John Curl's Blowtorch preamplifier part II

[scott wurcer]

Originally Posted by simon7000

Here are two theoretical circuits. When made with real components will they have the same noise level?

Is there any signal correlated noise?

Is this a trick question? You need to identify the op-amps and define signal correlated noise.

[simon7000]

Originally Posted by scott wurcer

Is this a trick question? You need to identify the op-amps and define signal correlated noise.

Pick the same op amp for all three. If you optimize op amps you can do better of course, but the issue is topology.

As to noise, does the noise level increase when signal is present? (Pick any real op amp for that question.) What contributes to the increase in noise level?

[scott wurcer]

Originally Posted by simon7000

Pick the same op amp for all three. If you optimize op amps you can do better of course, but the issue is topology.

As to noise, does the noise level increase when signal is present? (Pick any real op amp for that question.) What contributes to the increase in noise level?

The AD8099 has a flat input gm and so much BW that the operating point of the input devices changes only minutely so compared with the 1K input (4nV) it would be unmeasurable. In general any increase in noise in internal stages does not get back to the input to any measure. If that's what you are talking about, I'm never sure.

[jcx]

we could probably list a few 2nd order effects - most not unique to op amps - any transistors, resistors will show some level of noise modulation with signal

bulk metal foil R are "the best" on excess noise, TC - but we've been told they don't "sound the best"

and these 2nd order noise modulation effects can be orders of magnitude less than the Zoo of analog magnetic tape noise artifacts

[scott wurcer]

Originally Posted by jcx

we could probably list a few 2nd order effects - most not unique to op amps - any transistors, resistors will show some level of noise modulation with signal

bulk metal foil R are "the best" on excess noise, TC - but we've been told they don't "sound the best"

and these 2nd order noise modulation effects can be orders of magnitude less than the
Zoo of analog magnetic tape noise artifacts

It also depends where the excess noise occurs, if inside the feedback loop the contribution at the input must be calculated. In general the first high gain configuration would have the least opportunity for this since excess noise goes as the square of the voltage.

[simon7000]

Originally Posted by scott wurcer

The AD8099 has a flat input gm and so much BW that the operating point of the input devices changes only minutely so compared with the 1K input (4nV) it would be unmeasurable. In general any increase in noise in internal stages does not get back to the input to any measure. If that's what you are talking about, I'm never sure.

Thanks, I've ordered a few to play with.

The issue is that in a real circuit there is always noise. Some of it comes from the active components, others from the passive parts and still other comes in from the external connections, such as input and output leads, along with the well know power supply issues.

Which noise source is dominant? What happens when you have two noise sources and a non-linear channel?

With matching inverting gain stages some distortions cancel and the result can be better than predicted by simple feedback assumptions. In the case of uncorrelated noise the math is almost clear.

But in all real amplifiers there may be some bandwidth modulation due to signal and or slew rate limitations. Does this influence the noise? If there is a signal related correlated noise what is the bandwidth of such noise.

Can this noise conjugate with other stage noise to produce correlated artifacts back in the audio band?

At what levels do these issues occur?

[simon7000]

Originally Posted by scott wurcer

It also depends where the excess noise occurs, if inside the feedback loop the contribution at the input must be calculated. In general the first high gain configuration would have the least opportunity for this since excess noise goes as the square of the voltage.

Yes signal induced error in the feedback loop is a controllable issue when recognized.

[jlsem]

Originally Posted by scott wurcer

Never meant all of them. As SY mentioned Sheffield Labs even made "The Ride of the Valkiries" boring. Definately not the version I would play in my chopper.

Doesn't that have more to do with Erich Leinsdorf than Sheffield Labs? On the other hand, nothing the Los Angeles Symphony has done has ever excited me.

John

[scott wurcer]

Originally Posted by jlsem

Doesn't that have more to do with Erich Leinsdorf than Sheffield Labs? On the other hand, nothing the Los Angeles Symphony has done has ever excited me.

John

I guess that's all you can expect from a vanity lable. Lyrita had some excellent stuff in their catalog that I enjoyed a lot more. Nonsuch's Geroge Crumb series is still one of my favorites even though groove pre-echo is obvious on several tracks.

[Pano]

Originally Posted by jlsem

nothing the Los Angeles Symphony has done has ever excited me.

I heard them accompany Abel Gance's Napoleon live. It was pretty exciting.
Heard some LA stuff on the radio last week that I thought was good.