star ground for power tubes and ct, or soldered to chassis?

[Mick Bailey]

You stressed grounding is ALL about eliminating noise (your stress on ALL). Now you back-pedal when I suggest other factors by dismissing them. You obliquely refer to yourself as an amp designer, but "We don't talk V=IR here" Well you need to, because that's what everyone else does. Who is this "We"? it clearly doesn't include Me.

From your other posts there are such glaring omissions in your practical understanding of tube amps that you come over as an overly-defensive armchair theorist who's built a single amp and wants to tell the world 'The Way It Is According To Alan'. On the one hand you repeatedly tell us that you've spent decades in multi-layer board design amongst a sea of PhDs, and on the other it's clear you can't tell your shoes from your feet. That's fine, but don't try to be dismissive of others just because you find yourself in a corner.

[Alan0354]

You stressed grounding is ALL about eliminating noise (your stress on ALL). Now you back-pedal when I suggest other factors by dismissing them. You obliquely refer to yourself as an amp designer, but "We don't talk V=IR here" Well you need to, because that's what everyone else does. Who is this "We"? it clearly doesn't include Me.

From your other posts there are such glaring omissions in your practical understanding of tube amps that you come over as an overly-defensive armchair theorist who's built a single amp and wants to tell the world 'The Way It Is According To Alan'. On the one hand you repeatedly tell us that you've spent decades in multi-layer board design amongst a sea of PhDs, and on the other it's clear you can't tell your shoes from your feet. That's fine, but don't try to be dismissive of others just because you find yourself in a corner.

Back pedal what, this thread is about Fender grounding, it's about noise. read post #69. I just report. Obviously you want to start something that I am not interested. You want to talk ground, talk. Cut the bull.

And yes, I don't know much of tube amps, but this has everything to do with noise, nothing to do with sound. I stand by what I said here. You don't want to accept new things, it's your loss. Yes, electronics gone through quantum leaps and tubes stay in time in the bottle. Noise has been well studied in the last 40 years. Grounding has been well studied in the last 20 years. Books talking about this subject. What I said here is VERY VERY BASSIC that every signal integrity engineer must know.

I think I said my piece, you can keep the tubes and the old theory.

[Tooboob]

Alan, if you go back in time, you'll note that Mick admitted at the start that if he's cloning a Fender, he grounds like you and Leo, though he does star grounding with his own designs.

I'd like to know what happens to your noise floor when you plug in a signal with an earth-grounded shield from some external processor, creating an earth ground loop. You are also making lots of noise measurements, subjectively or objectively. What is your acceptance criteria? Do you have one? This is America, and if you like to spend your day minimizing noise in a one-off tube amp, hey, more power to you, but have you set a goal that will allow you to stop?

I have to worry very much about grounding. I do graphics output circuits. Yes, the science is well established, and there are still superstitions based on results and trade-offs.

Let's say that you're being hip, sending multi-GHz signals 30 feet down a cheap cable differentially. They start out differential on the board. You can route the signals comparatively close together and far from the ground plane, or vice-versa, and still wind up with 100 Ohms. You're fairly safe either way. Now you get to the output connector, and make the jump to the cable. It may have a more ground-influenced impedance, but it probably has less. Let's say it has less. The return currents in the ground change at the interface, along with the return currents in the opposite polarity inductor. We're 100 Ohms on either side. What happens at the connector? 100 Ohms on both sides. Are we done? I can assure you that most designers never worry about it. I'm wondering why not.

Not only is grounding just about noise (and I don't think that that's what you meant), noise isn't just about grounding. You point out grid-stop shot noise. For me, caps are out there with grounding - value, placement, and parasitics. I could probably re-locate one cap in your amp and resurrect much of the noise.

When we first started trying to get computers to stop radiating RF in the early '80s, the Tempest guys came out of the woodwork teaching seminars. Single-point ground to the chassis, seal it up, and the problem is solved, they said, and it worked great until we hooked up earth-grounded peripherals, providing a big loop for return currents, and an excellent antenna. We were failing testing at $1500 per day. Ground planes helped a bunch, but we started cutting them up to isolate noisy or noise-sensitive circuitry, and the problem got worse. Isolated grounds and judicious ground plane isolation and cutting has its place, but you really have to be the electrons (or in your case, be the fields) and understand what's going on.

If you star-ground, the resistance in the ground wires causes noise, and you still have loops that can pick up things like transformer noise where signals move between circuits grounded by different branches of the star. With your chassis grounding method, the ground points move apart, but the impedance is lower than the long ground wires in star grounding. If you run signals with significant current close to the chassis, you can reduce the area of the loops, and help avoid hum pickup and signal radiation.

Speaker outputs don't usually get grounded, and you're OK in the normal case where a floating guitar is hooked up with a cord, but if you receive earth ground from other equipment at the input, star grounding starts to gain appeal, as all the sins of the world appear as signals across the chassis, and how do you shield the radiation from the admittedly low voltage AC offsets in the chassis when the chassis can't be used because it's the source? Same thing goes for RF coming in from outside. Of course, guitar cord coax is a silly way to hook up a guitar. The guys who figured out how to hook up a microphone are all dead, and the guitar world still hasn't learned.

For me, it's as simple as seeing the loops, minimizing or shielding them, and reducing their effect. You can short a big loop and make two smaller ones with a cap, at least for power, or mess with inductance, but you have to understand the loop and put the cap or inductor in the right place.

So, I think you'll agree, the noise issue isn't really chassis grounding vs. star grounding. It's loops (and grid stop shot noise). That's why I stressed cap placement and loop management in a discussion about star- versus chassis- grounding. It's very easy to screw up either option.

I think I know a bit about the science, but I still find myself making hard choices in signal integrity. My religion works, but I've met people with other religions, and they can get good results too. You just need to be faithful yet open-minded.