Dumb negative feedback questions

[Zozobra]

I'm kicking about a design with the following topology:

input > EF86 (V1) > gain pot > ecc83 (V2a) gain stage > master vol pot > bridged T mid cut > AC coupled cathodyne PI (V2b) > crossline presence control > 2x6V6/6L6 cathode biased power amp

My apologies for the written topology, I drew a schematic up last night but forgot to grab a digital copy of it I'll add it here later today.

I want to use some global NFB from the poweramp to the cathode of V2a, however I'd also like to experiment with local NFB around V2a as well.
Will they fight each other and cause instability or odd behaviour with them being applied to the same node?
Is putting a bridged-t mid cut in the global loop just asking for trouble?

[pdf64]

Generally I think it would be a bad idea to mess with the gain / phase response of the system within the loop, at least within the bandwidth of interest.
However, I doubt that you will have enough gain for there to be a problem as such.
A master volume within a NFB loop is an obvious example of poor practice but within our application it's often used with good effect.
For your topology, the engineer in me prefers local feedback, eg consider a UL output stage.

[Mike Sulzer]

I'm kicking about a design with the following topology:

input > EF86 (V1) > gain pot > ecc83 (V2a) gain stage > master vol pot > bridged T mid cut > AC coupled cathodyne PI (V2b) > crossline presence control > 2x6V6/6L6 cathode biased power amp

My apologies for the written topology, I drew a schematic up last night but forgot to grab a digital copy of it I'll add it here later today.

I want to use some global NFB from the poweramp to the cathode of V2a, however I'd also like to experiment with local NFB around V2a as well.
Will they fight each other and cause instability or odd behaviour with them being applied to the same node?
Is putting a bridged-t mid cut in the global loop just asking for trouble?

The negative feed back will reduce the effect of your mid cut if included within the loop; that is, it will tend to make the frequency response flatter. So even if if does not cause instability it will partly undo what you are trying to do. Unless you do not have much feed back, but then why bother?

You might possibly consider doing the mid cut as part of the NFB circuit itself, but you really should have a very good idea what you are doing if you try that.