cathode bypass and grid stoppers/ when where and why

[Malcolm Irving]

... . 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?

The amp chassis should be solidly connected to mains earth for safety. The chassis should also be connected to signal ground to help with screening (at one point only - to avoid a ground loop). The best place to make the connection between signal ground and chassis is close to the input jacks, IMHO. This would be via a dedicated wire to a solder lug on the chassis (i.e. the jack socket itself is still isolated).

A lot of people recommend grounding at the output stage, but I disagree with that for the following reason: The non-zero resistance of any signal grounding scheme means that a small ac voltage will exist between the chassis and the signal ground at the opposite end to where the connection is made. I would rather have that small ac voltage at the power output stage where it will have negligible effect, rather than near the input stage where it will be amplified if it is linked into the signal path via stray capacitance.

[Malcolm Irving]

I suppose ‘signal ground to chassis close to the input jack’ only applies in conjunction with the ‘daisy chain of local star points’ method of grounding.
If people prefer ‘a single star point’ for the whole amp, that may work better with the star point at the power output stage. Otherwise, there would be some long ground wires carrying buzz current all the way to near the input jack and potentially linking into the signal path by stray coupling.

[ric]

.

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

You let me take up many hours of you're shop time over a year or two.

I also remember my cousin (also a tech) calling me back about voltage rating on a switch as I'm checking out at the hardware and telling him:"Thanks, Enzo already called back and told me that". And you pointed me at Ampage where "I answer these kind of questions". Yes you do.

[ric]

Anyone care to weigh in on an isolated input jack leaving the guitar's ground floating, not at earth ground?

[Enzo]

Chassis should be at earth, What we usually call "ground" is really the circuit common. On many amps those are the same, on many others it is not. For example many Marshall amps have a ground common to chassis link that is a 10 ohm resistor, a cap and a pair of diodes all in parallel.

But in all cases the earth and ground common are linked together, as Malcolm said. So the guitar is not floating, it is just not directly connected to chassis at the input jack. In some places the isolated jack ground is linked to earth (chassis) right at the jack, but through a link with a ferrite bead or some such. In other cases, the jack ground is connected to common which trails over to the power supply, where it is earthed to chassis.

So, no float, Pepsi.

[Mike Sulzer]

The amp chassis should be solidly connected to mains earth for safety. The chassis should also be connected to signal ground to help with screening (at one point only - to avoid a ground loop). The best place to make the connection between signal ground and chassis is close to the input jacks, IMHO. This would be via a dedicated wire to a solder lug on the chassis (i.e. the jack socket itself is still isolated).

A lot of people recommend grounding at the output stage, but I disagree with that for the following reason: The non-zero resistance of any signal grounding scheme means that a small ac voltage will exist between the chassis and the signal ground at the opposite end to where the connection is made. I would rather have that small ac voltage at the power output stage where it will have negligible effect, rather than near the input stage where it will be amplified if it is linked into the signal path via stray capacitance.

I do not think the issue is stray capacitance at 60 (50) Hz and significant harmonics. A connection between the low side of the supply (which has a transformer) and the chassis carries some current (blame it on Faraday). The voltage from the guitar should be the only voltage across the input of the first stage, which is the grid and the bottom of the cathode resistor/bypass capacitor. If the input jack is the connection of the power supply to the chassis then that current consisting of power harmonics flows in the wire connecting the low side of the jack to the low side of the first stage input, and thus makes a voltage in series with this input at the most sensitive point in the amp.

It seems that some really smart people can make that current very small with proper layout, etc. I prefer to be safe and make that connection at a high level point, preferring the center tap of the transformer itself, which its essentially the same as the output stage.

As discussed earlier, the jack isolated from the chassis is not good for rfi rejection. But there are ways around that.

[pdf64]

Anyone care to weigh in on an isolated input jack leaving the guitar's ground floating, not at earth ground?

Malcolm wasn't proposing that the guitar's ground be floating; in all cases it would be referenced to ground, either directly by means of a conductor (wire, chassis etc), or via a low value resistor//cap//diodes, the latter to mitigate for ground loop hum.
But really it's safer if the guitar (and hence the guitarist) isn't grounded.
If someone is playing out a lot, especially in venues where the electrical system may be less than ideal, then use of a wireless system is excellent electrocution mitigation.

[Malcolm Irving]

... . A connection between the low side of the supply (which has a transformer) and the chassis carries some current (blame it on Faraday). ... .

Can you elaborate a bit on that? Are you thinking of stray capacitance between the transformer winding and the core/frame and hence to the chassis? Or eddy currents induced somewhere?

[ric]

Can you elaborate a bit on that? Are you thinking of stray capacitance between the transformer winding and the core/frame and hence to the chassis? Or eddy currents induced somewhere?

It sounded to me like Mike is addressing what you mentioned in post#18, PA ground at the back end, only V1 grid and cathode/ bypass cap across input. Only at another angle/ perspective. I've been known to be wrong and I'm sure he can speak for himself, but that's what I understood.

[Malcolm Irving]

I’m sure Mike is right that there is some current between the PT secondary 0V side and the chassis, but I am not clear about the closed path which that current takes and how significant that current is.
If the connection from the PT secondary 0V and the chassis is the only such connection, we need to find some closed loop path for the current to return in. This path could be through inter-winding capacitance in the PT via the mains neutral and then the mains earth (which are connected somewhere by the power company). Or it could be via stray capacitance and/or insulation leakage to the transformer core/frame and hence to the chassis.
I don’t think that current would be a result of magnetic induction (i.e. Faraday's Law) but maybe I’m wrong.

[Malcolm Irving]

I’ve made a quick experiment. I have an amp which hums a bit more than I would like (100Hz hum). It’s a 2xEL84 class AB push-pull, with a 12AX7 LTP PI and an EF86 input stage. It has silicon rectification.
I tried 4 options for connecting the chassis to signal ground:
(1) near the input jack,
(2) at the 0V of the reservoir cap,
(3) both of the above (not recommended as it creates a ground loop),
(4) both of the above, but with the ground daisy chain broken between the LTP and output stage (i.e. separate pre-amp and power-amp grounds to chassis).
To cut a long story short, all 4 options gave me the same 23mV ac at the 8 ohm load, with nothing plugged in and the volume control at max.
The hum level is acceptable and is subjectively about the same as the hiss level. The input grid stopper is 10k, but an EF86 is known to be noisier than a 12AX7.
I suspect that the hum could be reduced by using better matched EL84s and/or replacing the 47uF reservoir with 100uF.

[Malcolm Irving]

OK. Tried increasing the reservoir by tacking in another 47uF in parallel with the existing one.
Output level (hiss plus hum) at full volume now reduced from 23mV to 13mV.
Still no measurable difference between options (1), (2) and (4). (Didn’t bother with option 3 this time.)

[ric]

OK. Tried increasing the reservoir by tacking in another 47uF in parallel with the existing one.
Output level (hiss plus hum) at full volume now reduced from 23mV to 13mV.
Still no measurable difference between options (1), (2) and (4). (Didn’t bother with option 3 this time.)

Aluminum or steel chassis?

[Malcolm Irving]

Aluminum or steel chassis?

Aluminium.

[Mike Sulzer]

OK. Tried increasing the reservoir by tacking in another 47uF in parallel with the existing one.
Output level (hiss plus hum) at full volume now reduced from 23mV to 13mV.
Still no measurable difference between options (1), (2) and (4). (Didn’t bother with option 3 this time.)

I agree with your analysis that leakage such as capacitance at the transformer would be necessary for the current I am hypothesizing. But nothing does it like an experiment! Some difference might be noticeable with a quieter first stage, but you have shown that it cannot be a big difference, if any at all.

If the hum is due to poor balance between the El34s, should it not be just as audible with the volume all the way down?

When you make the cap bigger, you are reducing the hum at all the successive stage of the RC filtering (if the circuit is the usual one or a variant), and so the question of where the hum is introduced is still open, I think.