Negative feedback take off point

[fglabach]

I have a stereo amp I am converting into a monoblock. See the attachment where I have crudely sketched the physical layout of the ouput stage and the global negative feedback take off point

Diagram A shows one channel of the original stereo amp with the take off point on the emitter resistor of the centrally located output transistor back to the input stage.

Diagram D is how it is configured now where it is a monoblock with 2 heatsinks each with 6 transistors. This appears to be suboptimal because the feeback from the right heatsink has to go through the the binding post of the output to return to the feedback point at the input. The feedback may not see the right heatsink transistors at all.

Diagram B would appear to be ideal because the takeoff point is centrally located where the positive output from both heatsinks join. This means I have to create a 3 wire T connection and it may make it hard to dissemble the amplifier

Diagram C is what I propose to do where there is a seperate feedback wire from each heatsink back to a common feedback point in the input stage. The wires will not be the exact same length. Is this the same as Diagram B from a feedback standpoint?

Does anyone know whether C is acceptable with 2 wires going back to the input stage or do I need to do D? Or is there even another suggestion for the global feedback take off point.

I take it everyone would agree that I cannot leave it like it is in Diagram D because the right heatsink may not be getting enough feedback even though the wires going to the binding post are 12 gauge and relatively short.

[AndrewT]

none of these are correct.
Read D.Self distortion 7
Do not tap into the trace/lead connecting the emitter resistors.
You must tap into the output lead.
Since there are two output leads, then it seems the NFB tapping should be where they meet.

[fglabach]

I did read Self and that is what informed my not terribly well formed thinking.

My drawing was very crude because I was in a hurry.

After looking at the circuit board more carefully the NF point is a spur off the output lead connection point (to avoid cross currents) and not from any one emitter resistor.

I appreciate you pointing out the error in my drawing.

Is there really a difference between running a wire between these spurs and then putting a T connection at the middle and then running a single wire back to the input stage or simply running two wires, one from each spur back to the input stage?

[Tim__x]

In case you didn't know. You can't simply parallel two channels of a stereo amp and get a monoblock. Very slight gain mismatches will cause self-destruction.

Now that that's out of the way. The only really important things are: to minimize the amount of inductance in the feedback loop for the sake of stability; and to take feedback from the same place you take the output.

[fglabach]

Tim: Thanks for the heads up.

I am using only one input, driver circuit to drive more output transistors. I did use a more robust driver transistor.

I know from a theoretical standpoint that the feedback point should be the output with minmal inductance back to the input.

I am just wondering if it matters from a physical location standpoint whether I need to take the feedback from where the 2 ouputs join near the binding posts, which is very high current, or can I simply run 2 wires, on from each board at the ouput spur feedback point back to the input stage.

[Tim__x]

The feedback isn't there for the sake of the output stage. It's there to detect errors at the output, in other words, it's there to measure the difference between the actual output and what was desired to be at the output.

Tie the two output stages together however you want, then wire the output connector to the output stages with a single wire leading to a single point, connect the feedback take-off wire to the very same point that the output wire connects to. It doesn't matter much where that is, so long as the output and the feedback are taken from the same point.

[Tim__x]

Borrowing the general layout of your drawings, here's a quick diagram showing what I mean.
Both are acceptable, the one on the right will allow slightly better current sharing between output stages.

[fglabach]

Here is my drawing as to what is already in place.

I am using multiple solid core wire and then using both relays to improve the current path.

I will look at your drawing now.

[fglabach]

This attachment is somewhat clearer

[fglabach]

Tim. Thank you so much. Your drawing is very clear.