Hoffman Amplifiers Tube Amplifier Forum
Amp Stuff => Tube Amp Building - Tweaks - Repairs => Topic started by: TubeGeek on November 22, 2014, 02:31:10 am
-
What is the purpose of the 0.05µF cap and two 22MΩ resistors on V1 across the grid and plate (pins 6 and 7)? I haven't seen this before.
Looks like some kind of feedback loop to me.
I am wondering if they are an important aspect of the D style tone.
I don't have the 22MΩ resistors in stock so I can't wire these parts in circuit until they arrive.
One of the layouts I found omits these parts all together. Also…in attachment 1, the parts are wired different than compared to the 3rd attachment. Why? Difference?
The layout I am following is: http://thenormster.com/DumbleLite/D-LiteLayoutFinal.jpg (http://thenormster.com/DumbleLite/D-LiteLayoutFinal.jpg)
:w2: :help:
-
It's LNFB (local negative feedback). It can be wired like the top or bottom layout. Some individuals say it's part of "the" sound. I've removed it or put it on a switch on every amp I've tried it on. It adds a cleaner tone on the clean channel. It's not been useful for me.
I've seen almost as many comments about people not liking it as thinking it's needed. You might consider having it on a switch on the back of the amp chassis.
with respect, Tubenit
-
Thanks for clarifying that.
Is the 22MΩ resistor value critical?
Can a lower value like 4.7MΩ be used? I guess the larger the resistance here would ensure a cleaner clean channel. I have some 1GΩ resistors (1/4W though), maybe I should try those. Or not.
edit: got some studying to do on this new to me topic of LNFB
-
Check this out:
http://www.valvewizard.co.uk/localfeedback.html (http://www.valvewizard.co.uk/localfeedback.html)
some opinions:
http://ampgarage.com/forum/viewtopic.php?t=5978&highlight=local+negative+feedback (http://ampgarage.com/forum/viewtopic.php?t=5978&highlight=local+negative+feedback)
-
Can a lower value like 4.7MΩ be used?
Sure, if you don't mind shattering the crystal lattice and destroying the fragile harmonics and writhing in agony on a lumpy, stained Motel 6 mattress instead of floating on a football field filled with marshmallows. Go right ahead.
-
:l2:
I feel like I showed up late to the party. I haven't spent much time reading up on the Dumble stuff. I am getting intrigued.
Thanks for the info, I'll get to reading now. I forgot how much fun this is. I took a few months off from building.
I finished wiring up my d'lite today, so now the experimenting can begin.
-
Tons and tons of Dumble stuff to read on ampgarage.com -- and don't make the mistake of only going back a year or two; a lot of the more open-ended, informative discussions will be 5+ years old at this point.
-
I like the LNFB on a switch located in the back of the amp.
I'm attaching a cool mod from Steve Ahola that touches on this LNFB.
-
What is the purpose of the 0.05µF cap and two 22MΩ resistors on V1 across the grid and plate (pins 6 and 7)? I haven't seen this before.
Looks like some kind of feedback loop to me.
I must not be as smart as Dumble; I don't see how they do anything.
1st-guess for the closed-loop gain of a negative feedback circuit is Rfb/Rg, where Rg is the series 220kΩ resistor in the Dumble and Rfb is 44MΩ (2x 22MΩ resistors in series with a cap for blocking d.c.).
But 44M/220k = 200, and we know the actual open-loop gain of a 12AX7 will only approach its Mu of 100 (with a crafty active load, like a transistor instead of a resistor load). True open-loop gain will be closer to Merlin's example of 64.
So the feedback resistors are too high a value to do anything; they're too large to feed any signal back to the grid to cause any feedback.
Or maybe I'm too blind to see the "Crystal Lattice." :l2:
Can a lower value like 4.7MΩ be used?
Yes. Lower values equal more negative feedback. If you source the feedback after the plate's coupling cap, you can even do away with the cap (but then you lengthen a wire run for feedback, maybe not a good thing).
Keep in mind that simply omitting the cathode bypass cap for that stage will create local negative feedback, and cut stage gain by about half. That's a simpler plan if you want ~1/2 gain in the stage.
If you're certain you need something more than 1/2-gain but less than full-gain (or also want the tone-shaping of a cathode bypass cap of 1uF or less to shave bass), then you can jigger the feedback resistor vs the series grid resistor. But the small value of Rg with a reasonable value of Rfb will create a low input impedance and make the stage hard to drive; conversely, large values of Rfb increase resistor noise.
For these reasons, Merlin suggested pretty low values of closed-loop gain (like 10 or less) to balance the trade-offs. I'm guessing you don't really want to drop gain that much, especially since the original circuit didn't drop gain at all.
What If?
What if the resistors are really 2.2MΩ (a pretty common value by comparison)? Now you'd have a 220kΩ working against 4.4MΩ for a gain of 20 (or a bit less when computed by Merlin's method). The 0.05uF cap blocks d.c. while placing the bass roll-off in sub-sub-sonic territory. The 500pF cap would work against the 220kΩ resistor to create a high-pass filter that's -3dB at about 1.5kHz. That would cause gain of the stage to rise from about 20 over most of the guitar's range up to full gain at a band of frequencies that only covers the guitar's harmonics (full-gain around 3kHz and up).
-
I don't see how they do anything
Regarding the LNFB, it genuinely changes the tone. I've done several with a spst switch. You can hear the difference pretty easily, IMO.
with respect, Tubenit
-
Regarding the LNFB, it genuinely changes the tone. I've done several with a spst switch. You can hear the difference pretty easily, IMO.
With the same 44M/220K and 0.05 cap?
Brad :dontknow:
-
22M, .047, & 22M.
-
There's one way I could imagine there being a change; but I'd need measuring equipment greater than what I have to test it. But I also believe Tubenit when he says he tried it and heard a difference.
What if the extra pathway adds capacitance from plate to grid, says 20-30pF? Such a low value is hard to measure directly; even the position of meter leads will change the measured value. But 2Deaf mentioned in another thread that he used caps with such a value on input stages to get rid of the harsh edge in multi-stage amps.
Maybe the stray capacitance of wiring such a circuit is more important than the actual parts...
-
I just wanted post an update… I did install the LNFB and do prefer it left in circuit.
I am happy with the clean channel on this amp but I am still going to play around with the OD channel, I am not quite satisfied yet. I am in no hurry, gonna let this amp burn in and study it some more.