[Amps] Preventing current flow on PEN/EGC (protective earth conductor/equipment grounding conductor).

Jim Brown jim at audiosystemsgroup.com
Thu Nov 28 23:30:25 EST 2024


On 11/28/2024 3:52 PM, Lukasz wrote:
> Hi Jim,
> 
> It is similar here in the EU as far as I know.
> 
> A typical way devices, are constructed here is that the ground 
> (including the case if made of metal) is not used as a normal current 
> path in the device. One sees current flow on that ground only in a fault 
> scenario.
> 
> This means RCD/GFCI protections can work as when it see any discrepancy 
> in the live/neutral current balance it shuts the circuit down.
> 
> These rules normally mean one always has a return current path within 
> the device that is separate from ground (neutral, DC minus etc), but 
> this is difficult when the amplifier is entirely made of metal and one 
> would prefer to have a current return plane (for example the floor if 
> the case) rather than route separate connections for DC minus and RF 
> ground.

There's an important principle that I learned many years ago at an IEEE 
EMC conference, with a corollary that I learned from the late Henry Ott, 
an expert on EMC who I got to know a bit.

 From that conference I was reminded that current follows the path of 
the lowest IMPEDANCE; above about 1 kHz, that path is a transmission 
line; with a connection to the mains, the transmission line is formed by 
the two current-carrying conductors. The reason that the protective 
earth is required to be run with them is so that the fuse or breaker 
will act more quickly. With modern PC boards that place traces above a 
another continuous layer, the trace forms a transmission line. The 
chassis would function in this manner only if the conductors for the 
current the chassis is returning were placed a few mm above it.

I'm long time Vice-Chair of the EMC Working group of  he Audio 
Engineering Society. One of the issues we got into in writing EMC 
standards was to view the chassis as a "shielding enclosure," that 
shielding only works if it is continuous, and that any conductor that 
passes through it will conduct RF and and out. That's why feed-through 
bypass capacitors were used in the old days. And in the world of pro 
audio, EMC means rejecting VHF and UHF transmitters, whether cell phone 
or high power broadcast.

The primary cause of RFI to/from equipment is a cable shield or bonding 
conductor that passes through an opening in the enclosure, wanders 
around inside either on discrete wires or a PC board before it finally 
hits the shielding enclosure. That's "The Pin One Problem" that you may 
have heard about, where Pin One comes from the fact that Pin One of pro 
mic connectors carries the cable shield.
> 
> This is not how OLD devices were made. I have a couple of old lathes and 
> mills that do use the steel frame as neutral/ground(one connector called 
> PEN) . This is how, it seems, all tube amps are built I saw pictures so 
> far.
> 
> So my question is, if one wants to build a tube amp in a modern way. 
> Properly grounded, is building a case within a case and electrically 
> isolating the two the only way? 

For many years, good engineering practice for low voltage power supplies 
has been to float the circuit board from the chassis, AND to float DC+ 
and DC-, and to provide an easy means for bonding DC- to the chassis if 
desired. Unless you need to do something different with your HV PSU, 
this seems like a good approach.

What does "grounding" mean to you? What do you expect it to accomplish?

If you mean a connection to earth, it is of value ONLY as part of 
lightning protection, and for lightning protection, extensive bonding is 
also required.

Do you mean what we call "bonding," which is connecting all earth 
connections and protective earths together?

> Or has anyone managed to build a tube 
> amplifier in a single grounded steel case. That works normally within 
> the usual electrical systems (with gfci/rcd protected circuits). If so, 
> please describe how are all the current return paths done.

Here in NA, our "house wiring" is 120-0-120, where 0 is the neutral, and 
big power supplies run between the two sides of 120V. Neutral is bonded 
to protective earth at the point where power enters our home. Here, 
GFCIs are required only in locations like kitchens and bathrooms, and 
would not be required on circuits feeding a ham shack, bedroom, or other 
living space. When RCDs were introduced here as a requirement, ARRL's 
EMC people went through a long period working with manufacturers to 
solve their susceptibility to our transmitters.

Since I don't build amps, this is as far as I can go.:)

73, Jim K9YC

> 
> 73,
> Luk, SP4IT
> 
> On Wed, 27 Nov 2024, 23:05 Jim Brown, <jim at audiosystemsgroup.com 
> <mailto:jim at audiosystemsgroup.com>> wrote:
> 
>     Hi Lukasz,
> 
>     In North America, the protective earth MUST be carried in the same
>     cable
>     with the current-carrying conductors, and MUST  be bonded to the
>     chassis
>     at the point where that cable enters. I'd be surprised if it's
>     different
>     in EU -- it must be done that way to be protective!
> 
>     There's another requirement in NA -- all grounds (earths) MUST be
>     bonded
>     together. Again, protective against both lightning and electrical
>     shock.
>     So when power goes through ANY transformer that's not part of
>     equipment,
>     protective earth MUST be carried from one side of the transformer to
>     the
>     other.
> 
>     I don't remember about EU, but I do remember that in the UK, no earth
>     connection to the power system is permitted -- that happens only within
>     the power distribution system.
> 
>     When thinking about this and antennas, remember that the earth is a big
>     resistor, and parts of antennas like radial systems are NOT intended to
>     couple to the earth, but rather to shield the EM field from the earth,
>     and to provide a low resistance path for the antenna's return current.
>     The only antennas that benefit from an earth connection are SOME
>     receiving antennas, like Beverages and loops.
> 
>     73, Jim K9YC
> 
>     On 11/26/2024 8:10 AM, Lukasz wrote:
>      > Hi,
>      >
>      > Is there some way to build a tube amp in a metal case, using the
>     case as RF
>      > ground, and not have current flow out of the amp through the
>     grounding
>      > conductor? (other than isolating that ground from the case
>     entirely, which
>      > presents it's own problems - for example capacitive coupling to
>     the case).
>      >
>      > I'm currently experimenting with an amp and I have to power it via an
>      > isolation transformer (the high voltage is 3 phase on its own
>     circuit with
>      > no GFCI temporarily) or my RCD will trigger.
>      >
>      > The high voltage power supply is an old military device that used its
>      > chassis as ultimate DC ground (hv ground connects to it through a
>      > overcurrent protection coil).
>      >
>      > Then I also used the chassis as ground for DC HV, screen and grid
>     supplies.
>      > This I could swap, but it's the RF ground (especially the Pi
>     tank) I'm
>      > concerned about.
>      >
>      > Is essentially isolating amp RF ground from the grounding conductor
>      > (PEN/EGC or whatever it's called where you are) the only way?
>      >
>      > 73,
>      > Łukasz, SP4IT
>      > _______________________________________________
>      > Amps mailing list
>      > Amps at contesting.com <mailto:Amps at contesting.com>
>      > http://lists.contesting.com/mailman/listinfo/amps <http://
>     lists.contesting.com/mailman/listinfo/amps>
> 
> 
>     _______________________________________________
>     Amps mailing list
>     Amps at contesting.com <mailto:Amps at contesting.com>
>     http://lists.contesting.com/mailman/listinfo/amps <http://
>     lists.contesting.com/mailman/listinfo/amps>
> 




More information about the Amps mailing list