Negative Feedback Question

[Cap217]

I am just learning here but I am curious about NFB on a Marshall or Friedman amp and how it works.

If the NFB is on the wiper, how does this effect the sound in 4, 8, 16 ohms?

If it is changed to the 4 ohm tap, what happens to the amp? And do you need to change the resistors while making this change? I see things like NFB to 4 ohm and 220k resistors or 100k 8 ohm or 47k 16 ohm. But I don't understand what you are going to hear and feel and how big of a difference there can be.

Then when you add a depth pot it makes the NFB variable but doesn't the output impedance selector kind of do this too?


And in this layout, so I can picture it, there is a NFB on the 4ohm tap to a what? Am I reading this right seeing a 220k with a .0047uf piggybacked to a 47k? Why does it have 220 and a 47k? And what does the .0047uf piggybacked do?

http://www.ceriatone.com/ceriatone/wp-content/uploads/2018/05/AH100-01May2018.jpg

 

[RedPlated]

I'm not sure what you mean by, " the wiper."

More NFB will cancel out frequencies in the power amp section. In simple terms, more NFB sounds darker and tighter. Less sounds brighter and more open or "raw."

The impedance selector is simply a rotary switch that selects either the 4. 8 or 16 ohm secondary taps of the OT and routes the selected tap to the speaker jack. Depending on which tap you route the NFB wire to, you'll have a greater or lesser amount of NFB. 4 ohm will have the least amount of NFB, 8 in the middle and 16 the most. You can change the NFB resistor so that moving from one tap to another is almost the same. If you have say a 47k NFB resistor on the 16 ohm tap, you'll need around a 33k NFB resistor on the 8 ohm tap for similar overall NFB. Multiply the resistor x 1.4 with each move from 4 to 8 to 16 to keep NFB approximately the same. A larger NFB resistor creates less NFB as it's blocking the signal.

A JTM45 is dark and tight feeling. It's a 27k NFB resistor on a 16 ohm tap. A TON of NFB. A JCM800 is 4 or 8 ohm tap with a 100k NFB resistor. Way less NFB (because of the tap being moved and much higher value resistor blocking the NFB signal) and is much brighter and more aggressive.

A depth pot is not just another NFB control. It's a resistor/ capacitor in parallel (with each other, and in series with NFB resistor) that removes certain frequencies from the NFB so that they aren't canceled out, but rather make it to the speaker jack. Commonly, a 1meg pot with a .0047 cap is used here. The capacitor determines the frequencies that are blocked and the pot adjusts how much of that frequency is being blocked, and therefore reproduced at the speaker. A pot is just a variable resistor...so, you can replace the depth pot with a resistor and set the depth to a fixed value. A common place to set a 1 meg audio pot is around 1 o clock. That's roughly 220k on an audio pot (this would be higher on a liner taper), so you can substitute the pot with a 220k resistor, or any value resistor that you choose, depending on how much fixed depth you want.

In that Ceriatone diagram, the 220k/.0047 combo is the fixed depth circuit that I just described above. The 47k is the NFB resistor. The circuit is: NFB signal from presence control to 47k NFB resistor to 220k/.0047 depth to 4 ohm tap on impedance selector.

 

[Cap217]

I'm not sure what you mean by, " the wiper."

More NFB will cancel out frequencies in the power amp section. In simple terms more NFB sounds darker and tighter. Less sounds brighter and more open or "raw."

The impedance selector is simply a rotary switch that selects either the 4. 8 or 16 ohm secondary taps of the OT and routes the selected tap to the speaker jack. Depending on which tap you route the NFB wire to, you'll have a greater or lesser amount of NFB. 4 ohm will have the least amount of NFB, 8 in the middle and 16 the most. You can change the NFB resistor so that moving from one tap to another is almost the same. If you have say a 47k NFB resistor on the 16 ohm tap, you'll need around a 33k NFB resistor on the 8 ohm tap for similar overall NFB. Multiply the resistor x 1.4 with each move from 4 to 8 to 16 to keep NFB approximately the same.

A depth pot is not just another NFB control. It's a resistor/ capacitor in parallel that removes certain frequencies from the NFB so that they aren't canceled out, but rather make it to the speaker jack. Commonly, a 1meg pot with a .0047 cap is used here. The capacitor determines the frequencies that are blocked and the pot adjusts how much of that frequency is being blocked, and therefore reproduced at the speaker. A pot is just a variable resistor...so, you can replace the depth pot with a resistor and set the depth to a fixed value. A common place to set a 1 meg audio pot is around 1 o clock. That's roughly 220k on an audio pot (this would be higher on a liner taper), so you can substitute the pot with a 220k resistor, or any value resistor that you choose, depending on how much fixed depth you want.

By wiper i mean the impedance selector has 4 lugs. 4 8 16 and the “wiper”? Not sure that’s correct but I assume the nfb wire coming off there and it changes according to what ohm you are set at. Sorry I’m not explaining it well.

 

[RedPlated]

By wiper i mean the impedance selector has 4 lugs. 4 8 16 and the “wiper”? Not sure that’s correct but I assume the nfb wire coming off there and it changes according to what ohm you are set at. Sorry I’m not explaining it well.

The NFB wire doesn't move on a wiper. It's fixed to whichever tap it's soldered to and stays there. The only thing that changes is which OT secondary tap is routed to the speaker jack. Regardless of which ohm tap is selected for speaker cabinet purposes, the NFB tap doesn't change. There is a 4th common lug on an impedance selector which routes the selected ohm tap to the speaker jack. I could see that being construed as a "wiper."

Some amps don't have impedance selectors and instead use separate speaker jacks for each ohm setting. The secondaries are simply routed to multiple jacks instead of a selector switch. In that case, the NFB wire attaches to the speaker jack with the impedance you want.

 

[Dino 939]

I thought you were on the fence about
leaving a negative comment in the
buyer/seller comment section.
I'm finally accepting that guitar forums
have changed the wiring in my brain...sorta

 

[Cap217]

I thought you were on the fence about
leaving a negative comment in the
buyer/seller comment section.
I'm finally accepting that guitar forums
have changed the wiring in my brain...sorta

Ha. Well you are right. I might get a ton of clicks expecting a different thread.

 

[glpg80]

I am just learning here but I am curious about NFB on a Marshall or Friedman amp and how it works.

If the NFB is on the wiper, how does this effect the sound in 4, 8, 16 ohms?

If it is changed to the 4 ohm tap, what happens to the amp? And do you need to change the resistors while making this change? I see things like NFB to 4 ohm and 220k resistors or 100k 8 ohm or 47k 16 ohm. But I don't understand what you are going to hear and feel and how big of a difference there can be.

Then when you add a depth pot it makes the NFB variable but doesn't the output impedance selector kind of do this too?


And in this layout, so I can picture it, there is a NFB on the 4ohm tap to a what? Am I reading this right seeing a 220k with a .0047uf piggybacked to a 47k? Why does it have 220 and a 47k? And what does the .0047uf piggybacked do?

http://www.ceriatone.com/ceriatone/wp-content/uploads/2018/05/AH100-01May2018.jpg

The tap is irrelevant. It’s the ratio of the tap to the NFB resistor. You can get the same ratios many different ways.

The NFB resistor is in the negative feedback path. Think of the tap as a series inductor with a specific frequency response. Each tap from 16 to 4 assuming the NFB resistor is constant is a 3dB decrease in NFB, meaning less NFB, meaning more gain, but a smaller gain bandwidth product. Amps that have more NFB generally have an extended frequency response highs/lows that sounds better.

The 47k resistor that’s in series sets the NFB ratio with the tap selected. The 220k//cap is creating a low pass filter to prevent highs from entering the negative feedback path meaning the bass will be made tighter and allowed to be attenuated with the NFB ratio selected and highs will not be canceled by the NFB path, meaning more aggressiveness.

 

[Cap217]

The tap is irrelevant. It’s the ratio of the tap to the NFB resistor. You can get the same ratios many different ways.

The NFB resistor is in the negative feedback path. Think of the tap as a series inductor with a specific frequency response. Each tap from 16 to 4 assuming the NFB resistor is constant is a 3dB decrease in NFB, meaning less NFB, meaning more gain, but a smaller gain bandwidth product. Amps that have more NFB generally have an extended frequency response highs/lows that sounds better.

The 47k resistor that’s in series sets the NFB ratio with the tap selected. The 220k//cap is creating a low pass filter to prevent highs from entering the negative feedback path meaning the bass will be made tighter and allowed to be attenuated with the NFB ratio selected and highs will not be canceled by the NFB path, meaning more aggressiveness.

Thanks. Very helpful. Especially at the 47k and 220k/cap. Are those wired together? In line? I’m trying to wrap my head around how a signal can go through 2 resistors but one acts as a low pass.

Also trying to wrap my head around less negative feedback is more gain but less bandwidth. This is my first day learning this so it’s a lot to grasp. Sounds like you need to find the perfect balance. Thanks.

 

[glpg80]

Thanks. Very helpful. Especially at the 47k and 220k/cap. Are those wired together? In line? I’m trying to wrap my head around how a signal can go through 2 resistors but one acts as a low pass.

Also trying to wrap my head around less negative feedback is more gain but less bandwidth. This is my first day learning this so it’s a lot to grasp. Sounds like you need to find the perfect balance. Thanks.

The low pass filter is in series with the 47k, so think of it as whatever frequency response the low pass filter has set, the 47k sets the NFB ratio with the tap from the OT as normal for those frequencies selected only.

NFB sacrifices gain for bandwidth. Ever heard amps that have piercing loud highs and weak quiet lows? Not enough gain traded for bandwidth. NFB is your friend to designing something that will cut in a live mix and not get buried.

 

[SpiderWars]

Some good info already posted. Maybe I missed it but another element of the NFB circuit is the PI ground. In older Marshalls the PI is grounded thru the 5k Presence pot. In later Marshalls it is the 4k7 resistor hanging off of the Presence pot. In either case, that resistor is part of a voltage divider for the NFB. The NFB resistor and that resistance to ground are a voltage divider for the NFB. In old Fenders without a Presence control that resistance to ground is only 100ohms (or 47ohms depending on whether it was a 4 ohm amp or an 8 ohm amp) so the NFB resistor is only about 1k (10x the PI ground resistance). Compare that to a Marshall with that 4k7 PI ground resistance and 10x that value is 47k...right about where it often is.

 

[Cap217]

The low pass filter is in series with the 47k, so think of it as whatever frequency response the low pass filter has set, the 47k sets the NFB ratio with the tap from the OT as normal for those frequencies selected only.

NFB sacrifices gain for bandwidth. Ever heard amps that have piercing loud highs and weak quiet lows? Not enough gain traded for bandwidth. NFB is your friend to designing something that will cut in a live mix and not get buried.

I follow. But I’m still trying to understand the filtering. A 220k/.0047 filter what frequency? And does this mean that anything below is being ignored or is being cut/bypassed?

I see that Friedman’s have 220k//.0047 and 47k. But old marshalls have either 33k, 47k, or 100k. No filtering.



Let’s dive deeper.... What are the sound/feel differences between assuming all 4ohm tap and no other changes.

1- 47k no filtering
2 - 100k no filtering
3 - 47k with 220/.0047
4- 47k with 220/.0047


I might be off and the 100k is only in its off the 8ohm tap.

 

[Cap217]

Some good info already posted. Maybe I missed it but another element of the NFB circuit is the PI ground. In older Marshalls the PI is grounded thru the 5k Presence pot. In later Marshalls it is the 4k7 resistor hanging off of the Presence pot. In either case, that resistor is part of a voltage divider for the NFB. The NFB resistor and that resistance to ground are a voltage divider for the NFB. In old Fenders without a Presence control that resistance to ground is only 100ohms (or 47ohms depending on whether it was a 4 ohm amp or an 8 ohm amp) so the NFB resistor is only about 1k (10x the PI ground resistance). Compare that to a Marshall with that 4k7 PI ground resistance and 10x that value is 47k...right about where it often is.

Great info. I overlooked the presence control pot value and didn’t know it mattered. Friedman now uses 5k so is this why he uses 47k resistors?


And it looks like all the old marshalls varied from year to year.

1967 - 27k to 5k presence off the speaker jack so I assume it depended on if you were on 4,8,or 16 ohm tap
68 - 4k7 to 5k pot 8ohm tap
69 - 4k7 to 5k pot 8 ohm tap
70 - 4k7 to 5k pot 4ohm tap
73 - 100k to 5k 8ohm tap
jcm 800 - 100k to 25k pot to 4k7 to ground 4ohm tap

Im trying to understand what the above differences make to the sound and feel. Can you explain?


And I read that the old fridmans were different but the new are 47k with 220k//.0047cap in series to a 5k presence. But he did something before 2015 because I read a lot about the NFB update to the 4k ohm tap but did the resistor value change too?

 

[glpg80]

I follow. But I’m still trying to understand the filtering. A 220k/.0047 filter what frequency? And does this mean that anything below is being ignored or is being cut/bypassed?

I see that Friedman’s have 220k//.0047 and 47k. But old marshalls have either 33k, 47k, or 100k. No filtering.



Let’s dive deeper.... What are the sound/feel differences between assuming all 4ohm tap and no other changes.

1- 47k no filtering
2 - 100k no filtering
3 - 47k with 220/.0047
4- 47k with 220/.0047


I might be off and the 100k is only in its off the 8ohm tap.

The filter cuts highs, meaning lows are sent through the NFB stage. It helps tighten amps up that may have a soggy low-mid/bottom end. Highs are blocked in the NFB path meaning they are as they are at the other side of the phase inverter and nothing is impeding them in the power section on the way to the speakers.

You can also experiment. I apparently use 68 feedback topology mixed with a jcm800 feedback resistor value in my 74 which makes sense. It’s a stock 1974 chassis with a JCM800 preamp board in it. It sounds good to me so I haven’t changed it to JCM800 spec, as that increases gain but also reduces GBWP.

 

[Cap217]

The filter cuts highs, meaning lows are sent through the NFB stage. It helps tighten amps up that may have a soggy low-mid/bottom end. Highs are blocked in the NFB path meaning they are as they are at the other side of the phase inverter and nothing is impeding them in the power section on the way to the speakers.

You can also experiment. I apparently use 68 feedback topology mixed with a jcm800 feedback resistor value in my 74 which makes sense. It’s a stock 1974 chassis with a JCM800 preamp board in it. It sounds good to me so I haven’t changed it to JCM800 spec, as that increases gain but also reduces GBWP.

When adding a depth knob, how does that make a difference in filtering? Does all the filtering happen on the pot with the .0047uf cap? I see a difference of no depth knob and it goes 4ohm tap to 47k with 220/.0047 to 5k presence knob. But with a depth it goes 4ohm tap, depth knob pot 1 meg with a .0047 to a 47k resistor to a 5k presence knob.

So is the .0047 on the depth knob making a 220k/.0047 not needed?


And should a depth knob come off the 4ohm tap vs the common? I have an amp with a depth knob that comes off the common of the impedance selector. Id like to make it more aggressive sounding so would moving it to 4ohm help?

 

[Cap217]

The filter cuts highs, meaning lows are sent through the NFB stage. It helps tighten amps up that may have a soggy low-mid/bottom end. Highs are blocked in the NFB path meaning they are as they are at the other side of the phase inverter and nothing is impeding them in the power section on the way to the speakers.

You can also experiment. I apparently use 68 feedback topology mixed with a jcm800 feedback resistor value in my 74 which makes sense. It’s a stock 1974 chassis with a JCM800 preamp board in it. It sounds good to me so I haven’t changed it to JCM800 spec, as that increases gain but also reduces GBWP.

I wanted this to be its own question...

What happens to the sound/feel if you change the 47k to a 100k? Or a 33k? I currently have a 47k and am wondering what Id expect to hear and feel.

 

[scottosan]

The filter cuts highs, meaning lows are sent through the NFB stage. It helps tighten amps up that may have a soggy low-mid/bottom end. Highs are blocked in the NFB path meaning they are as they are at the other side of the phase inverter and nothing is impeding them in the power section on the way to the speakers.

You can also experiment. I apparently use 68 feedback topology mixed with a jcm800 feedback resistor value in my 74 which makes sense. It’s a stock 1974 chassis with a JCM800 preamp board in it. It sounds good to me so I haven’t changed it to JCM800 spec, as that increases gain but also reduces GBWP.

That’s the end result, but it’s important to note is that the 220k/.0047 is a high pass filter, not low pass. It’s allowing high frequencies to be fed back from the output tap to the PI. The more of a specific frequency fed back the less gain there is for that frequency. So, more highs being fed back results in less highs and because the lows are attenuated from the OT tap results in more depth or thump. When using a depth circuit, the higher the NFB resistor, the the brighter and more edge the tone will have because it reduces the amount of highs and high mids being fed back, which increased the highs and high mids in the tone.

 

[Cap217]

That’s the end result, but it’s important to note is that the 220k/.0047 is a high pass filter, not low pass. It’s allowing high frequencies to be fed back from the output tap to the PI. The more of a specific frequency fed back the less gain there is for that frequency. So, more highs being fed back results in less highs and because the lows are attenuated from the OT tap results in more depth or thump. When using a depth circuit, the higher the NFB resistor, the the brighter and more edge the tone will have because it reduces the amount of highs and high mids being fed back, which increased the highs and high mids in the tone.

I follow, thanks.

So 220k/.0047 woks on what frequency? Is that how you look at it? This must be a standard because a lot of circuits use this.


For depth circuit, the higher the nfb resistor, the brighter the tone. So is 47k standard? If I want brighter I go to 100k or darker 33k. Are those big jumps that are noticeable? If you really wanted to push the low mids and lows, what would I look to use if I have a 47k now?

 

[scottosan]

I follow, thanks.

So 220k/.0047 woks on what frequency? Is that how you look at it? This must be a standard because a lot of circuits use this.


For depth circuit, the higher the nfb resistor, the brighter the tone. So is 47k standard? If I want brighter I go to 100k or darker 33k. Are those big jumps that are noticeable? If you really wanted to push the low mids and lows, what would I look to use if I have a 47k now?

On the Brad era Cameron CCV those values are variable and he has a dark switch that lets you choose between 33k/66k/99k, it pretty noticeable. It’s like Friedman’s response circuit. The higher the NFB resister in series with the depth circuit will add mids and high end aggression

The 4700pf /220k allows frequencies over 153.92hz

https://www.omnicalculator.com/physics/rc-circuit

 

[glpg80]

That’s the end result, but it’s important to note is that the 220k/.0047 is a high pass filter, not low pass. It’s allowing high frequencies to be fed back from the output tap to the PI. The more of a specific frequency fed back the less gain there is for that frequency. So, more highs being fed back results in less highs and because the lows are attenuated from the OT tap results in more depth or thump. When using a depth circuit, the higher the NFB resistor, the the brighter and more edge the tone will have because it reduces the amount of highs and high mids being fed back, which increased the highs and high mids in the tone.

It’s a low pass filter interaction with the presence cap and if you have a shunt 4.7k resistor across your presence pot then it can also act as a high pass filter regardless of potentiometer value. There’s no explicit shunt resistor to ground at the 220k/0047 so you’re right, only at that stage is it a high pass but that’s not the whole picture as to what’s going on as you know.

 

[Cap217]

On the Brad era Cameron CCV those values are variable and he has a dark switch that lets you choose between 33k/66k/99k, it pretty noticeable. It’s like Friedman’s response circuit. The high the NFB resister in series with the depth circuit will add upper mids and high end aggression

I have an amp that doesn’t really have the aggressiveness to it that I expect. It has a depth .0047 on it. 47k resistor nfb. But the depth is wired off the impedance common, not the 4ohm.

The depth knob is also sensitive. I can only turn it up between 2 and 4 (out of 10, not o’clock) before it gets muddy and woffy.


Would this be the first place to start before worrying about values?