Stacking Opamps??

[fmak]

Is it valid to stack 2 or more 'matched' opamps ro reduce noise, just like muti dacs?

 

[ghg]

http://www.linear.com/pdf/1028fa.pdf ( p. 15 )

 

[peranders]

Before you do anything like this, estimate how low noise you have INCLUDING signal source noise.

You will soon discover that this approach is rather pointless these days (not in 70's and 80's)!

Can you tell me what you are up to?

Some text from my homepage (scroll down a bit).
http://home.swipnet.se/~w-50719/hifi/qsxm2/schema_qsxm2.html#Anchor-The-49575

The text is about transistor pairs but the theory is the same.


2 opamps => SQR(2) => 3 dB
4 opamps => SQR(4) => 6 dB

etc.

If you use transistors in parallel the noise is lowered by a factor of square root of 8 (= 2.8 = 9 dB).

Noise improvements with extra pairs:

No of pairs Factor square of Factor dB
1 2 1.41 3
2 4 2 6
3 6 2.4 7.8
4 8 2.8 9
5 10 3.2 10
6 12 3.5 10.8

15 30 5.5 14.8

Check the table in original if you don't see the numbers clearly.

 

[fmak]

Before you do anything like this, estimate how low noise you have INCLUDING signal source noise.

You will soon discover that this approach is rather pointless these days (not in 70's and 80's)!

-----------------------------------------------------------------------------------
Why, when in 21st century high end designers are using parallel dacs to get 6dB or better improvement. At one point Accuphase used 16 dacs!

Opamps with very noise such as 5532 don't sound very good. In repalcing with better sounding ones, s/n suffers. So to address the imbalance, why not stack opmaps of choice for sound to get the dynamic range needed for 24/96 audio.

Some people seem to be stuck in the 16/44.1 er

 

[fmak]

Before you do anything like this, estimate how low noise you have INCLUDING signal source noise.

You will soon discover that this approach is rather pointless these days (not in 70's and 80's)!
------------------------------------------------------------------------------------
Before you say this is pointless, try it on low voltage version of your regulator. 2 AD817s can nearly match the noise level of the AD797 for higher voltage versions.

Yes, I tried it;see my post under Jung Regulator.