Preamp cathodes - question...

[EFK]

Excuse my ignorance on this but I'm having a hard time wrapping my mind around something: if both halves of a 12ax7 in the preamp are being used as gain stages for one channel as such: 1st half --->volume --->2nd half ---> tone stack..., does it make a difference if each of the cathodes 3 and 8 are tied together through one shared resistor and bypass cap (say 1.5K/22uf) or if each cathode has it's own 1.5K/22uf? This is assuming the split cathodes are loaded with identical resistor/bypass cap values. Why would one split the cathodes on a single 12ax7 driving a single channel? Try to keep it mind I'm new to this so you might have to dumb it down for me!

[pdf64]

If the 2 sections are in series, ie signal goes into 1 then the other, then best to split them.
If bypassing was perfect then wouldn't make any difference. However it never is.
So consider the situation where the cathode resistor is shared and not bypassed. The same signal would appear on the cathode as the respective grid, except that it's the same signal but amplified and inverted, with a bit of phase shifting caused by coupling cap time constant, bright cap, Miller capacitence.
You can see that the situation would be uncontrolled and would vary according to the volume control setting.
If one or the other or both got overdriven, then it would be even more difficult for the designer to determine how the circuit would respond.
Obviously if then bypassed, the issues would mostly go away, but as the cap aged and it's ESR increased, then it would behave ever more strangely.
So best to split them and have done.
Unless it's a deliberate choice to go for weird and undefined. Peter.

[EFK]

Got it! That makes perfect sense. Thank you!

[mooreamps]

Unless it's a deliberate choice to go for weird and undefined. Peter.

Oh, for the love of g**. !!

and I'm assuming you've cobbled up a circuit to prove all of this ???


you know, if it works for a pair of power tubes, don't you think it just "might" be OK for a pair of preamp gain stages ????


-g

[Enzo]

That is not a good example gary. Power tubes are all on the same page. two input stages share a common cathode in various familiar amps. And certainly the example of two triodes in parallel would be trivial. But combining the cathodes of say the input stage and the reverb driver? or the trem oscillator tube and the reverb return? Those may be far fetched, but I don;t think it unreasonable to think that in some circuits sharing a cathode might not be the best approach.

[Merlinb]

Enzo is right. If two stages are cascaded then having any kind of shared cathode impedance is positive feedback, and frequently leads to oscillation or ringing. (That's how a lot of valve signal generators were deisgned).

The two valves in a push-pull stage are not cascaded, they are a differential pair, so shared cathodes are ok. Same with the common Marshall input stage with shared cathodes; the two valves are not cascaded.

[pdf64]

'and I'm assuming you've cobbled up a circuit to prove all of this ???'
To test this for yourself, build a 5C3 type phase splitter (cascode?), but leave off the bypass cap. The gradually increase the signal level and you'll see what I'm on about.
I can't think of any amp with sequential gain stages where cathodes are shared? Try it (ie series up the input stages of a 5F6A) and you'll see why it really isn't a good idea. Peter.

[chipaudette]

65 Deluxe Reverb shares...

FWIW, the Fender blackface amps (Deluxe Reverb, Twin Reverb, Super Reverb) all implement their vibrato/reverb channels using pre-amp stages that share cathode resistors/caps.

http://www.fender.com/support/amp_sc..._Schematic.pdf

Check out V4A and V4B. V4A is the reverb recovery and V4B is the mixer of the reverb with the dry signal. They share a cathode resistor and cap. Note that they are sequential (cascaded) and it works just fine. But, I also agree that this isn't exactly the same situation as asked about in the original post.

I happen have a reissue Deluxe Reverb. I also happened to have modified it for some high-gain experiments. One of my mods happens to remove the shared cathode bypass cap but leaves the shared cathode resistor. Yes, that is weird...but it does inform this discussion somewhat. Removing the cap but leaving the shared resistor should be a disaster resulting in massive squeeling feedback. Interestingly, it doesn't squeel unless I turn the Reverb pot up to 8. I guess the circuit feedback isn't really that strong.

Fun experiment.

Chip

[pdf64]

'Check out V4A and V4B. V4A is the reverb recovery and V4B is the mixer of the reverb with the dry signal. They share a cathode resistor and cap.'
Fair point, Chip.
I guess that it's a 'not good practice' kind of thing, which you may get away with, or may cause some weird behaviour.
'Removing the cap but leaving the shared resistor should be a disaster resulting in massive squeeling feedback.'
No-one was suggesting that would be an automatic consequence. However, you would find that a volume control between the 2 stages would reach zero volume at about 3, and then the volume would increase again as it was turned down to zero. Peter.

[mooreamps]

For preamps, I think the real key here is, that for a shared cathode resistor, the signal going through both gain stages must be in phase with respect to each other. If not, then yes, one could expect the gain stage to be "unstable". Perhaps it was this point I should have made more clear. I also look at some of the old Matchless prints where they use parallel gain stages in the preamp, and share a common cathode R/C. Again, in this case the signal passing through both gain stages are "in phase" with respect to each other.

You can put this "in your book", I don't care.... But, if you do, just make sure you are understanding how this works.



-g

[pdf64]

As I see it, it's the potential for coupling that is the issue which the designer should factor in. The phase just determines whether that coupling (assuming it's the same signal) will be negative or positive. Peter.