LM3886 GC with LM4562 tone control

[Mooly]

Originally Posted by AndrewT

Use Dr Cherry tapping point for the NFB back to the -IN input.
That adds an extra pole of attenuation for RF injected into the speaker terminal.

Yes you could do. I think you mean taking the feedback from the other end of the 10 ohm in the network.

Tbh it's an area of design I have never really looked into in great detail.

[Mooly]

Originally Posted by blu_glo

'gis a clue .... who/whats Dr. Cherry???

Dr/ Professor ? Edward Cherry. I couldn't just turn up anything easily readable for you.

[mxriddler]

Originally Posted by Mooly

The resistor and cap to ground is a different thing altogether really. Many amps need those to ensure stability with no load (speakers) connected. Its a must have.

The output coil isn't always normally needed in practice. Its there to prevent small capacitive loads from causing instability... such as the long speaker wires blu_glo mentions. The 10 ohm across the coil damps any "ringing" caused when squarewave testing the amp. The coil isn't critical in value, just make sure its air spaced (not wound on a magnetic former) and of reasonably thick wire.

The latest diagram you have posted is a "single rail" configuration. Yes As long as you appreciate the differences there.

Thanks. I'll be sure to include the resistor and cap in my plans.

I am aware that the last schematic is a single rail. Only looking at it for the optional components, still planning the dual supply.

[mxriddler]

Originally Posted by blu_glo

Input impedance of an integrated amplifier is usually 10-47 K ohms; its so it doesn't unduly load equipment connected to it such as CD players etc.

See attached; it seems all the opamps you specify are unity gain stable so hould be ok, so it's up to you which to use or put a socket and choose afterwards! Make R1 = R2 in attached circuit the other resistor not really needed in audio (short "+ input" to 0V) but if you insist, when R1 = R2, R3= 0.5 x R1 (or R2!)

Suggested values R1 and R2 = 22K R3 if insisted apon, 10K is quite close enough.

Thanks for this. I hope to actually learn what each of these elements do while designing/building this amp.

While doing some reading on buffers, I noticed this is an inverting buffer. Is there a benefit over a non-inverting buffer? Also, R1 sets the input impedance, correct? The manual for my firewire interface that will feed the amp says it has an output impedance of 51k, is 22k an ok value to match my source? I do want the amp to be compatible with other sources as well, but for now the main source will be the firewire interface.

I also found some schematics with unity buffers on the input. Is there any benefit or negative to using a unity buffer here? If i understand properly, the unity buffer provides very high input impedance, would that be a better fit here?

Thanks again.

[blu_glo]

Originally Posted by mxriddler

Thanks for this. I hope to actually learn what each of these elements do while designing/building this amp.

While doing some reading on buffers, I noticed this is an inverting buffer. Is there a benefit over a non-inverting buffer? Also, R1 sets the input impedance, correct? The manual for my firewire interface that will feed the amp says it has an output impedance of 51k, is 22k an ok value to match my source? I do want the amp to be compatible with other sources as well, but for now the main source will be the firewire interface.

I also found some schematics with unity buffers on the input. Is there any benefit or negative to using a unity buffer here? If i understand properly, the unity buffer provides very high input impedance, would that be a better fit here?

Thanks again.

There is supposed to be a slight reduction in distortion using the inverting buffer due to it's virtual earth mode of operating. However that wasn't the primary reason I chose it! I actually chose it because the tone control is also inverting so I restored the "absolute phase" for you. It is not important, but some swear they can hear a difference and I would build my circuit this way, if only for "tidyness".

Yes R1 sets the input impedance. I'm surprised the firewire interface says output impedance is 51K? Does it mean its SUITED for input impedances of 51K?? Can you please check and copy the exact wording? Most designers worth their salt make output impedances as low as possible for driving reasonably long cable runs etc.....

You can freely use unity buffer if you're not fussed about the signal inversion from the tone control mentioned above. In all non-inverting buffers make sure the + input has a resistor to 0V normally chosen to set your desired input impedance.....

[mxriddler]

Originally Posted by blu_glo

There is supposed to be a slight reduction in distortion using the inverting buffer due to it's virtual earth mode of operating. However that wasn't the primary reason I chose it! I actually chose it because the tone control is also inverting so I restored the "absolute phase" for you. It is not important, but some swear they can hear a difference and I would build my circuit this way, if only for "tidyness".

Yes R1 sets the input impedance. I'm surprised the firewire interface says output impedance is 51K? Does it mean its SUITED for input impedances of 51K?? Can you please check and copy the exact wording? Most designers worth their salt make output impedances as low as possible for driving reasonably long cable runs etc.....

You can freely use unity buffer if you're not fussed about the signal inversion from the tone control mentioned above. In all non-inverting buffers make sure the + input has a resistor to 0V normally chosen to set your desired input impedance.....

That makes perfect sense. I'm all for being tidy and will go with your advice of using the inverting buffer. Thank you.

I realized that I misspoke and said 51k and not 51ohms for the output impedance. So I realize now there shouldn't be any issue there!

I'm putting together a full schematic and will post in the next day or so hopefully.

[blu_glo]

Originally Posted by mxriddler

That makes perfect sense. I'm all for being tidy and will go with your advice of using the inverting buffer. Thank you.

I realized that I misspoke and said 51k and not 51ohms for the output impedance. So I realize now there shouldn't be any issue there!

I'm putting together a full schematic and will post in the next day or so hopefully.

Oh, thanks, I thought madness had consumed all if manufacturers had started issuing equipment with such high output impedances!!!
Yes, 50 ohms or thereabouts is much more what I'd expect.

Would like to see how you get on with that schematic, btw what schematic package are you using?

[mxriddler]

Originally Posted by blu_glo

Oh, thanks, I thought madness had consumed all if manufacturers had started issuing equipment with such high output impedances!!!
Yes, 50 ohms or thereabouts is much more what I'd expect.

Would like to see how you get on with that schematic, btw what schematic package are you using?

I'm just laying it out in adobe illustrator (I'm a graphic designer after all)

Here's a question, would it be better to put the volume pot before the buffer or right before the lm3886 (after tone)? If i put the pot at the beginning, how does that play with the inverting buffer resistors? Also, if pot is at the beginning, should I keep the buffer in line when the tone is bypassed, or would it be better to bypass the buffer and tone altogether?

[blu_glo]

Originally Posted by mxriddler

I'm just laying it out in adobe illustrator (I'm a graphic designer after all)

Here's a question, would it be better to put the volume pot before the buffer or right before the lm3886 (after tone)? If i put the pot at the beginning, how does that play with the inverting buffer resistors? Also, if pot is at the beginning, should I keep the buffer in line when the tone is bypassed, or would it be better to bypass the buffer and tone altogether?

buffer-tone-volume-power amp, in that order. make sure suitable dc blocking caps on input and output of volume control.
you would want to bypass buffer and tone and if you make volume control 22K then input impedance still stays the same

[blu_glo]

are you doing switched inputs too?