Non-Buffered Non-Inverted Gainclone (LM3875) fpr 47K Pot Circuit Wanted

[krishu]

Hi Folks,

sorry for asking (i tried searching the forum ).

Can anyone please post a schematic or a link to a GainClone circuit which works in non-inverted mode and uses a LM3875 and 47K pot (this is what i have here...)

Lots of schematics can be found on the www but only a few of them are made based on the experiences you folks made. Unfortunately i can not judge which circuits out there are the good ones ...

Thanks in advance

Cheers
Christian.

[theChris]

47k linear or logrithim scale pot?

the simple non-inverting circuit will work, placing the pot before the input terminal. i think there are otehr ways of doing it too.

[krishu]

Hi,

47k linear or logrithim scale pot?

both on my desk ...

Where can I find the latest circuits?

Cheers
Christian.

[GregGC]

This is one example. I'm trying a bit different values. R6=22k, R4=470K to minimize the output DC offset to minimum.

[Ricren]

Originally posted by GregGC
This is one example. I'm trying a bit different values. R6=22k, R4=470K to minimize the output DC offset to minimum.

Amazing. I was about to try this option, for exactly the same reason: lowering the offset.
Have you listen to it yet? I was going to use 220K in feedback and 6.8 to ground. I know that I should match 22K in the other input, but still, is a closer match than the circuit that is proposed to date, (22K in non inverting input and 680 ohms in inverting, right?
After all, we have been using high feedback resistors in the IGC for a long time, so the situation sould be an improvement. I've seen 1 meghom feedback used in some designs, but I feel is a little extreme.
We'll see.

Ric

[krishu]

Hi

@ GregGC,

Thank you very much. This is about what I was looking for. Should the pot be log or lin?

@ Ric

As far as I know* a opamp should "see" the same impedances on both inputs when used in inverted mode. When used in non-inverted mode this is not so important. Am I right? Is this true for power opamps, too?

Great forum.

Cheers
Christian.


*My knowledge is collected on the www ...

[theChris]

the final result should be somewhat logarithmic. either by manufacturer's design, or by compensation. a linear pot can be made to approximate a log - pot, and many cheap log pots are not really logaritmic either.

[krishu]

Thanks,

I got

> a dual 47K logarithmic pot here,
> two large heatsinks,
> a 2x20V transformer,
> lots of 1/2W metal film resistors,
> four 2.200uF electrolytic caps
> two 2.2uF Wima MKS for input
(i know that this smaller value will cost me some bass)
> two LM3875, of course
> two 0.1uF MKP caps
> some connectors
> some fast diode rectifiers

here. Hope this fits.

I will mount everything on the heatsink and a piece of MDF first.

Before I ruin my beloved speakers, I will measure DC on output, and try the gainclone on cheap speakers. Is there anything else I can do to make sure not to blow my equipment?

Thanks again,

Best regards
Christian.

[Ricren]

Cristian,

As far as I understand it, it does not matter if it's inverting or not, the same principle aplies to any op amp, power or not.

I wonder if the difference in sound that Peter Daniel hears between NI and IGC has to be more with the resistor values than the topology.
I'd really like to know if he'd like the NIGC with 220K in feedback and 6.8K to ground.
He is the only one that can carry this experiment,but I think is asking him too much.
cheers
Ric

[GregGC]

Originally posted by Ricren


Amazing. I was about to try this option, for exactly the same reason: lowering the offset.
Have you listen to it yet? I was going to use 220K in feedback and 6.8 to ground. I know that I should match 22K in the other input, but still, is a closer match than the circuit that is proposed to date, (22K in non inverting input and 680 ohms in inverting, right?
After all, we have been using high feedback resistors in the IGC for a long time, so the situation sould be an improvement. I've seen 1 meghom feedback used in some designs, but I feel is a little extreme.
We'll see.

Ric

Exactly the way I was thinking. I haven't listen to it yet.

Originally posted by krishu
Hi

@ GregGC,

Should the pot be log or lin?

Christian.

As someone already said: The pot can be linear. Actualy preferebly linear so that you get a closer match between left and right chanal, especially in stereo version. The 22k in paralel will create a sort-of-log pot out of it. If you use 100k pot instead you'll be closer to the log curve, but not such a big deal.

The parts list looks fine to me.
If you have SM resistors try this.
As Ric mentioned above, try playing with R4/R6, keeping the ratio the same see what combination will give you less offset at the output and see how it'll sound. Try 3.3k/68k or 6.8k/150k or 10k/220k, 15k/330k and tell us what you get. The goal is min DC on the output (single digit in mV) with the min res. values of R4/R6.