Cathode By Pass Cap Question

[SilverFox]

Referring to Merlins Book I've got a question regarding the cathode bypass cap.

If a guitar frequency range cuts off around 80 Hz, what good is placing a capacitor value any larger then 5uf on the cathode? I've seen and used values as high as 330 on the JTM 45 design and I see the Sunn Sceptre uses a 250 on the input.

What is the purpose of so large a capacitor?

silverfox.

[David Root]

It doesn't "cut off", there can be subharmonics, and the bypass cap just produces a shallow 3dB per octave drop in response. So, if it's 3dB down at 80Hz, that is 6dB down at 40Hz, and so on.

Increasing the cap size shifts the initial slope to lower frequencies, leaving more bass apparent. These days 250 or 330uF is considered overboard, but can still be managed further down the line if done on the first stage.

Historically those high value bypass caps trace back to the '59 Bassman 5F6A, which was intended as a bass amp, thus the big cap. Turned out it was a poor design for bass but great for guitar. The 4x10" speakers and open back cabinet were partly responsible, along with Fender's deliberate bass attenuation further down the line because the small speakers and deliberate undersizing of the OT mitigated against full bass reproduction. ie down to 40Hz.

[xtian]

David's answer is good, but I hear you asking what good is using a large value bypass cap? Because you get a boost in gain. Also will help decouple one triode from the other in the same glass package.

[Firestorm]

I once read another reason for the overkill, but can't remember it. Maybe current stability or the imprecise nature of electrolytics.

[Malcolm Irving]

One other reason is that any noise on the cathode coming from the heater will be well shunted to ground by the extra large cap.
I think that might only apply to tubes with a heater-to-cathode problem. I did a few tests by switching on a big cap in parallel with the usual small cap and I couldn't notice any change in the noise level.

[Firestorm]

One other reason is that any noise on the cathode coming from the heater will be well shunted to ground by the extra large cap.
I think that might only apply to tubes with a heater-to-cathode problem. I did a few tests by switching on a big cap in parallel with the usual small cap and I couldn't notice any change in the noise level.

I think once you get well enough above the 60Hz breakpoint, the heater hum is decoupled. In split cathode Marshalls, some tubes will hum on the .68 side. Others won't.

[tubeswell]

Having as much frequency bandwidth boost as you can in V1 helps with getting a good S:N ratio. (All the better to be able to manipulate frequencies with later in the signal path).

[Leo_Gnardo]

If a guitar frequency range cuts off around 80 Hz, what good is placing a capacitor value any larger then 5uf on the cathode?

Lots of good answers, including "Tradition . . . . tradition!" but IMHO you're right - with 5 uF across a 1500 Rk, plenty enough low end for guitar and hum reduction in pre tubes that have that problem. Beyond there, you don't really need to go.

[SilverFox]

Yeah these are very usable replies. I'm starting another build and will probably use the first two stages out of the Dumble circuit. Unfortunately if the last build had been a medical patient I would have thrown in the towel and walked out of the operating room. The guts were in a mess and the patient was in v-fib! As a Medicare patient it would have paid $97.62 but in this cae that's about what it cost me, plus time.

Dr. Fox.

[pdf64]

..the bypass cap just produces a shallow 3dB per octave drop in response. So, if it's 3dB down at 80Hz, that is 6dB down at 40Hz, and so on...

Just to note that cathode bypass is a shelving type filter, rather than roll off; so the bass cut from it can't fall below the lower (unbypassed) shelf.
With regular circuit values, the unbypassed shelf may be ~6dB below the bypassed shelf; that shelf response will carry on down to 0Hz.

ECC83 info (eg Philips) refers to the hum and noise spec requiring the cathode to be sufficiently decoupled http://www.mif.pg.gda.pl/homepages/fran ... /ECC83.pdf

Also to note that when calculating the frequency breakpoint, the 'R' term should take account of the tube's equivalent cathode resistance, which is in parallel with the cathode resistor. From Aiken, it seems that the plate load resistor affects that! http://www.aikenamps.com/index.php/desi ... amplifiers

[bal704]

Here's an article I bookmarked on the subject:

http://el34world.com/Forum/index.php?topic=17617.0

[Structo]

Seems like some lower powered guitar amps used larger cathode bypass caps on the first gain stages.

I can't remember the exact number but a 25uF cathode bypass cap is said to fully bypass the cathode.

So it is letting all frequencies through. (although the tube capacitance does affect this)

Many amps have fully bypassed first stages.

Later on, some amps attenuated this later in the gain stages

While other amps seem to throttle the frequency response early on.

Tube amps are like a box of chocolates.........

[Firestorm]

One of the common myths is that changing the cathode resistor changes the gain of the stage. This only true if the resistor is not bypassed. What this change does do is shift the operating point of the tube so that the signal might be clipped at the top or the bottom of the swing (or centered). There is a lot to be gained sonically from these manipulations.

[bal704]

The basic gist of the article I linked is that the large cap reduces hum.

The "heater-to-cathode capacitor" and the oversized cathode bypass cap can be viewed as a capacitive voltage divider. The source voltage is the heater a.c. voltage; the "heater-to-cathode cap" is a very small capacitance and is in series with the cathode bypass cap, which completes the circuit from cathode to ground. The cathode itself is the mid-point, or output, of this capacitive voltage divider. In a resistive voltage divider, you reduce output voltage by making the resistance from output to the ground lower. In a capacitive voltage divider, you do the same by making the reactance from output to ground lower, which is done by making the cap value larger.

The heater-to-cathode cap has some very-low capacitance, so it has a high reactance at 60Hz. At 60Hz, the 25μF cap has a reactance of 106Ω; going 10x bigger to the 250μF reduces the reactance by 1/10th to 10.6Ω. And cuts the hum output.

[SilverFox]

I read Bluetubes post on EL34 and found it most informative. I've read some of his posts in the past and found them to be technically informative and a reliable source of information. In light of the hum reduction benefit of the larger capacitor and not mass producing the next build I figure why not, (as long as it fits the frequency response profile). Thanks for the link.

Varying the cathode resistor to change the bias point is also good to know. Although I've read it before and even recently, that post made it much more applicable.


silverfox.