NFB mods

[dusty712]

I recently finished building a Sunn Model T clone and I'm quite happy with the results. I was looking at taking this opportunity to learn more about the NFB loop and potentially add a couple of mods, and I'd like to get some help to understand them a bit better. I tried looking through old posts and I think I understand each mod on its own, but I'm having trouble understanding how to combine them.

One of the mods is to just add a pot to control the amount of negative feedback. The stock amp uses the 16 ohm tap and a 22k resistor, so I was thinking of simply adding a pot in series (maybe a 50k) to control this. I tested it out and it seems to do the correct thing (increasing the resistance makes the amp louder and the feel is tighter/more modern).

The other mod I wanted to understand and try out is adding a depth/resonance control. This is what the mods usually look like from what I saw, but I have a couple of questions:
1. does the 1M pot replace the NFB pot I mentioned above, or can they both work in series? I'm having trouble understand how they could both work in series properly, thinking that the depth pot would have a large effect on the amount of NFB allowed through to the NFB pot.
2. I wanted to confirm that the 0.1 cap is required for my amp if I were to perform this mod
3. would it make sense to add a switch that allows the choice of two different cap values and a bypass setting where the entire circuit is bypassed?

The Model T is really bassy as it is, so I'm not sure the depth mod would do anything useful but it looks easy enough to try and gives me an opportunity to learn a bit more about this circuit.

Thank you for the help.

[Willabe]

Last year we had a member who had a hard time understanding the difference and how to put in a resonance control, a meat control and a presence control. Here's the link, it's long but read it and watch the Uncle Doug video. And yes an amp can have any 1 or combo of 2 or all 3 of these in the NFB loop. 

Then we don't have to go all through it again. 

https://el34world.com/Forum/index.php?topic=31929.msg353041#msg353041

[dusty712]

Perfect, thank you so much for the link. I didn't see that thread in my search so I'll definitely be going through it.

Last year we had a member who had a hard time understanding the difference and how to put in a resonance control, a meat control and a presence control. Here's the link, it's long but read it and watch the Uncle Doug video. And yes an amp can have any 1 or combo of 2 or all 3 of these in the NFB loop. 

Then we don't have to go all through it again. 

https://el34world.com/Forum/index.php?topic=31929.msg353041#msg353041

[stratomaster]

Last year we had a member who had a hard time understanding the difference and how to put in a resonance control, a meat control and a presence control. Here's the link, it's long but read it and watch the Uncle Doug video. And yes an amp can have any 1 or combo of 2 or all 3 of these in the NFB loop. 

Then we don't have to go all through it again. 

https://el34world.com/Forum/index.php?topic=31929.msg353041#msg353041

I ain't doing it again. Someone else's turn this time. 

[Willabe]

I ain't doing it again. Someone else's turn this time. 

Oh come on, that was fun!     

[dusty712]

I really do appreciate both of yours patience and advice you had throughout that whole thread. It gave me some materials to look through later on.

I definitely don't want you to repeat or go through any of that again. I think what was tripping me up was the interaction between the resonance control and the NFB pot. After drawing it out and seeing more clearly that the resonance pot is in parallel with the cap it made more sense. I attached the diagram of what I hope is the right way to connect the two, and also how to wire a selector to either bypass the resonance circuit or choose two different cap values (I think either a 3 way switch would work or a 2 way with a pot that has a built in spst switch). If you wouldn't mind having a quick look in case I got it all wrong, I'd appreciate it.

[Willabe]

Looks right to me.

But you don't need the bypass on the switch. When you turn the 1M pot up all the way, it's the same as by passed.

The 50K pot in the adjustable NFB should be 1MA to take the NFB out completely. Many call it  a 'meat' control.
 
You can see that you can have all 3 of those mods in series with the NFB loop.

[dusty712]

Thank you for checking it over. The only reason I was thinking of having the bypass is to eliminate any potential effect the capacitor may have. I'll take a listen first to see if it's worth it as the bypass is mainly from a "purist" perspective where I still want to be able the have the Model T circuit as was originally.

Looks right to me.

But you don't need the bypass on the switch. When you turn the 1M pot up all the way, it's the same as by passed.

You can see that you can have all 3 of those mods in series with the NFB loop.

[Willabe]

Look at my edit on my last post.

Thank you for checking it over. The only reason I was thinking of having the bypass is to eliminate any potential effect the capacitor may have.

Look at your drawing. When you turn the 1M pot up all the way, 0 resistance, it's exactly the same as being by passed. Or should I say all the way down? Pot set at 0 resistance.

[dusty712]

Yep I know what you mean and in theory that's true. I just thought, and I could be very wrong here, that in practice there could still be a bit of current that would flow through the capacitor (because the pot will likely not be completely 0 and also I thought that current still will follow all paths, but in different proportions). Again, most likely negligible and I'll take a listen first.

Regarding the "meat" pot, I did try a 1M pot there initially I just noticed that after the first bit of travel it had no impact on the sound at all and the 50k seemed to be just fine. I'll give it another listen with the 1M though.

Look at your drawing. When you turn the 1M pot up all the way, 0 resistance, it's exactly the same as being by passed. Or should I say all the way down? Pot set at 0 resistance.

[Willabe]

Yep I know what you mean and in theory that's true.

It's not theory, it's mechanically fact.

I just thought, and I could be very wrong here, that in practice there could still be a bit of current that would flow through the capacitor (because the pot will likely not be completely 0 and also I thought that current still will follow all paths, but in different proportions). Again, most likely negligible and I'll take a listen first.

So measure the pot's resistance set at 0, wiper to O end of the pot.

The signal will take the path of least resistance. In this case their shouldn't be any resistance at all with the pot set to 0. But if there's a few ohm's, you'd never hear any difference.

[dusty712]

Willabe, I hope you know that I wasn't disagreeing or arguing with you and that I do appreciate your help. I'm a very new member but the community has been really helpful. Regarding my thinking here, I just remember in the past being corrected that current will flow everywhere inversely proportional to the resistance/impedance, not just the path of least resistance. In this case I thought that even if a smallest amount of current were to leak into the capacitors then maybe eventually it could have an effect. Again though, it's not something I worry about here. If there were any reason to want to have a separate bypass path it would mainly be just so that the signal doesn't have to go through unnecessary components, like an extra pot.

I wired it all last night and it all looks to work and makes the amp quite a bit more versatile. However, for the purposes I built it, which is their infamous use by doom bands, I much prefer the stock version as I find it takes pedals up front much better.

One question that maybe you or someone can help me understand (or point me to another thread if it was discussed before). If I have the master volume cranked (say around noon) then reducing the NFB or increasing resonance even by a quarter turn or less results in a lot of noise and feedback. Did I make a mistake when adding these in, or is this possibly expected because the power amp section is designed for a lot of NFB?

[pdf64]

... If I have the master volume cranked (say around noon) then reducing the NFB or increasing resonance even by a quarter turn or less results in a lot of noise and feedback. Did I make a mistake when adding these in, or is this possibly expected because the power amp section is designed for a lot of NFB?

That seems weird, by 'feedback' do you mean oscillation? Could you record the issue, upload to YouTube / soundcloud etc, and provide a link.
Do the channel volume settings make a difference?
https://el34world.com/charts/Schematics/files/Sunn/Sunn_model_t.pdf
I expect UL to reduce the output stage gain a bit, compared to an equivalent regular pentode arrangement.
Even so, it does seem a high NFB ratio.

Was your 1M NFB pot audio or linear taper? Did you wire it so that turning it clockwise increased the volume?

[Willabe]

Regarding my thinking here, I just remember in the past being corrected that current will flow everywhere inversely proportional to the resistance/impedance, not just the path of least resistance. In this case I thought that even if a smallest amount of current were to leak into the capacitors then maybe eventually it could have an effect. Again though, it's not something I worry about here.

For someone who is not worried about it, you keep talking about it. And you are arguing with me. 

"even the smallest amount of current" It's not that sensitive of a circuit.

How could it leak into the caps as part of the circuit? The switch disconnects 1 end of the caps from the OT. It chooses either the 0.0047 cap, the 0.01 cap or the 1M pot. And any signal/acv from after the 1M pot if it goes into the caps, it has no place else to go. So NO losses.   

... current will flow everywhere inversely proportional to the resistance/impedance, not just the path of least resistance.

     You're making this way over complicated. It's just a switch, 2 caps, 1 pot and only 1 signal path. 

This can only be if the signal is being fed to more than 1 circuit path and there's a 2nd, 3rd, etc. path to ground. The way you have the switch drawn the signal will go where it finds a path to flow, it has no other choice.

[dusty712]

I'm really not arguing, please don't take it as that, but if you do I apologize.

I'll ask this because my diagram may have been confusing, but the 1M pot is meant to be in parallel with the capacitors, so the way I wired it is that it's always connected to the OT. With a 2-way switch as I have now, one of the caps will always be connected to the OT. That's why in my diagram I had a third straight wire as a bypass if I had a 3 way switch, so that neither of the caps are connected. If this doesn't change anything then consider the issue dropped.

How could it leak into the caps as part of the circuit? The switch disconnects 1 end of the caps from the OT. It chooses either the 0.0047 cap, the 0.01 cap or the 1M pot. And any signal/acv from after the 1M pot if it goes into the caps, it has no place else to go. So NO losses.   

[dusty712]

First I'll mention that the noise went away when I plugged straight into the amp. I could turn the input channels volume at max, the master volume at max and the nfb/resonance pots at max and no feedback/noise (just white noise).

Previously I was going through my pedalboard, which clearly has some noise issues. For completeness, this is the noise I was hearing: https://soundcloud.com/user-981938832/model-t-with-mods-noise

I used a 100k pot for the NFB and 1M for the resonance; both I think were linear as that's all I had but other than ramp-up I wouldn't think it would a difference.

That seems weird, by 'feedback' do you mean oscillation? Could you record the issue, upload to YouTube / soundcloud etc, and provide a link.
Do the channel volume settings make a difference?
https://el34world.com/charts/Schematics/files/Sunn/Sunn_model_t.pdf
I expect UL to reduce the output stage gain a bit, compared to an equivalent regular pentode arrangement.
Even so, it does seem a high NFB ratio.

Was your 1M NFB pot audio or linear taper? Did you wire it so that turning it clockwise increased the volume?

[Willabe]

I'm sorry, I got mixed up, you have the resonance control drawn wrong. Both caps are always connected. And the 1M pot's wiper is tied to 1 end of itself.

Look at this again, you'll see it;

https://robrobinette.com/Voicing_an_Amp.htm#Resonance_Control