Parallel Cathode Caps and Shared cathode Caps/Resistors

[fiftynine]

Why do some amps have 2 cathode caps in parallel (0.1/250 and a 25uf/25 on the DC-30 for example) on the first gain stage? Is it just to bleed off high frequencies? Because the first 12AX7 is paralleled? Both?

Ta.


Edited to add, the typical Blackface circuit uses a dedicated cathode cap and resistor for V1 triode 1 but shares a cathode cap and resistor for V1 triode 2 with V2 triode 2. If you remove the second channel and build a 1 channel BF amp, is it best to have a dedicated cathode cap and resistor for V1 triode 2 or can it be shared with V1 triode 1?


As jojokeo pointed out below, if one 12ax7 is paralleled (like the DC-30 example) the cap and resistor value are effectively halved. So, when you parallel, shouldn't you raise the cap and resistor value so the response is similar to a single triode (to retain the Vox feel for example)? And when you build a single channel BF, shouldn't you share the cathode V1(T2) with the simulated equivalent V1(T1) as it was originally shared with V2(T2) to retain the BF feel?

[jojokeo]

Here's a non-technical answer; a .1uF  poly-type non-polar cap will do a better job of bypassing the upper frequencies than the much larger polar e-cap? This may be more noticeable when two stages are paralleled like that also? Guessing it will give a smoother less grainy type of response when driven hard? Know that when the tube is paralleled like that the effective load & cathode resistance value of each stage is effective half the value used. So in the schem each stage is effectively a 110k / 12.55uF biased stage.

[fiftynine]

Here's a non-technical answer.

Sounds pretty good to me.

[Willabe]

Hi-Fi guys like to use small poly caps on tube K's and on the B+ filter caps to bypass the larger E-caps. (I think the Hi-Fi rule of thumb is ~10% of the E-caps uF value for the poly bypass cap.)

Some guys use just a low voltage poly cap for the K bypass without the E-cap. 

Some say they can hear a slight difference, others say they can't. They say it's a little smoother, less grit and a little faster in response to picking.       

[Merlin]

It has no effect. It is probably a habit copied from other realms of electronics where putting a plastic or ceramic cap in parallel with an electrolytic decoupling cap does have an effect, because the smaller cap does all the work at high (above audio) frequencies where the e-lytic has too much self inductance. In a valve amp, espcially a guitar amp of all things, it's just another way to spend an extra 10 cents. Audiophools often do the same thing with coupling caps in hi-fi amps thinking somehow the high freqiencies will 'go around' the horrible e-lytic. They don't. There is no improvement at audio frequencies.

NB: If there is a switch to select one capacitor or the other, that's different.

[jojokeo]

"Audiophools..."

 

[jjasilli]

It has no effect. It is probably a habit copied from other realms of electronics where putting a plastic or ceramic cap in parallel with an electrolytic decoupling cap does have an effect, because the smaller cap does all the work at high (above audio) frequencies where the e-lytic has too much self inductance. In a valve amp, espcially a guitar amp of all things, it's just another way to spend an extra 10 cents. Audiophools often do the same thing with coupling caps in hi-fi amps thinking somehow the high freqiencies will 'go around' the horrible e-lytic. They don't. There is no improvement at audio frequencies.

NB: If there is a switch to select one capacitor or the other, that's different.

It sounds different to me, so I do it. 

[fiftynine]

Edited to add, the typical Blackface circuit uses a dedicated cathode cap and resistor for V1 triode 1 but shares a cathode cap and resistor for V1 triode 2 with V2 triode 2. If you remove the second channel and build a 1 channel BF amp, is it best to have a dedicated cathode cap and resistor for V1 triode 2 or can it be shared with V1 triode 1?


As jojokeo pointed out above, if one 12ax7 is paralleled (like the DC-30 example) the cap and resistor value are effectively halved. So, when you parallel, shouldn't you raise the cap and resistor value so the response is similar to a single triode (to retain the Vox feel for example)? And when you build a single channel BF, shouldn't you share the cathode V1(T2) with the simulated equivalent V1(T1) as it was originally shared with V2(T2) to retain the BF feel?

[sluckey]

Edited to add, the typical Blackface circuit uses a dedicated cathode cap and resistor for V1 triode 1 but shares a cathode cap and resistor for V1 triode 2 with V2 triode 2. If you remove the second channel and build a 1 channel BF amp, is it best to have a dedicated cathode cap and resistor for V1 triode 2 or can it be shared with V1 triode 1?

Use separate cathode R/C. Do not share cathode R/C with two consecutive gain stages. You may just create an oscillator.

As jojokeo pointed out above, if one 12ax7 is paralleled (like the DC-30 example) the cap and resistor value are effectively halved. So, when you parallel, shouldn't you raise the cap and resistor value so the response is similar to a single triode (to retain the Vox feel for example)? And when you build a single channel BF, shouldn't you share the cathode V1(T2) with the simulated equivalent V1(T1) as it was originally shared with V2(T2) to retain the BF feel?

The reason you half the cathode resistor value when two similar gain stages share that resistor is to maintain the same bias point.

For example, a single 12AX7 triode with 100K plate resistor and 1.5K cathode resistor will have about 1mA flowing through it. This current develops a 1.5V drop across the cathode resistor. Since the grid operates at zero volts, this 1.5V is the bias point for that triode. And that's a happy bias point for that tube. Now if you tack on another identical triode but share that same 1.5K cathode resistor, each tube will still flow about 1mA for a total of 2mA flowing through the shared 1.5K resistor. Now the shared resistor drops 3V and this becomes the bias voltage for each tube. That's not such a happy bias point. So we half that resistor value. It should be 750Ω but 820Ω is the closest easily available value. Now the 2mA flowing through the shared 750Ω will cause 1.5V to be dropped, making the bias point for each tube happy again.

[2deaf]

As jojokeo pointed out above, if one 12ax7 is paralleled (like the DC-30 example) the cap and resistor value are effectively halved. So, when you parallel, shouldn't you raise the cap and resistor value so the response is similar to a single triode (to retain the Vox feel for example)?

When you strap the triodes in a tube as they did in the Matchless DC30, the internal resistance (plate resistance) drops to something on the order of one half for the pair.  It does not drop exactly in half because the overall DC current changes which changes the bias point and the plate voltage which changes the operating point resulting in a different internal resistance for the individual triodes.  But for the sake of argument, let's say the internal resistance drops in half when strapped.

With the application of Ohm's law, the output impedance drops when strapped, but not nearly as much as one half.  With a constant B+ the DC current for the whole circuit increases a little bit when strapped.  The current through each individual triode decreases when strapped which makes it look to me as though the individual triode is seeing a larger resistor connected to the plate.

Anyways, if you want a strapped pair to act like a single triode, you need to half the cathode resistor and the plate resistor and double the cathode bypass cap. 

   

[jjasilli]

To clarify, I do it at the last filter stage.

[fiftynine]

Cool. I reckon the next question is, if you were switching between a paralleled triode tweaked sluckey/2deaf style, a DC-30 as is and a single triode Vox style, what's the difference in tone and feel?

[sluckey]

About 50 cents.

[fiftynine]

That made me laugh.

[sluckey]

Good. That was my intent.

[Willabe]

Why do some amps have 2 cathode caps in parallel (0.1/250 and a 25uf/25 on the DC-30 for example) on the first gain stage?

I tried a poly 6.6uF @450v cap, it measured 7.2uF, for the 1st gain stage K bypass by it self, No E-cap, on my bread board Champ set up today. I had it on a switch with a 10uF E-cap to make it easy to hear any difference.

There was some difference to my ears, a little smoother. The difference I could hear was on the hi end only.