Idealized grounding arrangement per Merlin!

[NorCalTuna]

I'm reworking a build to closely follow Merlin's excellent scheme which has been floating around here for a bit- thanks Merlin, book order shortly following!

My question pertains to his initial connection to the "reservoir" cap. I'm pretty sure he is presuming a bridge rectifier, sans center tap. The transformer I'm using is a fender-style that has one. If I'm correct, the center tap forms that lower leg and must, to work with this scheme, be connected at the base (-) of the res cap forming the reservoir gnd point.

Am I on the right track? I'm using a bus wire from the input Jack, Where the bond is made, back to a main spur and the res ground, per Merlin, and a mains earth right off the plug, unconnected to anything else, also per big M.


Thanks forum!

[R.G.]

My question pertains to his initial connection to the "reservoir" cap. I'm pretty sure he is presuming a bridge rectifier, sans center tap. The transformer I'm using is a fender-style that has one. If I'm correct, the center tap forms that lower leg and must, to work with this scheme, be connected at the base (-) of the res cap forming the reservoir gnd point.

I haven't read Merlin's book, but the result is the same for every rectification scheme. No matter how the rectified current pulses get into the first filter cap, you want the rectifier outputs going to the first filter cap terminals directly. You do NOT want to connect them to other "ground" points along the way. This will cause hum/buzz because of the high currents causing shifts in the voltage on the signal grounds. If you have a CT power transformer, the two cathodes of the diodes go to the + side of the cap and the CT, which carries both of the diode return current pulses, goes directly to the - side. With a FWR, the '+' side of the bridge goes to the '+' side of the cap, the '-' side of the bridge goes to the '-' side of the cap, and the '-' side of the cap is treated the same in both cases: a wire leads from there to the clean ground point.

Am I on the right track? I'm using a bus wire from the input Jack, Where the bond is made, back to a main spur and the res ground, per Merlin, and a mains earth right off the plug, unconnected to anything else, also per big M.

Again, haven't read the book, but grounding the chassis to the input jack and isolating it from all other "grounds" except the safety ground is a good thing to do. The mains safety ground can tie to any place on the chassis, but for safety sake, it must be tied to a place used only for that - no you can't stick the ring terminal onto the power transformer mounting bolts - and the paint must be sanded away. A toothed washer goes between the chassis and ring terminal, another over the ring terminal and under the nut. The wire must be long enough that it's the last thing to go if the AC mains entry or cord is pulled out hard enough to break the wires.

This is all right out of the playbook for good grounding practice.

[NorCalTuna]

wow, thanks RG for the speedy and clear response! I had a look again at some fundamentals of recs and feel much more up to speed.

I also like the tip about the long enough gnd, it brings some some funny (and not so funny) images of people tripping over a cord or not unplugging the amp to move it. I use an IEC plug so it hopefully just pulls out- and because it allows longer and shorter cords etc... safety first for maximum rock! : ^)

thanks again!

[R.G.]

wow, thanks RG for the speedy and clear response! I had a look again at some fundamentals of recs and feel much more up to speed.

Glad I could help.

I also like the tip about the long enough gnd, it brings some some funny (and not so funny) images of people tripping over a cord or not unplugging the amp to move it. I use an IEC plug so it hopefully just pulls out- and because it allows longer and shorter cords etc... safety first for maximum rock! : ^)

The IEC connector is good. The ground pin is deliberately longer for just that reason.

[Wilder Amplification]

One HUGE benefit to star grounding and having only one tie point from the circuit ground to the chassis is that it makes it possible to install a ground lift switch. Not sure if you've ever played any clubs, but lots of them don't have the best wiring schemes around, and some are even fed from two separate mains panels. If you end up plugging some stuff into one outlet while plugging other stuff in your rig into another outlet and one of the outlets happens to be on a separate mains panel, you've got ground loop city.

Some guys do the wrong thing and either use those "3 prong to 2 prong" adapters to lift the mains ground and some even go so far as to cut the 3rd prong off the cord! So there you no longer have a mains safety ground.

But with a single tie point from circuit ground to the chassis, you can install a switch to lift that tie point off the chassis in the event of an external ground loop. This removes the ground loop while leaving your mains safety ground intact.

[soundmasterg]

Again, haven't read the book, but grounding the chassis to the input jack and isolating it from all other "grounds" except the safety ground is a good thing to do. The mains safety ground can tie to any place on the chassis, but for safety sake, it must be tied to a place used only for that - no you can't stick the ring terminal onto the power transformer mounting bolts - and the paint must be sanded away. A toothed washer goes between the chassis and ring terminal, another over the ring terminal and under the nut. The wire must be long enough that it's the last thing to go if the AC mains entry or cord is pulled out hard enough to break the wires.

This is all right out of the playbook for good grounding practice.

Here's a link to Merlin's grounding article if you have the time to read it R.G.

The Valve Wizard

Its pretty good, and is written in a way that is easy to read, which is always good!

Greg

[NorCalTuna]

One HUGE benefit to star grounding and having only one tie point from the circuit ground to the chassis is that it makes it possible to install a ground lift switch. Not sure if you've ever played any clubs, but lots of them don't have the best wiring schemes around, and some are even fed from two separate mains panels. If you end up plugging some stuff into one outlet while plugging other stuff in your rig into another outlet and one of the outlets happens to be on a separate mains panel, you've got ground loop city.

Some guys do the wrong thing and either use those "3 prong to 2 prong" adapters to lift the mains ground and some even go so far as to cut the 3rd prong off the cord! So there you no longer have a mains safety ground.

But with a single tie point from circuit ground to the chassis, you can install a switch to lift that tie point off the chassis in the event of an external ground loop. This removes the ground loop while leaving your mains safety ground intact.

I do play some scuzzy clubs. In fact, the one this amp is destined to be played in one of the scuzzy bars you see in the background of "American Graffiti", the beloved Gale's Central Club in Petaluma, Ca. We set up there one time and a loose outlet breakered-off half the bar off when I plugged the PA in and the socket grounded against the box. Yay I thought the "patrons" were going to drag me into the street and....

Am I correct in thinking you're saying that you can lift what merlin calls the "Ground To Chassis Connection" with a switch in event of loop? Do you couple it to earth through a cap at that point with a on-on switch, or just let it float? Is the amp less safe functioning like this?

Further, if the entire return path from input to reservoir is on the bus, and what i'm saying is correct and it will function floating, why bond it to the chassis anyway? aren't there amps out there that the circuit floats entirely in the box and isn't ever tied to the chassis? Am I totally missing something?

Thanks!

[R.G.]

One HUGE benefit to star grounding and having only one tie point from the circuit ground to the chassis is that it makes it possible to install a ground lift switch. Not sure if you've ever played any clubs, but lots of them don't have the best wiring schemes around, and some are even fed from two separate mains panels. If you end up plugging some stuff into one outlet while plugging other stuff in your rig into another outlet and one of the outlets happens to be on a separate mains panel, you've got ground loop city.

Not to mention sudden-death city if you get it wrong.

Some guys do the wrong thing and either use those "3 prong to 2 prong" adapters to lift the mains ground and some even go so far as to cut the 3rd prong off the cord! So there you no longer have a mains safety ground.

If those guys just got more play time, Darwin would start cleaning up the gene pool.

But with a single tie point from circuit ground to the chassis, you can install a switch to lift that tie point off the chassis in the event of an external ground loop. This removes the ground loop while leaving your mains safety ground intact.

There's a problem with that, too.

If your signal ground is not tied to your safety ground, then you can indeed get away from hum from inside your amp. The problem is that you can now die from touching your own amp. Lift the signal ground to chassis ground connection inside your amp and you're free - free of the safety ground. Touch a mike whose ground in another piece of equipment is also lifted, but which has an AC leak from power to signal ground inside, and then touch your amp's front panel, and now you're the path from arm to arm across your chest for the AC leakage. In the EE biz we call this a term from another industry. It's "sudden death overtime".

If you're going to play lift-the-ground, you're vulnerable to not only your amp but all the other stuff you may touch. There is a dirty trick that probably will not pass formal safety testing but may help. Use a low resistance, perhaps 10 ohms, paralleled by a 0.01uF, 1kV ceramic cap, paralleling a big, burly diode bridge. The AC terminals of the bridge are tied to the ends of the resistor, and this whole mess goes between safety ground and signal ground. Your ground lift switch goes across this.

The cap keeps out RF buzz. The diodes prevent low level hums, but will conduct heavily (use a 25A bridge here!) and blow fuses if it comes to that, but will keep the voltage difference between chassis and safety ground down to +/-1.4V for bigger, solid faults.

There's a better answer. Isolation transformers. Your amp probably doesn't pull more than, say, 500W from the wall. A 500W isolation transformer can be bought new, retail for under $150, and with some looking can be had as cheaply as $50 by haunting surplus places, craigslist and ebay. This thing safely breaks your amp from the AC line entirely . Sure, they're heavy and expensive. So are coffins.

A quick look at ebay shows a 2KVA (yep, two kilowatt) isolator that's agency approved and can operate from either 120 or 230 inputs for $160. It's heavy, but it's more safety even in dodgy bars. You can run a lot of stage gear from one of these.

If you're seriously in the game of playing music, you can get regulating isolators. These things not only isolate you from the AC line, they regulate the AC voltage and filter out line noise. Again, expensive (couple hundred, surplus) but an incredibly good tool for substandard bars and playing outdoors from generators.

[mooreamps]

I don't know why he calls it a smoothing cap ; when it's actual function is "de-coupling". Further, it looks like is all word for word from Kevin's book TUT1. Since Kevin's book come first, no surprise here, it looks like just another re-print.


-g

[jmaf]

RG, I don't think an amplifier chassis earth will protect you from a series potential, as in the microphone example, unless the microphone is 0 ohm to your guitar amp chassis, but then it's not an issue with the amp. In fact, in that particular example, the low impedance path to earth, through the amplifier, would increase the danger if the microphone was charged.

And there lies the problem....is every piece of equipment grounded to the same potential(0 VAC earth)? In 99% of the small gig stages I've seen the answer is no.

[jmaf]

Here's a link to Merlin's grounding article if you have the time to read it R.G.

The Valve Wizard

Its pretty good, and is written in a way that is easy to read, which is always good!

Greg

From the article:

Most guitar amps are built in a metal chassis. Even if it is enclosed in a wooden box, it is usually possible for the user to touch the metal somehow...

Yes "somehow"...like...playing the guitar...

[R.G.]

RG, I don't think an amplifier chassis earth will protect you from a series potential, as in the microphone example,

You're right - it won't. I may not have been clear.

unless the microphone is 0 ohm to your guitar amp chassis, but then it's not an issue with the amp. In fact, in that particular example, the low impedance path to earth, through the amplifier, would increase the danger if the microphone was charged.

Yes. If you ground your amp chassis and also hard connect the signal ground to the safety ground, you are now putting yourself at risk if you hold your guitar and grab a microphone which is connected to an ungrounded PA amp chassis.

The whole point of safety grounds is that ALL of the equipment is grounded for safety purposes to the AC power line safety ground. That makes the microphone nominally "0 ohm" to the guitar amp chassis, and to your guitar, although we all know that there is some resistance. The resistance is specified to be and tested to be quite small and capable of supporting IIRC 25A of current without opening.

In a setup like that, the house wiring is supposed to take care of holding the "ground" wires at the same potential. If one of a number of amps is ungrounded to make it quieter, then it becomes hazardous because its chassis is not not connected to safety ground, and an internal short of the AC power to chassis does not blow the fuse. Instead, it waves the chassis "ground" around at line voltage, and is an electrocution trap.

Necessarily, ungrounding a piece of equipment to make that one piece quieter is penny wise and pound foolish. In some ways it could be interpreted a selfish - the user of the ungrounded equipment is putting others at risk. It sets a trap for others.

And there lies the problem....is every piece of equipment grounded to the same potential(0 VAC earth)? In 99% of the small gig stages I've seen the answer is no. If they are all grounded to the AC safety line - and critically, the house wiring is also set up correctly, then there can be no dangerous levels on any of the grounds.

However, if the house grounding lets a few volts of difference sneak into the ground wires, then the equipment is still safe, but will hum because the various pieces of equipment all are connected to different "grounds" and they have no common mode rejection. The amps think that a waggling ground is the same as signal.

Pro audio had the answer a long time ago - use differential signal wires. Imagine how much less a problem this would be if we used one whole 12AX7 as a differential amplifier at the input to an amp, and not just 1/2. The differential pair can now reject many db of common mode signal. A lone triode has zero rejection of ground noise.

But back at safety. All grounds solidly connected to AC power safety ground is safe. But bad electrical wiring may make it hum. One way an amp can cause hum is to leak AC so badly that it makes enough ground current to not quite blow its fuse. The ground line then is carrying the current-induced voltage and hums when you connect it to something else with a non-current carrying ground wire.

So we cut the connection to AC safety ground on the leaky box. Now it does not hum due to ground currents. But its chassis is live to the AC line with enough current to wave its signal ground around by many volts. Grab your (grounded) guitar and this ungrounded, but quiet mike, and you get a shock. How much shock depends on the details.

If you had "lifted" the ground with resistor/cap/diodes, then the signal ground may be quiet, but it can't get more than 1.4V away from the safety ground, and the current is enough to blow the line fuse if a harder short develops. It (probably!) is low enough voltage to hum less, and probably can't support enough current to kill you.

This is still not good, and as I said, probably won't pass safety testing, but may solve the hum problem without causing electrocutions.

How'd I do that time? Was that clearer?

[jmaf]

You're right - it won't. I may not have been clear.

Yes. If you ground your amp chassis and also hard connect the signal ground to the safety ground, you are now putting yourself at risk if you hold your guitar and grab a microphone which is connected to an ungrounded PA amp chassis.

The whole point of safety grounds is that ALL of the equipment is grounded for safety purposes to the AC power line safety ground. That makes the microphone nominally "0 ohm" to the guitar amp chassis, and to your guitar, although we all know that there is some resistance. The resistance is specified to be and tested to be quite small and capable of supporting IIRC 25A of current without opening.

In a setup like that, the house wiring is supposed to take care of holding the "ground" wires at the same potential. If one of a number of amps is ungrounded to make it quieter, then it becomes hazardous because its chassis is not not connected to safety ground, and an internal short of the AC power to chassis does not blow the fuse. Instead, it waves the chassis "ground" around at line voltage, and is an electrocution trap.

Necessarily, ungrounding a piece of equipment to make that one piece quieter is penny wise and pound foolish. In some ways it could be interpreted a selfish - the user of the ungrounded equipment is putting others at risk. It sets a trap for others.

And there lies the problem....is every piece of equipment grounded to the same potential(0 VAC earth)? In 99% of the small gig stages I've seen the answer is no. If they are all grounded to the AC safety line - and critically, the house wiring is also set up correctly, then there can be no dangerous levels on any of the grounds.

However, if the house grounding lets a few volts of difference sneak into the ground wires, then the equipment is still safe, but will hum because the various pieces of equipment all are connected to different "grounds" and they have no common mode rejection. The amps think that a waggling ground is the same as signal.

Pro audio had the answer a long time ago - use differential signal wires. Imagine how much less a problem this would be if we used one whole 12AX7 as a differential amplifier at the input to an amp, and not just 1/2. The differential pair can now reject many db of common mode signal. A lone triode has zero rejection of ground noise.

But back at safety. All grounds solidly connected to AC power safety ground is safe. But bad electrical wiring may make it hum. One way an amp can cause hum is to leak AC so badly that it makes enough ground current to not quite blow its fuse. The ground line then is carrying the current-induced voltage and hums when you connect it to something else with a non-current carrying ground wire.

So we cut the connection to AC safety ground on the leaky box. Now it does not hum due to ground currents. But its chassis is live to the AC line with enough current to wave its signal ground around by many volts. Grab your (grounded) guitar and this ungrounded, but quiet mike, and you get a shock. How much shock depends on the details.

If you had "lifted" the ground with resistor/cap/diodes, then the signal ground may be quiet, but it can't get more than 1.4V away from the safety ground, and the current is enough to blow the line fuse if a harder short develops. It (probably!) is low enough voltage to hum less, and probably can't support enough current to kill you.

This is still not good, and as I said, probably won't pass safety testing, but may solve the hum problem without causing electrocutions.

How'd I do that time? Was that clearer?

As always with your ellaborate replies RG, clear as can be. Thanks.

[Merlinb]

My question pertains to his initial connection to the "reservoir" cap. If I'm correct, the center tap forms that lower leg and must, to work with this scheme, be connected at the base (-) of the res cap forming the reservoir gnd point.

Breaking it into four simple (and more or less universal) steps;

You should consider the transformer, rectifier and reservoir cap to be a single circuit block, using short thick connections. (Whatever the type of rectifier, it doesn't matter). You don't even need to consider it as part of the 'ground scheme', it's just a small, independent piece of circuitry. This is nothing special to tube amps; it works this way for all equipment using a linear power supply.

Then you have your actual 'ground scheme', which incorporates all the audio amplifying stages and their associated smoothing/decoupling caps.

The ground scheme then gets wired up to the reservoir cap/rectifier block (which you built first, right ).

Then the audio input is connected to the chassis.



(Gary, 'decoupling cap' is a generic term, 'smoothing cap' is a type of decoupling cap used in power supplies. I have not yet read TUT1)

[JHow]

... 'decoupling cap' is a generic term, 'smoothing cap' is a type of decoupling cap used in power supplies...

And I was wondering if "smoothing" was a more British way of speaking, or is it accepted engineer-speak? I am aware of all those boots, bonnets, tappets, spanners (and even valves!) that don't mean what I thought they meant.