LM3876 MUTE as volume control

Tcheko said:
Hello,

I was wondering it the mute pin could be of any use for controling volume. The documentation shows a nice ramp from 0dB/0.5mA to -100dB/0.01mA. The ramp is not really linear but a tweaked log pot may do the trick.

Any insightful though about this?

Czk.
The mute function is intended as mute and my gut feeling says that this is not a good idea. :no:
 
Dxvideo said:
No...
Because the op-amp becomes unstable under 10 V/V gain. Just a question;
Why do you need a volume control except "normal" ways? :)

Ordinary ways are not fun. ;)

I started electronic hobby one month ago and I am still in the learning curve. I am used to look at odd ways to do things. That's why. :)

Ordinary ways seems to me a bit odd too. Dividing the voltage with a stepped attenuator or potentiometer means for low level a bad SNR.

Would it be more efficient for adjusting volume to change the gain ?

LM3876 is not designed to do such thing. Is it possible to design such volume control? I guess it is possible but there is certainly some tradeoff.
 
Hi,
if you adopt the inverting topology, you can alter the gain by inserting a pot in the feedback loop to replace the upper and lower resistors.
BUT,
you must ensure the amp remains stable at low gains. How would you do this?
If the minimum gain is 10 and the maximum is 50 then that is a range of just 14db, not much of a volume control. It will however allow adjustment for high/low sensitivity to suit different source voltages.

There is one more problem with gain adjustment in the NFB loop. If the pot goes momentarily open circuit the gain goes to max, then the speaker cone/VC goes to heaven.

Cambridge used NFB as volume control in some of their amps. But I believe they had a reliability problem.
 
peranders said:
It's quite uncommon to change the gain because the amp will also get more unstable at lower gains and in performance it's a waste to have the amp stable at 0 dB gain. If this had been a good idea it would have been more common, I'm sure.

He. Sure. Windows was not a brillant idea (remember old good 3.0 days?) and we see it everywhere nowdays...

Looking futher about this idea, I am just talking about VCA or variable-gain amplifier.

The idea behind this is to not alter the audio signal path but the amplification gain. Lowering the voltage of the audio signal reduces the SNR which is bad. It is better to work with a 2VRMS instead of 150mV... But I am maybe wrong about this.

If this kind of volume control is not common, it is maybe due to the tradeoff involved with this solution, distortion, linearity, component matching etc...

A divider is far easier to build. I will go for that with my gainclone.

This is my first electronic device and so far, I am quite surprised to have succeded. No hum, no shhhhh and 11mV at output :)

Thanks for your advice.
 
AndrewT said:
fit a shorting plug to the input RCA and re-measure the output noise.
Aim for <0.2mVac

No noise at output : I can't ear anything from the speaker even if my ear is close.

11mVdc offset at output. And I can't mesure under 1mVac with my current equipment.

How can I remove the 11mVdc offset? Do this come from the slight difference between +V and -V rails?
 
Tcheko said:
How can I remove the 11mVdc offset? Do this come from the slight difference between +V and -V rails?
no,
it comes from the input offset current which flows from both the inverting and the non-inverting inputs.
These currents pass through the resistances attached to the inputs and generate a voltage.
The difference in the input offset voltages appears at the output multiplied by the gain of the amplifier.

If you need to reduce the output offset, then you can manipulate the input terminal source resistances to get that balanced input offset voltage. But as the chip changes temperature the input offset currents will probably change giving rise to temperature induced drift of output offset.

Output offset of 11mVdc is pretty good. How far does it drift with temperature?
 
AndrewT said:
fit a shorting plug to the input RCA and re-measure the output noise.
Aim for <0.2mVac


AndrewT said:
no,
it comes from the input offset current which flows from both the inverting and the non-inverting inputs.
These currents pass through the resistances attached to the inputs and generate a voltage.
The difference in the input offset voltages appears at the output multiplied by the gain of the amplifier.

If you need to reduce the output offset, then you can manipulate the input terminal source resistances to get that balanced input offset voltage. But as the chip changes temperature the input offset currents will probably change giving rise to temperature induced drift of output offset.

Output offset of 11mVdc is pretty good. How far does it drift with temperature?

I didn't mesured it long enough to have an idea. Last time I did, I had less than 8mVdc just after start (cold chip).

here is a picture of the proto : http://tcheko.hostarea.org/P1010184.JPG

Air wired, centered ground star, short wiring. The ground star is hidden by the cinch. Banana plugs are output. We can see the NFB resistor mounted directly on the LM. The MX22 shows the Vac of the output. The chip is driven by a sin wave from a LFG.

I will do a burn to the see how it drift.
 
measures

I did some measures after two hours running :

9mVdc and 0.09mVac at output. I was not aware that my metrix MX22 could be that precise.

I tried the amp against my other amps (all crap). Clear winner is the LM3876. The sound is crystal, vocals, instruments are well separated. I feel that common consumer products worth not their price... The amp costed around 80€. Bargain for the improvement. I still need to build a case.

As a first diy project, I feel rewarded. A must. Gainclone rocks!