Global NFB & Back EMF

[jjasilli]

Anyone care to chime-in?  Can't get my head around this.  Is back EMF necessary as the source of global NFB?  IOW, if an amp with a global NFB network were played into a dummyload, would there still be an NFB signal? 


If yes, what is the effect on NFB signal if a speaker were connected in place of the dummyload?


If a passive attenuator is placed between the OT and the speaker, does this affect NFB voltage?


If an electronic, virtually active, attenuator circuit (as opposed to, say, a Weber Mass) is used, does this affect NFB voltage?


Thanks!

[HotBluePlates]

... Can't get my head around this.  Is back EMF necessary as ...

If it's confusing to you then forget about it & pretend it doesn't exist.  Negative feedback will do the same things for you it always did whether you're adept at explaining "back-EMF" or not.

... Can't get my head around this.  ... back EMF ...

Think of it like this:
It's really obvious a 200-pound person exerts a 200-pound downward force against the earth.
When that person stands on solid-ground, that ground exerts a "200-pound normal force" upward to support them.
If the "normal force" didn't exist, they would sink into the ground (kind of like with water, but if buoyancy didn't exist).

For "EMF" applied to the speaker, there is also a "back-EMF."  It's the "earth" to the amp's "200-pound person."
But negative feedback is just taking "a % of voltage applied to the speaker" and applying it to the amp somewhere.  It will work whether we think about back-EMF or not, because it's still just a a voltage-divider sampling a percentage of the amp's output.

[acheld]

Is back EMF necessary as the source of global NFB?

   No.  It's all about the voltage developed in the NFB loop (very little current flows into the signal path).   So, however that voltage is developed, that's all that matters.    Now, obviously the impedance of a speaker changes quite a bit with frequency, just to complicate the situation. 

what is the effect on NFB signal if a speaker were connected in place of the dummyload?

  If the impedance (at each frequency) were the same for each device (NOT likely), then no change.   

If a passive attenuator is placed between the OT and the speaker, does this affect NFB voltage?

Yes, to the extent that the passive attenuator changes the impedance of the speaker system and thus has an effect on the NFB voltage.  I would not expect much of a change, but haven't done any calculations on this scenario.

If an electronic, virtually active, attenuator circuit (as opposed to, say, a Weber Mass) is used, does this affect NFB voltage?

The type of the load -- eg, passive or active -- does not matter, subject of course to variance of impedance over frequency. 

[tubeswell]

‘Back EMF’ happens in an inductor. It occurs because changing current causes induction, which in-turn causes (other) changing current.


A Transformer works by induction. Back-EMF occurs in a transformer which has changing current, so it will be present whether or not you’re considering NFB sourced from the OT secondary.


On a separate note, if your NFB source is in phase with the NFB insertion point into the main signal, you will get positive feedback.

[shooter]

confused here;
is counter EMF = back EMF?, I don't recall the term "back EMF"


I understand HBP's explanation, even used a similar argument to explain the physics of 'lil Johnny running headfirst into a tree n getting knocked on his Keyster(sp?)
thanks
dave

[HotBluePlates]

... if an amp with a global NFB network were played into a dummyload, would there still be an NFB signal? ...

A Deluxe Reverb's negative feedback loop consists of an 820Ω and 47Ω resistor.
47Ω / (820Ω + 47Ω) = 47 / 867 = ~5.42 %  --->  5.42% of the speaker-signal is applied as feedback.

A Deluxe Reverb supposedly delivers 22 watts to its 8Ω speaker.
√(22 watts x 8Ω) = 13.27 volts RMS
13.27v x 0.0542 = ~0.72 volts RMS of feedback.

If the dummy load exhibits 8Ω of impedance to the amp-output, there will be something like 0.72v RMS of feedback at max clean output power.

... what is the effect on NFB signal if a speaker were connected in place of the dummyload?

If the load were "always 8Ω" then the feedback voltage would always be a constant 0.72v RMS at max clean output.  But the negative feedback will be a varying voltage with a speaker used as the load.

The speaker is not "8Ω" at all frequencies.  It has reactance, and behaves as those it is an inductive  load for some frequencies (between 0Hz & its bass resonant peak, and again for all frequencies higher than its "nominal impedance") while behaving as a capacitive load for other frequencies (between its bass resonant peak & its "nominal impedance").  This changing Impedance Curve is shown in the dark line below:



Pretend the amp-output is a constant current, regardless of the load impedance.  Volts = Current x Impedance, so the voltage at the speaker terminals will be higher wherever impedance is higher than the nominal "8Ω".

   - Wherever speaker-voltage would tend to be higher (due to higher impedance), it is also "louder."
   - The feedback circuit samples the voltage across the speaker terminals, which counteracts amp-output.
   - So feedback counteracts  an effect where higher speaker-impedance tends to cause higher-amp-output.

   - This is especially effective at reducing speaker-flap at the bass resonant frequency (where speaker impedance is something like 140Ω in the diagram).
   - Feedback also counteracts "rising treble response" due to the speaker's impedance, though the speaker also has a treble roll-off for other reasons that counteracts this effect.

If a passive attenuator is placed between the OT and the speaker, does this affect NFB voltage?

If an electronic, virtually active, attenuator circuit ... is used, does this affect NFB voltage?

"Passive" and "Active" are the wrong terms to use.  We care about "Resistive" and "Reactive".

A resistive attenuator is a mostly-constant impedance.
It doesn't cause an "apparent treble/bass boost" the way a speaker's changing-impedance does.
Fletcher-Munson compounds this by reducing apparent treble/bass when loudness is reduced.

A reactive load/attenuator attempts to mimic a speaker's varying impedance.
If the reactances are juggled correctly, it causes the same "rising bass/treble" that a speaker changing-impedance might cause.
Fletcher-Munson still reduces apparent treble/bass with reduced loudness, so some reactive loads (like Fryette's Power Station re-amper) offer various settings to slightly bump up bass & treble to accommodate.

______________________________________________

We make the simplifying assumption that "8Ω speaker is always 8Ω."  It isn't.
Negative feedback normally acts to squash the impacts created by changing-impedance (by sampling & changing amp-output).

Resistive loads don't react the way a speaker does, and if we sampled the voltage across them to listen something sounds "not quite right."

Reactive loads attempt to mimic the speaker, so the sampled signal sounds "the way we expect."

[tubeswell]

is counter EMF = back EMF?

Yes. Same thing

[jjasilli]

Thanks, guys!