voltage dividers vs. lower value pots?

[EFK]

If I am trying to attenuate some degree of signal between stages, what are the ramifications of using varied methods? For example, I currently have a first gain stage going into a 1M volume pot, but I really need to bleed off some gain before hitting the second gain stage (it sounds like a jmp on '10' at about 2 on the volume pot...). For example, I could hang a 470K resistor to ground off the input to the next stage grid, but if I understand correctly, isn't that then putting 1M and 470K to ground in parallel, for a total resistance to ground of @ 300K. So, let's just say I dug up a 300K volume pot - is this the same thing? Is there any advantage to doing it one way vs. the other insofar as treble retention, clarity, effectiveness of attenuation etc? Trying to understand this. If I use a resistor to ground vs. simply reducing the value of the volume pot, is it noisier ins some way as I am dropping AC signal over two resistances vs. just one lower value pot? Just curious about the varied implications. Thanks!

[EFK]

As a side note, most Supro amps used a divider setup off the input triode wherein they would run the signal off the plate through a .005 cap, then a 270 resistor to ground, then another .005 cap before hitting the volume pot. Why not just use a single .002 (approximate) cap then the 270K? Was it merely a result of common part availability (they used a lot of .005 caps)? Or is there a benefit to attenuating the signal through a "multiple stage filter" (seems like a good way to describe it)? Or a purpose?

[ThermionicScott]

Using a smaller-value pot reduces the gain of the previous stage without introducing as much series resistance (as you would have with a resistor divider or volume pot at part-rotation) to the next stage.

[pdf64]

'For example, I could hang a 470K resistor to ground off the input to the next stage grid, but if I understand correctly, isn't that then putting 1M and 470K to ground in parallel, for a total resistance to ground of @ 300K. So, let's just say I dug up a 300K volume pot - is this the same thing?'

Say you use a 2n2 cap feeding the 1M pot. At low settings, the -3dB will be about 70Hz. However at max, the resulting 300k effective resistance will give a -3dB of about 240Hz. Useful for stopping the tone flubbing out at high gain, whilst preventing it thinning out too much at low gains.

'is there a benefit to attenuating the signal through a "multiple stage filter"'
It should result in a steeper attenuation slope, -12dB/octave for a 2 stage filter, rather than -6dB/octave for single stage. I think vox use a 4 stage filter to remove the vibrato modulation signal.
Pete.

[Enzo]

I think you are confusing a couple issues EFK.

If you want less signal, and have a 1meg pot to ground, adding parallel resistance will lower the total resistance, but other than loading it down, won;t change the signal voltage so much.

What you want to do is add a high value resistor in series on the way in. Let's say the signal came from the previous stage plate, through a cap of course to block the B+ voltage. Instead of connecting that cap directly to the 1M pot, run it through a 1M resistor on its way to the pot. Now that resistor and the pot make a voltage divider, and the voltage at the top of that pot will be half what was coming out of that previous plate.

Too much reduction? Ok use a 470k resistor and the 1M pot, now the pot will look at 2/3 the original voltage. And so on.

[EFK]

This is all very helpful! Will give me a few options to explore. Thermionic Scott your comment was something I was especially thinking about. Couple of questions -

pdf, I assume you are using an equation to figure the rolloff figures. I am terrible with math, however I'd like to know what formula or calculator you're using to determine these approximations as it certainly would be useful in predicting tonal effects to some extent. This is for a 6sj7, not a 12ax7 - I have a feeling the preceding stage affects the formula in some way?

Enzo, really clear explanation, but would sticking something like a 470K-1M resistor in the path between blocking cap and volume pot be robbing treble response? Maybe - I think - a bypass cap could be used over this resistor, some kind of small pf value I'm sure, although I guess it would have to be a trial/error approach to value unless there is one again some math that could be employed to determine the bypass frequency.

Thanks guys!

[jpfamps]

Using a resistive voltage divider (as described) will add noise, and will potentially reduce treble response due to having a large resistance in series with the next stage's grid. It will also limit grid current if the next stage is heavily overdriven.

The later two results may be desiable but the extra noise may not be.

My preferred method of reducing gain between stages is to use a split load resistor on the plate of the preceding stage. This will give lower Johnson noise and would allow experimentation with differnt grid stops for tone rather than attenuation.

Two good examples of this approach are on the stage before the PI of the Fender Brown Tolex Deluxe (6g3), and on the first stage of the Fender 6G15 Reverb unit.

[EFK]

Glad you mention it! One side of this tw0-channel setup is basically a 6G3 circuit (my favorite circuit) albeit cathode biased and with the original Supro trem in-progress (tied in via the first stage cathode resistor), and I'm entirely familiar with the split load setup off the second stage of the 6G3. So there is no down side to this? Voltage divider math I can easily handle! I usually juice the 6G3 divider up just a touch, using 82K/33K instead of the stock 100K/15K. I could easily experiment with the signal coming off this 6sj7 in the first channel (which then feeds the 6G3 second stage), because frankly it is currently making so much gain that I doubt it needs a second stage at all. Also have to fiddle with the NFB as yet.

[jpfamps]

The only real down side of the split load plate resistor is any gain stage using it is less immune to any noise on the B+; however in my experience this is not a majpr issue.

[ThermionicScott]

You can also change the operating conditions for less gain from the get-go. Lower the plate resistor value (say, to 47k or 56k) and then rebias via the cathode resistor. You risk being visited by the ghost of Leo Fender for that one, so I can't be held liable...

- Scott

[EFK]

As long as he's bringing the beer I'm alright with that!

[R ski]

Back in 93 in a freinds studio I found a 10 watt ma and pop Hifi amp and went ahead and cascaded a 12AX7 together using a 1 meg linear pot into the next stage, I will tell that gain cell sucked big time. So later on I copied the configuration of a Marshal plexi and noticed a commonality of a cascade configuration.

So what works to tame the stage is a voltage divider of two 470K ohm resistors in series from pin 7 of V2 to ground. Using the wiper of your 1 meg gain to feed the center tap of that voltage divider will tame the nastyness of that stage.

Cathode Ac signal by pass capacitors on either pin 3 of V1 or Pin 8 of V2 can be tweaked to gat some degree to get that preamp to pump, capacitive values of 0.5 Mfd to 1.0Mfd works, voltage leakage of these caps is not as critical as the coupling capacitor connecting to the plate of any tube stage.