Gainclone Grounding and schematic comparisons...

[MrDragon1]

Hi All,

I have built a pair of LM3876 amps in one project, and am planning to build a smaller setup with a LM1876 instead for small 4" bookshelf speakers. I have seen multiple schematics for various different chip-amps and have a few questions.. I have some basic electronics background but am mostly left taught so need help at times..

I see a couple of conflicting schematics with regard to the feedback loop grounding. The NatSemi data sheet for example shows a 1K resistor in series with a 22uF non-polar capacitor then ground connected to the -Input on the IC. I have seen other asthmatics with just a 680 Ohm resistor and no capacitor. What will this change?
(for examples see here)
LM1876 Overview
DIYAudioprojects.com Beast

I know this has probably been explained several times before, but the star ground seems tricky. Is it easiest to just run a wire from every ground symbol to one point? If yes, should wires for some things be equal length?

Volume control Pots seem difficult to find balanced at lower volumes. I have tried a couple from Radio Shack (100k) and they are horrible (both in balance and sound quality) but easy to find. I tried linear 10K and it sounded very good but was as unbalanced (L/R) as the other. Lastly I gutted some dead PC speakers and settled for a tiny 10K Log pot for now. I think its an Alps but not sure. It sounds ok, but is balanced very well. I found the following and wondered if anyone had used it yet? Carbon Volume Potentiometers - Audio Taper for Volume Control - 10kohms | Diy HiFi Supply

I want to keep this as basic as I can. It will be driven primarily by a PC sound card so as long as that works I'll be happy.

Thanks all!!

[Redshift187]

Adding the capacitor to the feedback ensures any DC on the output is returned to the -ve input, reducing the DC at the output.

[tangmonster]

the difference between the 1 k and the 680ohm versus the 20kohm resistor sets the amplifier gain.

google overture.xls

it gives a simple way of changing these two and seeing the gain difference.

High gain is great but if you don't need it use a lower gain, it is easier to get more noise at the output.

in one of my setup's i used 1.5k for a much lower gain ,since the preamp gave out much more than 1V.

Do not make the gain too low, lmxxxx doesn't like gains lower than 10v/v

[AndrewT]

post1 link2 is a DC coupled non-inverting schematic for a chipamp (or any opamp).
post1 link1 is a mixed AC and DC coupled amplifier. Never build this until you understand and can work out all the consequences of this mixed coupling.

I do not recommend that a beginner ever build a DC coupled amplifier.

As a first build and more probably first dozen builds, stay with AC coupled amplifiers.

This requires a DC blocking capacitor at the input to form a high pass filter (attenuates low frequency and blocks DC).
It also requires a DC blocking cap in the feedback loop. The cap in series with the 1k or 680r reduces the low frequency gain of the amplifier and makes the DC gain of the amp = 1times (+0dB).

I do not recommend inverting topology for your early builds.

You must also fit a low pass filter to the amplifier input.
Once you know how and why your builds work/operate then you will have the information/knowledge to make decisions on whether you are ready to start adopting inverting topology or DC coupling.
Finally, you must follow the manufacturer's recommendations for all the optional components that help to ensure stability.

[MrDragon1]

My first project actually more matched this here:
Synergy - LM3875 Gainclone Amplifier (Chip Amp)
With the exception that I added a 4.7uF in series with the input resistor and 22k connected between them to ground. The end result should be blocking any DC coming in, but still only has the 680 ohm resistor to ground in the loop. If I am understanding this right, the 22uF capacitor in series with it would then just be to block any difference in voltage between the +V and -V?

I really do appreciate the help :-) I know I don't have many posts yet, but thats mostly because I try to read as much as I can before asking..

Thanks again!

[AndrewT]

post5 link fig01 still shows a DC coupled amplifier.
And worse it has a variable source resistance to upset the output offset.
Do you want my advice or not?

[Sebastiaan]

Originally Posted by AndrewT

post5 link fig01 still shows a DC coupled amplifier.
And worse it has a variable source resistance to upset the output offset.
Do you want my advice or not?

Good point. I hear some reports from DIY-ers who build amplifiers with the LME driver chips or the LM power opamp chips with attached a potmeter like above straight to the input. In some cases I heard reports about oscillation on different volume settings. This will not be discovered without scope. Personally I would never add a potmeter direct on the positive input like this. You really want to keep input impedances at the + and - arms equal.

With kind regards,
Bas

[Redshift187]

Originally Posted by AndrewT

post5 link fig01 still shows a DC coupled amplifier.

Yes, but he said he added an input filter capacitor, which means it's actually AC coupled on the input.

[MrDragon1]

Originally Posted by Sebastiaan

You really want to keep input impedances at the + and - arms equal.

With kind regards,
Bas

Just for clarity's sake, does that mean if I use a 680 ohm to ground on the -input that I should also have a 680 ohm to ground on the +input then matched to the DC filter capacitor? Does that count the +input resistor in the total between input and ground?

Sorry to not have a drawing of this, I hope I am making sense..

Thanks!!

[AndrewT]

an opamp and a differential input power amp both rely on matching currents and matching voltages at the +IN & -IN input pins.

To achieve this you must imagine yourself sitting on either of these pins looking out towards the outside electronics world. Both seats should have the same view.
If you additionally want balanced inputs then both seats should have an identical view of the world.