 man :grin: If the gain in the power amp is too high (the only gain available when not using an active pre) then we just end up having to attenuate the signal at line level more and throw that voltage away.

I believe lowering gain in a power amp (all else being equal) leads to reduced levels of noise which I assume can only be good . The level of noise produced by the power amp is not altered by the level of the volume control as it is after it in the chain. If we run with the volume control producing very little attenuation and a power amp with gain JUST enough to produce the volumes we will require (and enough headroom) with all sources (not much volume in my case - small room, low listening levels) then surely we are preserving as much of the signal to noise ratio as possible ?

When you said "the gain stages have to be somewhere" I obviously totally agree. But surely it makes sense to have it before the volume control as the signal to noise ratio remains constant? Having ALL the gain in the power amp which is after the volume control means even when the volume is low the power amp still contributes the same amount of noise as it would be if the volume control was on 11.

Maybe I have got this all muddled up :shifty:

Anyway, I am intrigued by ways of reducing noise and reducing power amplifier noise by reducing gain seemed like a logical thing to do if it could be done by increasing feedback and not straying outside of phase shift margins which would mean oscillation, instability, tears and a nasty smell.

I already got rid of unnecessary gain by ditching active preamps and bypassing the output stage of my CDP. There is more to do :grin: If I had a source which outputted 5V (it is known these days) and had incredibly sensitive speakers then I would be just shoving an output buffer current amplifier on the end with no voltage gain.
DIY inspired by Richard "The Doc" Dunn RIP

User avatar
Lindsayt
Posts: 3529
Joined: Fri Jul 13, 2012 9:06 pm
Has thanked: 345 times
Been thanked: 262 times
VATICAN_CITY

Re: Why so much gain ?

Unread post by Lindsayt »

I agree. It's bonkers to have so much gain that you chop most of it back with a volume control.

It all depends whether you could get away with one less amplification stage or not. Or if you could get a better sounding solution by using a different amplification stage that has less gain.

For marketing purposes, I can see why it might be attractive to have too much gain or a volume control that's calibrated such that the system gets loud rapidly over the first few notches on the volume knob. Because then it's like a car where the throttle opens most of the way with the first two millimetres of pedal travel.

Although having said that, having too much gain is better than having too little. It's annoying when max volume is still too quiet.

Vinyl-ant
Posts: 575
Joined: Sat Jul 01, 2017 12:51 pm
Location: South yorkshire
Has thanked: 0
Been thanked: 153 times
Contact:
Great Britain

Re: Why so much gain ?

Unread post by Vinyl-ant »

The volume control does not alter the gain, it reduces the voltage that the stage sees.

The term power amp is also something of a misnomer, it just does not have the volume control and source switching in it. Thats the only difference befween a so called intergrated and power amp.

You will typically have an op amp on the output of a source which sets the output voltage, a low gain gain stage in a preamp, boosting the voltage the 'power amp' sees to 4-5v, then a power amp with a voltage gain stage and a current amplifier stage to drive speakers.. a massive oversimplification but it illustrates a possible system.

So you have a gain structure with a total of something like 4 stages. To give a total system gain of x

An alternative is to have less stages to get the same x.

Or more stages to get the same x

This gain structure is designated by the manufacturer based on expected perameters such as typical speaker sensitivity/impedance, and typical source voltage and output impedance, the bias characteristics of the devices used, the most linear operating point of a device used, the topology of the circuit, wether its differential or single ended, op amp or dircrete transistor stage, ect ect.

Richard based his gain structure on these things, his own speakers, and a typical source.

Quad did the same with their ranges, so did naim, so did musical fidelity et al.

Input sensitivity is set by the designer based on their typical source output, biasing the first amp stage to give its maximum just above the maximum expected input voltage, or its lowest distortion, or below the clipping point of a stage. Or some other reason.

They are designed around the least worst set of compromises to the designer.

There is always a reason why the gain structure is how it is.

There is no spoon
I am (barely....) in the hi-fi trade
Status: Manufacturer
Company Name bte designs
Www.bte-designs.weebly.com

DaveyTed
Posts: 22
Joined: Mon Apr 26, 2021 10:01 am
Has thanked: 16 times
Been thanked: 25 times
Great Britain

Re: Why so much gain ?

Unread post by DaveyTed »

Some years ago it was explained to me, by an amplifier designer, that a volume pot should be compared to the brakes of a car. He asked - would you drive your car with your foot permanently on the brake pedal. Even allowing for different pot characteristics he considered volume settings below about 12 o'clock to be undesirable.
I don't know if this theory is correct but to my layman's brain it made sense.
These users thanked the author DaveyTed for the post:
karatestu (Wed Sep 01, 2021 7:02 pm)

User avatar
karatestu
Posts: 4790
Joined: Sun Jan 08, 2017 4:40 pm
Location: North Yorkshire
Has thanked: 1048 times
Been thanked: 616 times
Great Britain

Re: Why so much gain ?

Unread post by karatestu »

DaveyTed wrote: Wed Sep 01, 2021 6:43 pm Some years ago it was explained to me, by an amplifier designer, that a volume pot should be compared to the brakes of a car. He asked - would you drive your car with your foot permanently on the brake pedal. Even allowing for different pot characteristics he considered volume settings below about 12 o'clock to be undesirable.
I don't know if this theory is correct but to my layman's brain it made sense.
That's a good analogy imo. I would bet most people have too much gain available to them and end up hard on the brakes. Yes, designers have to cover all bases so a potential customer doesn't end up with not enough. The average person (not nutters like us) probably think that having all that gain available is a bragging right, look how powerful my system is :roll:

Richard designed his amp circuit with the amount of gain it has to make up for using a passive pre. But even with an average source component voltage and nva cubes (which aren't very efficient) there is still too much gain for me in a small room at the levels I listen at. Lowering the gain in the power amp makes the most sense to me. It could be done (increased feedback) but probably with a few alterations here and there to keep things stable.
DIY inspired by Richard "The Doc" Dunn RIP

Geoff.R.G
Posts: 1238
Joined: Thu Aug 11, 2016 2:46 pm
Location: Denham UK
Has thanked: 64 times
Been thanked: 248 times
Great Britain

Re: Why so much gain ?

Unread post by Geoff.R.G »

Please bear in mind that I am used to devices with variable gain so some people what I have to say may not be completely applicable to hi-fi.

Gain is normally adjusted to make the levels at the fader broadly the same for each input; as an example a trained singer will need less gain than an untrained one, a guitar may need attenuation (pad). In hi-fi the manufacturer will design an active pre amp to achieve a broadly similar level at the volume control from each input device. This will be achieved by having the gain of each input set to match the common output of the device for which it is designed, in hi-fi there is no standard output level.

Professional audio kit, mixers specifically but also effects units and processors, are designed for +4db u which makes life easier.

Now, with a passive pre amp the maximum output is determined by the input level from each attached device. The gain must be sufficient to allow low level input devices to drive the power amp to a suitable listening volume but with sufficient headroom to prevent clipping from devices with a higher output (there are limits which makes the user responsible for keeping the volume in check).

The more recent trend is for higher output levels from pre amps meaning that less sensitivity is required for the power amp and thus the noise figure is lower. Purely marketing hype, to get from a moving coil cartridge to, say, 50 watts output, requires the same gain whether the power amp needs 500mV or 2V input for that output. The power amp that requires 2V will have less gain but the associated pre amp must have more.

A result of the current trend is that some people are modifying vintage power amps for lower gain (sensitivity). Great if you want to use a recent pre amp but if, like me, you are using a vintage pre amp the mod has to be reversed. As Vinyl-ant says, the amp is designed the way it is for a reason, only the designer knows the reason. It may also be worth considering that source device output levels have been rising so that what was an appropriate level of gain in 1990 or 2000 now appears excessive.

User avatar
karatestu
Posts: 4790
Joined: Sun Jan 08, 2017 4:40 pm
Location: North Yorkshire
Has thanked: 1048 times
Been thanked: 616 times
Great Britain

Re: Why so much gain ?

Unread post by karatestu »

DIY inspired by Richard "The Doc" Dunn RIP

Post Reply