cathode bypass and grid stoppers/ when where and why

[ric]

Looks like there's a consensus here that the first stage cathode should be fully bypassed to minimize heater noise. Are we talking 22uf on a 100k plate 1k5 cathode? Are we saying smaller cap values are still fully bypassing as long as no resistor is in series with the cap?

Also the grid stop question of how much and why, if at all.

As a jumping off point I'll mention from my own experiments with 3 gain stage to concertina and 4 gain stage with reverb to LTP I liked 22k single input stopper, 47k second stage and 22k the rest of the way. With the cathode bypasses varying from adjustable up to 220uf to none.

I'd like to know what the prevailing wisdom around here is, and why.

[eschertron]

Be sure to understand that these are two separate engineering issues, so they can be discussed apart from each other without confusion.

I'm in the 'bypass the cr@p out of the first stage' camp. 22u takes the 1.5k cathode resistor's knee frequency to just about 5Hz. I have used some 10u/50v caps because they were available, the knee freq is about 10Hz. Probably still OK.

What about forward-biased LEDs? Anyone have a sense of how this topology resists heater hum?

[Chuck H]

Fully bypass the first gain stage cathode. If you need to mitigate LF from the first stage you should use the coupling cap. If you need to mitigate LF in the overall preamp you can usually partially bypass subsequent gain stages without notable consequence. The caveat here is that there is a big difference between vintage preamps, amps that use overall overdrive to achieve distortion and modern preamp based distortion generators that typically run the power amp clean. For the later it's best to keep the first two stages fully bypassed or just use DC for the filaments and partial bypass as you like on any stage.

[loudthud]

A Cold Clipper stage will act badly if the cathode is bypassed when that stage is hit by a big signal. The bias will want to move when a signal is present and may be slow to recover between notes. An unbypassed cathode will recover instantly.

[nsubulysses]

What would you consider a cold clipper stage to be, maybe anything higher than 3.3K cathode resistor, or maybe 10K?

I have been messing around a little with what seems like ampeg style component values where it's something a little weird, something like 6.8K cathode resistor and 150-470K plate resistor.

What is the advantage or difference with this compared to just doing 1K cathode 100K plate resistor which is way more typical?

[eschertron]

As the anode R goes up, the cathode R can go up too. Looking at a load line chart will help identify what's too far from center. Marshall's 100k anode, 10k cathode, no bypass cap (2203/2204) is a good example of cold clipping in action.

[Malcolm Irving]

I think there are two different types of heater noise (maybe more). One type occurs outside the tube, and is caused by unwanted coupling between the heater circuit and the signal circuit. This will be affected by cathode bypass capacitors in the usual way (low-cut shelving filter).
The other type occurs inside the tube and is caused by unwanted leakage current from the heater to the cathode. This will be affected by a cathode bypass cap in a different way. A big capacitor in this case will short-circuit the ac hum voltage to ground and prevent it being amplified by the tube. I think in this case there is no shelving effect, and it’s actually a high-cut filter (with regard to the heater noise getting into the signal), albeit with big attenuation at 50/60Hz. A very high valued bypass cap, such as 200uF, can then have a beneficial effect on this type of hum/buzz.

[Malcolm Irving]

Also the grid stop question of how much and why, if at all.

The grid stopper on V1 is usually the main source of background hiss in an amp. Reducing the resistor value reduces the hiss level, but the lower you go the more chance there is of picking up radio stations. (Could be either humorous or embarrassing in a gig situation!)

A grid stopper of around 10k can be a good compromise. An additional small cap from grid to ground can also be introduced to allow a small grid-stopper while retaining the required high cut filtering to block radio interference.

[ric]

Malcolm, what range of sizes do you use for input cap-to-ground?

[Malcolm Irving]

I would say about 300pF. The theory and technical details are covered in Merlin's book (which is highly recommended):
'Designing Tube Preamps for Guitar and Bass', Second Edition, Merlin Blencowe

[pdf64]

It goes horribly against the grain with me to add capacitance to the input; fine with the guitar at full volume, but it would surely muddy the tone at mid settings, ie when the equivalent series impedance of the source is >100k? Assuming a 'plug straight in' no pedals scenario.
Also not an issue for the players who just use the instrument volume as a mute, sometimes even replace it with a switch.

But for the players such as myself, who spend time and money sourcing low capacitance guitar cables in order to minimise the 'muddy at mid volume' thing, it seems counter-productive to then add capacitance at the amp input.

[Mike Sulzer]

I would say about 300pF. The theory and technical details are covered in Merlin's book (which is highly recommended):
'Designing Tube Preamps for Guitar and Bass', Second Edition, Merlin Blencowe

A filter such as you describe is most effective if the interfering signal is coming in on the center conductor since the R and C act together in this case. But it is the shield of the cable that picks up most of the interference. The C still has some effect, of course. Guitar amps usually have jacks that are isolated from the chassis. This is what you need to prevent ground loops resulting in power line buzzing, but it leaves you open to rf on the shield. One solution is to put a .01 microf cap from jack ground to the chassis, making the jack and chases at nearly the same rf potential, while leaving them isolated at the frequencies of significant power line harmonics. Post #6 here (https://www.thegearpage.net/board/in...ignals.281178/) describes this. Ferrite might be useful if used properly, like those lumps you see on some USB cables.

[ric]

It goes horribly against the grain with me to add capacitance to the input; fine with the guitar at full volume, but it would surely muddy the tone at mid settings, ie when the equivalent series impedance of the source is >100k? Assuming a 'plug straight in' no pedals scenario.
Also not an issue for the players who just use the instrument volume as a mute, sometimes even replace it with a switch.

I'm picking up radio on one amp and plan to include this feature on future builds. I would, however, make it switchable.

A friend and trusted source of info who retired from the local university after a career of building and working on tube amps there suggested no more than .001. 300p seems quite reasonable and conservative.

Guitar amps are a special case, and it's cool to be able to get all these different perspectives from people all over the planet that build and play them. For my specific guitar related questions I used to go to the local tech, some guy who called himself "Enzo". He really seemed to know quite a bit and was willing to answer questions.

[ric]

A filter such as you describe is most effective if the interfering signal is coming in on the center conductor since the R and C act together in this case. But it is the shield of the cable that picks up most of the interference. The C still has some effect, of course. Guitar amps usually have jacks that are isolated from the chassis. This is what you need to prevent ground loops resulting in power line buzzing, but it leaves you open to rf on the shield. One solution is to put a .01 microf cap from jack ground to the chassis, making the jack and chases at nearly the same rf potential, while leaving them isolated at the frequencies of significant power line harmonics. Post #6 here (https://www.thegearpage.net/board/in...ignals.281178/) describes this. Ferrite might be useful if used properly, like those lumps you see on some USB cables.

Mike, this brings up something else I wonder about. An isolated input jack leaves the guitar ground floating from earth ground. We want the amp chassis unearthed for safety, wouldn't we want the guitar grounded for the same reason?

[Enzo]

For my specific guitar related questions I used to go to the local tech, some guy who called himself "Enzo". He really seemed to know quite a bit and was willing to answer questions

.

Thank you for the kind thoughts. Now retired, I am trying to pass along the accumulated lore of the trade to local fellow RJ.