Why shouldn't the Mains Ground connect to the Reservoir cap negative?

[christarak]

I'm working on a Marshall 2203. I understand how Marshall has muddled up the overall grounding scheme, but I don’t understand why the general advice is to give the “Mains to chassis” ground its own independent chassis connection and not be common with the chassis Negative connection of the first reservoir multicap can & power supply? I would have thought it would still provide an uncompromised safety ground to the chassis as well as connect to that first (ideally star) chassis ground from a hum perspective. What have I missed?

[g1]

The chassis safety ground is for safety only. It must be independent and should be the last wire connected if the AC cable is ripped from the unit.
If it only serves one purpose, there is less chance of it getting disconnected by accident during service of other parts. For example, using an existing transformer bolt/nut is not considered acceptable.

[Chuck H]

This ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

It might also help to know that an amp will operate and can even be hum free without an earth ground. The 0V circuit reference ground for signals and amplification is it's own element. In fact it doesn't even have to be 0V if all other circuit parameters are made relative. The earth ground is a different circuit, sort of. But because both "grounds" are (typically) 0V and both (typically) use the chassis people tend to think they are the same thing. Even though they really aren't.

[christarak]

For example, using an existing transformer bolt/nut is not considered acceptable.

Aha! I hadn't considered the accidental service disconnection possibility. Yes, I ensure that its the longest of the 3 mains wires. With the owner's permission, I usualy remove the ridiculous nut'n'bolt of mains earth and make a secure solder connection to the chassis. Thank you for taking the time to reply, G1.

[christarak]

This ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

The earth ground is a different circuit, sort of. But because both "grounds" are (typically) 0V and both (typically) use the chassis people tend to think they are the same thing. Even though they really aren't.

I think this is where I got confused, Chuck. When you see the Mains earth and the circuit's earth both connected to chassis often only an inch apart, I could not see why they couldn't be joined. at the same point. Thanks Chuck.

[Chuck H]

I've used the safety ground for the first filter ground in early builds with the erroneous belief that somehow that could be more ideal. Shorter path taking ripple and hum OUT of the amp, right? Wrong. There's no harm in it circuit wise. But it doesn't actually help anything and it compromises the safety ground if any tech needs to alter the connection. So I had the belief that they were the same thing too. It was only after years of reading and re reading ground related information here and elsewhere that their independence started to sink in. And it had more to do with understanding circuit grounds and how current moves through them. When I was a lot greener I thought that current made it's way out to the earth. Shaking the speaker on it's way there. In actuality things are moving the other direction and electrons are drawn from earth. Even though as I said before you don't actually need an earth for amplifier operation. I expect this is because the electron supply from the AC neutral is ample. I still haven't got it fully sorted but I'm closer.

[christarak]

Do you have any idea as to why there is a 22ohm resistor (and D6&D7) between the chassis ground and the circuits star earth in the JCM900 schematic? Its related to why the 2 grounds are not the same...but I can't see the reason for it.

[Chuck H]

Well it's unclear just where the connections are because of line gaps in the image. But the short answer is no. I've seen the small resistance elevation in other amps too and I still don't understand it. On any simulations I've attempted on CAD programs it seems to cause instability. The parallel capacitor value of 100n seems anemic across a 22r resistor for AC bypass. This is still beyond my pay grade. Hopefully someone else will see this and chime in because I'm interested too.

[christarak]

Well it's unclear just where the connections are because of line gaps in the image. But the short answer is no. I've seen the small resistance elevation in other amps too and I still don't understand it. On any simulations I've attempted on CAD programs it seems to cause instability. The parallel capacitor value of 100n seems anemic across a 22r resistor for AC bypass. This is still beyond my pay grade. Hopefully someone else will see this and chime in because I'm interested too.

Thanks a lot, Chuck. If no one sees the question in the next day or 2, I might post it as a new question.

[g1]

Do you have any idea as to why there is a 22ohm resistor (and D6&D7) between the chassis ground and the circuits star earth in the JCM900 schematic?

Apparently to deal with ground loops plus a safety backup: https://music-electronics-forum.com/...438#post439438

[Chuck H]

Wow! Fat reading for tonight (I'm already have two beers in ) but any time you get R.G. Keen and Randall Aiken on the same grounding discussion it's worth perusing sober. So thanks for this.

For any one not familiar with soundguruman you can skip any reference to him or responses to his replies. They don't matter. Just focus on what the big cats are saying

Clearly this elevated ground to avoid loops is developed in situ.?. You pretty much wear a white lab coat if you're conceptualizing or designing with this circuit. In my experience elevating circuit grounds, on the decoupling caps as well as shown in the Marshall schematic, causes instability. But it's been practiced by a lot of well respected manufacturers that have futsy, house hiding nerds figuring these things out at the bench. The best we can do is try to keep up. I'll be reading it tomorrow with more clarity but I make no promises about how much I'll learn. Too often the deepest understandings go over my head even on good days.

[christarak]

Wow! Fat reading for tonight (I'm already have two beers in ) but any time you get R.G. Keen and Randall Aiken on the same grounding discussion it's worth perusing sober. So thanks for this.

For any one not familiar with soundguruman you can skip any reference to him or responses to his replies. They don't matter. Just focus on what the big cats are saying

Clearly this elevated ground to avoid loops is developed in situ.?. You pretty much wear a white lab coat if you're conceptualizing or designing with this circuit. In my experience elevating circuit grounds, on the decoupling caps as well as shown in the Marshall schematic, causes instability. But it's been practiced by a lot of well respected manufacturers that have futsy, house hiding nerds figuring these things out at the bench. The best we can do is try to keep up. I'll be reading it tomorrow with more clarity but I make no promises about how much I'll learn. Too often the deepest understandings go over my head even on good days.

I'm with you, Chuck. I've got Merlin Blencowe's Preamp Book, and yesterday I ordered his new edition of Power Supplies. By the time I have fully absorbed those, I will either have a propeller sprouting from the top of my head, or I will be locked up in "Happy Valley". We'll compare notes.

[christarak]

Apparently to deal with ground loops plus a safety backup: https://music-electronics-forum.com/...438#post439438

Wow!. Thanks G1. That link has answered questions that I haven't even thought of yet. Well, that takes care of what's left of my Sunday!

[pdf64]

... With the owner's permission, I usualy remove the ridiculous nut'n'bolt of mains earth and make a secure solder connection to the chassis...

Why ridiculous?

My understanding is that whilst electrical safety regs dont preclude a soldered safety earth connection to a metal chassis, it needs to be combined with a mechanical fastening. eg crimped, and then the crimp eyelet soldered to the chassis. Then the eyelet also screwed to the chassis.
A connection where solder is the only means by which the wire is attached to the chassis doesn’t comply.

Reason being that fault current can cause solder to melt, and the wire may become detached.

A dedicated connection acts to eliminate dual use of chassis fasteners, the other function of which might put the fastener under significant mechanical stress. A prime example of that being a mains transformer mounting bolt, which seems to be the default choice of DIY amp builders.
Fasteners that go through a transformer lamination stack are particularly unsuitable, as the stack can compress over time / vibration / AC magnetostriction, allowing the fastener to become loose.

A dedicated connection also helps to reinforce the point that the safety earth is about the only one in an amp that’s safety critical.

[Chuck H]

Well I do most of my work in aluminum chassis so it's always a bolt for me. I always use star washers and goop the connection with the anti oxide paste on the premise that it can't hurt. So far no reports of failed grounds. Though I didn't know the code was a crimp I always did it since that's how the terminal attaches anyway. Crimped and soldered.