Ground/earth solution for class 2 amplifier with some voltage on the chassis

[IanForumUser]

Hi folks, I think I understand the issue at hand here, but not having any formal training in electronics I thought it best to run my problem and proposed solution via the ASR crowd!

Problem: I have a Denon PMA-2020ae. Great amplifier, and it's in perfect working order. However, there is a very low level of voltage detectable on the chassis (I haven't put a meter on it, but you can feel it on the casework during operation as a very mild static-like field). I understand this is common in class 2 electrical devices, especially those with large transformers (the Denon has two big EI types), and what I'm feeling is just some harmless capacitive coupling. Assuming that's correct...

Proposed solution: The amp has a signal ground lug for the unused phono input. If I run an earth wire from the signal ground to the earth pin of an otherwise empty UK mains plug, that elimates the "problem". I should add that the amp isn't connected to any class 1 devices, just other class 2 devices.

Question: Is the proposed solution safe? Safe for me and safe for the amp? I can't think why it wouldn't be given the "problem" isn't a safety concern to begin with (based on what I've read, including some bits on ASR). But I am very aware that I don't know what I don't know!

 

[Old_School_Brad]

Class II devices, as you've noticed, commonly produce a slight electrical sensation when touching the chassis. HThis occurs at a very high impedance, meaning the current is extremely small and not hazardous.

The GND terminal on the back is intended for signal grounding to help reduce noise from other sources. It is not a safety ground.

 

[IanForumUser]

Thanks Brad. So it's accurate to say I've not made anything safer by connecting the signal ground to mains inlet/safety ground? But it does eliminate the electrical sensation on the casework. Does it in anyway introduce safety concerns? I can certainly live with the "problem" if so!

 

[DVDdoug]

I'd be "slightly concerned". I don't see any harm in leaving the ground but if-possible I'd use an existing screw that's solidly-connected to the cabinet-chassis.

...When I was in high school my mother said, "I'm feeling a tingle from the waffle iron." I didn't feel anything and I thought maybe she was imagining it. But I knew a little about electricity and I was wearing sneakers. So I reached-over and briefly touched the water faucet. BLAM! Huge jolt through my chest! And I said, "That was stupid!" and I suggested she get rid of the thing.

 

[gwing]

As I understand it a double insulated class 2 device does not *require* an earth but there is nothing stopping you adding one for some degree of extra safety.

Actually, if you have all class 2 devices in your system it seems to me to be a good idea to have at least one class 1 device or if not to earth the signal ground on *one* of your class 2 devices so as to reduce noise *but* I don't claim any qualified or comprehensive knowledge here.

 

[Old_School_Brad]

A Class II device should not be connected to ground or protective earth because it's built to keep any faults inside the device from causing dangerous voltages on exposed parts. This is done through double or reinforced insulation, which means grounding isn’t necessary. Connecting the device to ground or PE wouldn't help, as the devices design doesn’t allow enough current to flow to trigger a safety breaker. This would give a false sense of security and could be against safety rules and thus non-compliant.

 

[AnalogSteph]

Actually, if you have all class 2 devices in your system it seems to me to be a good idea to have at least one class 1 device or if not to earth the signal ground on *one* of your class 2 devices so as to reduce noise *but* I don't claim any qualified or comprehensive knowledge here.

I concur.

If the entire setup is left floating, you will have the mains leakage currents of the entire bunch to contend with. These things add up - a few nF here and there, and suddenly you get a decent tingle / hum on open inputs. The amp may not even be the main culprit.

So the recommended number of earth/ground connections is exactly one per setup. In the olden days, that role might have fallen to the external FM aerial, for example.

Since the leakage current is still minuscule in the grand scheme of things (equivalent to several tens of kOhms to mains at the very least), you don't need massive wire gauge or anything. This is not protective earthing and doesn't have to be.

 

[IanForumUser]

A Class II device should not be connected to ground or protective earth because it's built to keep any faults inside the device from causing dangerous voltages on exposed parts. This is done through double or reinforced insulation, which means grounding isn’t necessary. Connecting the device to ground or PE wouldn't help, as the devices design doesn’t allow enough current to flow to trigger a safety breaker. This would give a false sense of security and could be against safety rules and thus non-compliant.

That sounds pretty clear to me. I don't want to make anything less safe in pursuit of eliminating benign capacitive coupling.

 

[IanForumUser]

I concur.

If the entire setup is left floating, you will have the mains leakage currents of the entire bunch to contend with. These things add up - a few nF here and there, and suddenly you get a decent tingle / hum on open inputs. The amp may not even be the main culprit.

So the recommended number of earth/ground connections is exactly one per setup.

Since the leakage current is still minuscule in the grand scheme of things (equivalent to several tens of kOhms to mains at the very least), you don't need massive wire gauge or anything.

Thanks for this, which does make sense to me. But then so does Brad's "operate as designed" approach!

 

[Old_School_Brad]

That sounds pretty clear to me. I don't want to make anything less safe in pursuit of eliminating benign capacitive coupling.

The point is, it wouldn’t make the system any safer or more dangerous.

 

[Speedskater]

a Class II double insulated device that has a total plastic enclosure, should not be modified to ground its internal metal chassis.
But if the amp has an exposed metal chassis, that means that just the power supply is double insulated.
That phono ground lug can be connected to the chassis of another audio component that has a metal chassis and a 3 wire AC cord.

 

[IanForumUser]

a Class II double insulated device that has a total plastic enclosure, should not be modified to ground its internal metal chassis.
But if the amp has an exposed metal chassis, that means that just the power supply is double insulated.
That phono ground lug can be connected to the chassis of another audio component that has a metal chassis and a 3 wire AC cord.

But not direct to the outlet? Only via a class 1 devices?

 

[AnalogSteph]

A Class II device should not be connected to ground or protective earth because it's built to keep any faults inside the device from causing dangerous voltages on exposed parts.

The former does not follow from the latter, IMHO.

A Class II device does not need explicit grounding to be safe, but there is nothing keeping you from adding it. I have been advocating fitting vintage Yamaha A-x60 series integrateds with 3-conductor mains cords with the PE used for nothing else but draining the mains filter leakage currents. Chassis even remains separated from PE via a Y class capacitor in this case so not even blatant outlet miswiring would make things unsafe. The procedure is a bit of a chore with US/CAN models but quite trivial with others (220/240V, General) as you just need to solder the PE to a wire bridge and that's it.

But not direct to the outlet? Only via a class 1 devices?

As you may be suspecting already, it doesn't really make a difference.

If you want to get really fancy, you could include something like a 4.7-10 nF Y class capacitor paralleled with a 1 Megohm resistor in the connection, then things would remain safe even if your ground connection were suddenly connected to live. It probably wouldn't be necessary though, as outlets tend to not spontaneously miswire themselves.

 

[solderdude]

Proposed solution: The amp has a signal ground lug for the unused phono input. If I run an earth wire from the signal ground to the earth pin of an otherwise empty UK mains plug, that elimates the "problem". I should add that the amp isn't connected to any class 1 devices, just other class 2 devices.

Yep, it will route the leakage currents to ground and should remove the tingling feeling.
That ground-lug is always the chassis and is intended to ground the tonearm and chassis of a turntable.

The point is when you have multiple class-II devices, especially those with those lightweight wall warts or bricks, then the total leakage currents could really add up.
Grounding is not really needed nor essential but it will help in this case to get rid of the tingling.
As a test you can rub your finger on metal parts of the chassis and if you feel it 'bouncing' there is too much leakage.
Also what could be interesting is to disconnect all connected devices and do the rub test. Then connect 1 device, rub, connect the second one, rub.. and see if one of the devices is the main culprit.

Question: Is the proposed solution safe? Safe for me and safe for the amp? I can't think why it wouldn't be given the "problem" isn't a safety concern to begin with (based on what I've read, including some bits on ASR). But I am very aware that I don't know what I don't know!

Yep, it is safe to do so.
When you get a hum you could use the trick from the post above (AnalogSteph) with the 1M and Y cap in parallel.

 

[IanForumUser]

Yep, it will route the leakage currents to ground and should remove the tingling feeling.
That ground-lug is always the chassis and is intended to ground the tonearm and chassis of a turntable.

The point is when you have multiple class-II devices, especially those with those lightweight wall warts or bricks, then the total leakage currents could really add up.
Grounding is not really needed nor essential but it will help in this case to get rid of the tingling.
As a test you can rub your finger on metal parts of the chassis and if you feel it 'bouncing' there is too much leakage.
Also what could be interesting is to disconnect all connected devices and do the rub test. Then connect 1 device, rub, connect the second one, rub.. and see if one of the devices is the main culprit.


Yep, it is safe to do so.
When you get a hum you could use the trick from the post above (AnalogSteph) with the 1M and Y cap in parallel.

Super interesting, thanks! The "tingle" / "bouncing" / "static-like" sensation is present with nothing but the power cord connected. It's not made noticeably worse with other devices (a Denon CDP and a Wiim Ultra) connected; and neither of those devices exhibit it themselves. I won't be doing any mods to the amplifier, so I think safety grounding the signal ground is what I will continue to do, on the basis it has no deleterious effect on performance or safety.

 

[Speedskater]

But not direct to the outlet? Only via a class 1 devices?

Well you could attach (bond) it to the wall outlet, that would be safe.
But it's more effective to attach it to an audio component in the same system.
From an audio point of view having all the chassis's at the same potential is more productive.

 

[sam_adams]

Hi folks, I think I understand the issue at hand here

All your questions will be answered by reading through Bill Whitlock's presentation below:

Bill Whitlock - An Overview of Audio System Grounding and Interfacing - Indy-AES-2012 (PDF).

You can measure the Inter-Chassis Current—capacitance induced circulating current—using the instructions below, also from Bill Whitlock:

From: Bill Whitlock - Hum And Buzz In Unbalanced Interconnect Systems - Jensen AN-004 (PDF).

A Pomona 1330-0 dual banana jack, a 1 KΩ resistor, and a .15 µF film cap will allow you to construct the adapter to test the device for the circulating chassis current level.

 

[Speedskater]

The above measurement assumes a correctly wired AC power circuit. No swapped wires, no bootleg grounds. Those little 3 LED testers won't catch most of those problems.