[Andrew_H]

Hi,
First post and what a great site !

On to my question.
Would it be possible to build the basic Cmoy with, say an OPA4132 and use 2 of the channels for the buffer, the other two for the stereo amp ?
Does a schematic exists ?

Many thanks,
Andrew.

 

[Bjornboy81]

using one as an amplifier and one as a buffer won't get you anything because the opamp output current is still limitied to XXmA. One thing you could do to get higher output current (assuming that is what you want) is to parallel two of them per channel like this: http://www.intersil.com/data/an/an1111.pdf (figure 1).

Welcome to Head-fi BTW! Watch out...it's addicting

 

[Andrew_H]

Thanks for the info Bjornboy81,

I`ve put together a basic Cmoy using an NE5532 on a breadboard, and although it`s a poor bipolar opamp, the audio is OK (ish).
Your right, I am seeking higher output to drive low impedance Headphones.
Would I be right thinking I could use a second paralleled NE5532 in the same circuit ?
Also, do bipolar opamps not like an input cap ? I`ve fitted everything from quality film to ceramic and electrolytic (0.47uF - 1uF) on the input to this amp and it almost completely destroys the audio.

Thanks again, and yes I`m `bitten`.

 

[OtterMarc]

Andrew,
The NE5532 is an old staple in the audio trade. Much of our music from the 70's 80's etc was mixed on that op amp or similar ones. It isn't up to the standard head-fiers appreciate today.

Having a bipolar input means that bias currents must be taken into account.

Here are two links to get you started in your understanding.

Working with Cranky Op-Amps

Op Amp Bias Currents

The book that made it crystal clear to me (after re-reading it a couple times - haha) is "Electronic Circuit Guidebook Vol 3 - Op Amps, by Joseph Carr Jr, Prompt Publications, ISBN 0-7906-1131-7.

It cost me about $30 here in Canada at Active Electronics a few years ago, a company related to Future Electronics. It's not up to date re the current breed of op amps we like, but the basics are all there and I found the circuits to bring the understanding are well done.

If you still struggle, keep asking, and I'm sure other head-fiers will steer you to other resources.

You have two choices - experiment (struggle along) till you understand, or switch to jfet input op amps.

For me, playing with circuits on breadboards made all the difference. There are limits to what breadboards will do with your circuit, though.

Also, to find resources here re buffering and such, try searching on "stacked", stacking, stack. It's what they call it when they piggyback opamps or buffers to gain more current (and drop distortion). A double bonus.

Hope that helps.

 

[ericj]

Quote:

Originally Posted by Bjornboy81 /img/forum/go_quote.gif using one as an amplifier and one as a buffer won't get you anything because the opamp output current is still limitied to XXmA. One thing you could do to get higher output current (assuming that is what you want) is to parallel two of them per channel like this: http://www.intersil.com/data/an/an1111.pdf (figure 1). Welcome to Head-fi BTW! Watch out...it's addicting

Yes, pretty much identical to Figure 9 in the opa2604 datasheet, the original inspiration for the A47 amp.

http://www.ti.com/lit/gpn/opa2604

And sijosae does have an A47 on quad-opamp layout somewhere. I once breadboarded it with an opa404 and it worked ok.

 

[Seaside]

As eric said, that would be A47. Original A47 at headwize by Apheared (HeadWize - Project: Apheared's Project Scrapbook by Michael Shelton) employees two dual OPs, each dual OP as one channel and buffer. Sijosae modified that as what you asked. 1 dual OP for amp, other dual OP for buffer. So, his A47 would work like a Cmoy if you remove a OP that does buffer. You can see Sijosae's A47 design at here.

 

[Andrew_H]

Many thanks for the replies.
Certainly plenty to digest.

I think I`ll keep to the single opamp design for now, and carry on tinkering with the breadboard.
As a first `improvement` albeit mainly current output increase, I`ll stick in a NJM4556AD and have a play.
I want to experiment and learn the basics before going deeper. We`ve all got to start somewhere.
I`ll, no doubt, have a few more questions on the way shortly.

Thanks again,
Andrew.

 

[Bjornboy81]

baby steps, baby steps.....

 

[joneeboi]

I just built the A47 per headwize library thread-spec, though my resistors probably aren't too tightly matched, and tried the JRC4556/NJM4556 opamps in both positions. I kept the buffer and amp within the same chip, but that's beside the point. With the JRC4556 opamps, I came out with 33.5mV DC offset per channel, though the OPA2134, OPA2107 and such produced lowly 2mV-5mV offsets. What are the consensus opamps in the A47 anyway? OPA2227 seems to be popular, though not particularly well suited to audio. OPA2134 also is pretty well-liked in the A47, but I can't really gather much from the few threads on head-fi and headwize.

 

[ericj]

Quote:

Originally Posted by joneeboi /img/forum/go_quote.gif I just built the A47 per headwize library thread-spec, though my resistors probably aren't too tightly matched, and tried the JRC4556/NJM4556 opamps in both positions. I kept the buffer and amp within the same chip, but that's beside the point. With the JRC4556 opamps, I came out with 33.5mV DC offset per channel, though the OPA2134, OPA2107 and such produced lowly 2mV-5mV offsets. What are the consensus opamps in the A47 anyway? OPA2227 seems to be popular, though not particularly well suited to audio. OPA2134 also is pretty well-liked in the A47, but I can't really gather much from the few threads on head-fi and headwize.

The 4556 is a bipolar-input opamp and needs carefully selected feedback resistors to avoid DC offset.

When i built my A47, with carefully calculated resistors, it had less than 4mv of DC offset with the 4556 in the gain stage.

If you build chip-per-stage rather than chip-per-channel, you can stick anything you want in the buffer stage without much concern for offset.

It also allows you to mix chips - for example putting a relatively weak (but nice sounding) 2107 in the gain stage with a 4556 in the output stage.

of course, in that configuration, you could also remove the resistors that bridge together the outputs of both opamps and just take the output from the 4556, which has much higher current output.

 

[~n00beR]

Here is a link to a korean diy site that Sijosae has posted his A47 layout for a quad opamp.

A47 Layout for Quad Op Amp

Hope this helps

n00b

 

[Andrew_H]

EricJ,

Would your choice of resistors differ greatly from the final diagram (bottom of page) in Tangents `Working with Cranky Opamps` ?

I`d appreciate a little guidance in setting this up in a Cmoy type config.

Strangely, my NE5532 gave a DC Offset of <4mV when using the standard Cmoy resistor values but lacked much of its bass. Adding an input cap destroyed the audio almost completely. I`ll continue to play with the resistor values via the various calculators, although I`m sure there is a known-good set of component values whicj can`t really be improved upon.

Cheers,
Andrew.

 

[Pars]

Quote:

Originally Posted by ericj /img/forum/go_quote.gif If you build chip-per-stage rather than chip-per-channel, you can stick anything you want in the buffer stage without much concern for offset. It also allows you to mix chips - for example putting a relatively weak (but nice sounding) 2107 in the gain stage with a 4556 in the output stage. of course, in that configuration, you could also remove the resistors that bridge together the outputs of both opamps and just take the output from the 4556, which has much higher current output.

Yes, I agree. I highly recommend if you go dual opamps, to do the chip-per-stage rather than the original chip-per-channel as it allows for much higher flexibility. You can even pull the 2nd opamp and listen to it as a Cmoy if you wish... and I do like the OPA2107.