Question RE: NFB tap

[12AX7]

Lets say you run a 100k resistor from the NFB to the speaker and you're using a 8 ohm speaker off the 8 ohm tap. You could also run a 47k from the NFB to the 4 ohm tap while using the speaker off the 8 ohm tap still. The sound the same, but the question is, will there be any benefit to the NFB being run off a different tap than the speaker? I ask because as i understand it, when you crank up an amp at some point the NFB starts having less effect and eventually at a high enough volume falls apart altogether. And i feel like i hear it happen to some extent at least well before stage volume. Will running off a different tap than the speaker have any effect on how loud you can go b4 the loop falls off? And as a second question, if not, IS there a way to keep it from happening till a louder point?

[kagliostro]

NFB consist of a portion of the output signal routed back to a stage of the amp in antiphase, something like a brake and you have more headroom

the level of the output signal, in volts, depends on the tap from which you take it, so different tap, different voltage level back to the circuit

to tailor this level usually are used resistors or pots, sometime a tone control (this to have control on the frequency that go back to the circuit)

The higher is the signal you return to the circuit the higher is the "brake" action

Hope this can be of some help, I'm not able to explain in a better way

Franco

[12AX7]

Not sure what you mean exactly but i do understand the basic idea of how NFB works. I just don't understand electronic theory enough to understand the why's and whether it's possible to lessen the issue.

[kagliostro]

Will running off a different tap than the speaker have any effect on how loud you can go b4 the loop falls off?

NO, it depends on the level of the signal you run back and this can be set using a higher or lower resistor

--

IS there a way to keep it from happening till a louder point?

To have more headroom you must increase the level of the signal you route back, so lowering the in-series resistor give you more headroom (but there is a limit beyond which you can not go and depends on the characteristics of the amplifier)

Franco

[shooter]

it's possible to lessen the issue

you mean this;

to some extent at least well before stage volume

if so, the R n Tap, from my understanding determine how much signal comes back, decreasing overall gain, but like you said, until it can't.  I very seldom use NFB, but you should be able to measure the AC at the speaker and "see" how more or less NFB changes the AC @ speaker, with a scope you can probably even determine how it affects the freq spectrum.

[jjasilli]

Lets say you run a 100k resistor from the NFB to the speaker and you're using a 8 ohm speaker off the 8 ohm tap. You could also run a 47k from the NFB to the 4 ohm tap while using the speaker off the 8 ohm tap still. The sound the same, but the question is, will there be any benefit to the NFB being run off a different tap than the speaker? I ask because as i understand it, when you crank up an amp at some point the NFB starts having less effect and eventually at a high enough volume falls apart altogether. And i feel like i hear it happen to some extent at least well before stage volume. Will running off a different tap than the speaker have any effect on how loud you can go b4 the loop falls off? And as a second question, if not, IS there a way to keep it from happening till a louder point?

This is not entirely correct.  The criteria is the amount of NFB voltage injected back into the signal.  Traditionally, NFB was used to flatten frequency response; and to damp the speaker to reduce back EMF.  In guitar amps it's come to be used as a form of tone control. 


The voltage on the 8Ω tap is NOT 2X the voltage on the 4Ω tap.  It's 2X √2 = 2X 1.414.  To get the NFB effect you want you need the correct final voltage to be injected at the return point of the NFB loop.  To get this voltage you need to know the source voltage at the OT secondary tap; then use the proper resistors (which together form a voltage divider) in the NFB loop to yield the desired final NFB voltage.  Traditionally manufacturers tend to use the highest tap to source NFB.  But you can source from any tap you choose. 


I'm not sure how to address your second question. 

[12AX7]

it's possible to lessen the issue

you mean this;

to some extent at least well before stage volume

if so, the R n Tap, from my understanding determine how much signal comes back, decreasing overall gain, but like you said, until it can't.  I very seldom use NFB, but you should be able to measure the AC at the speaker and "see" how more or less NFB changes the AC @ speaker, with a scope you can probably even determine how it affects the freq spectrum.

Thats over my head, but all i wanna know is if it CAN be made to drop out at a louder volume, be it by using a different tap than the speaker or whatever. I understand how it works as far as the tap and resistor giving you more or less NFB. Thats not the issue tho UNLESS using a ton of NFB will make it happen later. But i can't test that at home due to the volume so i thought i'd ask.

[12AX7]

Lets say you run a 100k resistor from the NFB to the speaker and you're using a 8 ohm speaker off the 8 ohm tap. You could also run a 47k from the NFB to the 4 ohm tap while using the speaker off the 8 ohm tap still. The sound the same, but the question is, will there be any benefit to the NFB being run off a different tap than the speaker? I ask because as i understand it, when you crank up an amp at some point the NFB starts having less effect and eventually at a high enough volume falls apart altogether. And i feel like i hear it happen to some extent at least well before stage volume. Will running off a different tap than the speaker have any effect on how loud you can go b4 the loop falls off? And as a second question, if not, IS there a way to keep it from happening till a louder point?

This is not entirely correct.  The criteria is the amount of NFB voltage injected back into the signal.  Traditionally, NFB was used to flatten frequency response; and to damp the speaker to reduce back EMF.  In guitar amps it's come to be used as a form of tone control. 


The voltage on the 8Ω tap is NOT 2X the voltage on the 4Ω tap.  It's 2X √2 = 2X 1.414.  To get the NFB effect you want you need the correct final voltage to be injected at the return point of the NFB loop.  To get this voltage you need to know the source voltage at the OT secondary tap; then use the proper resistors (which together form a voltage divider) in the NFB loop to yield the desired final NFB voltage.  Traditionally manufacturers tend to use the highest tap to source NFB.  But you can source from any tap you choose. 


I'm not sure how to address your second question.

I understand that, i wasn't suggesting certain values as being accurate, just to demonstrate my point. As above, i understand all that. What i don't understand is whether NFB can somehow be made to last longer, IE: to a higher volume before falling apart, and if so, how.

[PRR]

> whether NFB can somehow be made to last longer

No.

When an amp distorts, its gain falls off. NFB needs gain to work.

[12AX7]

I think i understand. You're saying it happens when the output section can no longer get louder because it's gone into distortion? If that's the case then it not what i'm hearing i guess. Can i assume until the output starts to distort the NFB will remain intact?

> whether NFB can somehow be made to last longer

No.

When an amp distorts, its gain falls off. NFB needs gain to work.

[HotBluePlates]

... You're saying it happens when the output section can no longer get louder because it's gone into distortion? If that's the case then it not what i'm hearing i guess. Can i assume until the output starts to distort the NFB will remain intact? ...

Long version of what PRR said:

 -  NFB throws-away/uses-excess gain in the feedback loop to keep the amp clean.  The amount of gain available for this purpose is the gain-difference between open-loop (no feedback) and closed-loop (with feedback) performance of the amp.

 -  When you play an amp with NFB and crank it until the output section finally does start distorting, you've actually slammed it hard enough to exceed the feedback's ability to keep the amp clean.  So that's it, feedback done, the amp is now behaving as though there were no feedback (until the driving signal drops again).

... as i understand it, when you crank up an amp at some point the NFB starts having less effect and eventually at a high enough volume falls apart altogether. ...

The middle part of this statement is false; the effect of the feedback doesn't "fall away"... It's in full-force, or it's gone entirely because the driving signal finally overwhelmed it.  In this way, feedback contributes to abrupt transition from clean to distortion, rather than a gradual change (and was a classic gripe of some in the hi-fi community; i.e. "abrupt clipping" vs "smooth overload").

... What i don't understand is whether NFB can somehow be made to last longer, IE: to a higher volume before falling apart, and if so, how.

... IS there a way to keep it from happening till a louder point?

You can cause there to be greater feedback voltage applied back at the phase inverter.  You could alter resistor values in the voltage divider from speaker to phase inverter, or use a higher-impedance tap (which increases the amount of voltage before it gets divided down).  If you increase the feedback too much, you risk making the amp unstable, and the "negative" feedback will become positive feedback.

Pick an amp and link its schematic; you'll need to be pointed at concrete components in the circuit to know what to change.

[robrob]

Speaker Tap and Feedback Levels
Speaker Tap     5F6A Light Feedback   Moderate     JTM45 Heavy

2 ohm                    27k  5F6A               20k             10k

4 ohm                    39k                        27k             13k

8 ohm                    56k                        39k             20k

16 ohm                  75k                        56k             27k  JTM45

Moderate is 41% more feedback than Light. Heavy is 2.8 times more feedback than Light.

[12AX7]

... You're saying it happens when the output section can no longer get louder because it's gone into distortion? If that's the case then it not what i'm hearing i guess. Can i assume until the output starts to distort the NFB will remain intact? ...

Long version of what PRR said:

 -  NFB throws-away/uses-excess gain in the feedback loop to keep the amp clean.  The amount of gain available for this purpose is the gain-difference between open-loop (no feedback) and closed-loop (with feedback) performance of the amp.

 -  When you play an amp with NFB and crank it until the output section finally does start distorting, you've actually slammed it hard enough to exceed the feedback's ability to keep the amp clean.  So that's it, feedback done, the amp is now behaving as though there were no feedback (until the driving signal drops again).

... as i understand it, when you crank up an amp at some point the NFB starts having less effect and eventually at a high enough volume falls apart altogether. ...

The middle part of this statement is false; the effect of the feedback doesn't "fall away"... It's in full-force, or it's gone entirely because the driving signal finally overwhelmed it.  In this way, feedback contributes to abrupt transition from clean to distortion, rather than a gradual change (and was a classic gripe of some in the hi-fi community; i.e. "abrupt clipping" vs "smooth overload").

... What i don't understand is whether NFB can somehow be made to last longer, IE: to a higher volume before falling apart, and if so, how.

... IS there a way to keep it from happening till a louder point?

You can cause there to be greater feedback voltage applied back at the phase inverter.  You could alter resistor values in the voltage divider from speaker to phase inverter, or use a higher-impedance tap (which increases the amount of voltage before it gets divided down).  If you increase the feedback too much, you risk making the amp unstable, and the "negative" feedback will become positive feedback.

Pick an amp and link its schematic; you'll need to be pointed at concrete components in the circuit to know what to change.

Thats good stuff, thanks ! I get it now. So i guess that wasn't what i hear when i crank it and i won't have to worry about it unless i crank the amp louder than i usually do. So thats good to know. It's basically a 2204 with 6V6's, but it;s the same circuit and values. Here the schematic... http://schematicheaven.net/marshallamps/jcm800_lead_mstvol_50w_2204.pdf

speaker Tap and Feedback Levels
Speaker Tap     5F6A Light Feedback   Moderate     JTM45 Heavy

2 ohm                    27k  5F6A               20k             10k

4 ohm                    39k                        27k             13k

8 ohm                    56k                        39k             20k

16 ohm                  75k                        56k             27k  JTM45

Moderate is 41% more feedback than Light. Heavy is 2.8 times more feedback than Light.

Thats a great chart but i have to ask, are you sure those are typical? First, it depends on the other resistor that makes up the voltage divider so that can;t apply to all amps. Shouldn't It should show a resistor value for the other part of the divider? Marshalls often use a 100k. I have mine switchable between 100 and 56k and always thought 56k off the 8 ohm was probably considered a lot of NFB but according to that chart it would seem thats normal and 100k is too little. I only switch to 100k at low volumes. 

[PRR]

> Shouldn't It should show a resistor value for the other part of the divider?

Yes; but it seems to be for 5F6a and the near-copy Marshall. These traditionally originally used a 5K pot as both the bottom of the NFB network and the Presence cap trim. So write "against 5K" on there.

Part of "the Marshall sound" is moving Fender's NFB source up the secondary for a tighter sound. Instead of moving to a tap you do not have, the chart suggests ways to get a similar sound with a resistor change.

But this depends mightily on the speaker and the genre. And some amps like Ampeg VT-40 ran no NFB at all. IMHO it should be tuned by-ear after smoke-test and debug. With the speakers in the room where it will be played (or a good approximation... not a Eight in an apartment unless that IS where you play).