Voltage noise in opamp inputs.. how much is too much?

[Svart]

Hey Ted, can you give us some insight into current feedback? I believe you used this topology in some of your equipment, right?

 

[Samuel Groner]

Thanks for the test, Ted!

Using 3k3 feedback resistors and 10 channels, the noise attributable to the mix amp would be -106 dB (using NE33078)... Or have I got my sums wrong? It could be 6dB better than that.

I'd say it is just the -126.8 dB figure, no? 10 3k3 resistors in parallel behave like a single 330? :?

Samuel

 

[TedF]

Apologies guys.... I made a mistake in my 'tests' posting; the ic type that came out best is LM33078 (I said it was NE33078 which was wrong). For those who don't know, the LM33078 has conventional silicon front end, and a very simple 'lopsided' output stage; it operates predominantly in class A which is why I like it.
I can't agree with Svart; I think it sounds good, and use it everywhere.... but a discrete amp made of BC109s still sounds better!! :grin:

'Mediatechnology' asked about the feedback resistors in the test circuit... yes they were 3K3; low enough to perform well in a medium impedance environment, and high enough to avoid too much distortion.

Re current feedback.... I did already publish my balanced mix circuit here previously.
I'm not clever enough to find it and point you there, but perhaps one of you flash guys will do that. :grin:

 

[Svart]

I did already publish my balanced mix circuit here previously

I missed that!

someone please point it out to me!

:thumb:

 

[Mlewis]

woohoo! i'm Flash!!

http://www.theopiumdenproductions.com/balmix.jpg

and

http://www.groupdiy.com/index.php?topic=14512&start=15&postdays=0&postorder=asc&highlight=superbal

 

[Svart]

http://www.theopiumdenproductions.com/balmix.jpg

Oh ****! that's my website!

I guess I DID see that before, but I only remembered it when i saw it!

 

[Svart]

I run all my opamps in forced class A, possibly a reason the 33078 doesn't sound as good for me..

 

[TedF]

Hmmm.... I think you may be right Svart. I believe that the LM33078 has a simple lopsided output arrangement, so it would not respond well to being dragged towards one of the rails. used 'au naturel' I think it sounds nice. My philosophy is to avoid driving the chips higher than about 20dB gain; and where they are used as impedance changers or just followers, then there's nothing wrong with good old TL072s! :razz: :guinness:

 

[rodabod]

[quote author="Svart"]I run all my opamps in forced class A, possibly a reason the 33078 doesn't sound as good for me..[/quote]

Svart, we've talked about this before - are you still using the simple resistor trick (ie. 10K or so...)? Or are you using any of the other techniques?

I wasn't sure whether to try it, as I thought it maybe seemed a bit like a hack.

Would you recommend it? Maybe I should do a blind a/b.

 

[Svart]

Hey I'm using the simple resistor trick. I tried many of the other types but they are largely cost prohibitive. You can indeed hear a difference between the forced A version and the normal version of a 5532 but it'll be personal taste in the end if you like it or not.

and I used 3k from outputs to -V