Audio system grounding

[wayne325]

See attached:

Discussion ????

[Speedskater]

As I just wrote:

When in doubt, believe:

Bill Whitlock of Jensen Transformer Co.
Jim Brown of Audio Systems Group
& the technical papers at Rane Corp.

any other way is wrong.

Pin #1 of an XLR Connector is not signal ground!

[wayne325]

I did what a Rane corp paper says as a starting point, thinking
they knew of what they wrote. THe paper in question is
called "Sound System Interconnection" and it has a label
RaneNote 110, copyright 1985 and 1995. Figure 1a in that
paper is called "The right way to do it".

When I connected my system in that configuration, the 60 Hz
noise was 20 to 40 dB louder than the signal. That is, the
system was not usable.

The configuration in question would be identical to my paper
except the two XLR pin 1's would be connected to the grounding
post rather than the signal ground.

The problem with that configuration is obvious if you draw it...
It creates ground loops potentially several square meters in
area. Currents travelling through those loops generate voltages
in the loop.

When I finally got to the end of my rope doing things people
like Rane were telling me to do and having a system that
generated 60 Hz noise signals really well (we're talking enough
noise at the preamp input to blow out my speakers if I turned
up the volume to maybe half of full volume), I stopped and looked
at pictures I took of my Pearl before I took it apart and replaced
it with my "Xono Clono". When I connected the Xono the same
way I did my Pearl - lo and behold, all the problems went away.
The system is dead silent now. The last thing I did was move
the XLR pin 1's from chassis ground to signal ground.

I'm not going to say there are no other ways to connect the
grounds. But I did try quite a few, and one of the worst was
to connect the various chassis together using pin 1.

W.

[luvdunhill]

I guess I'm confused as well. I have pin 1 on both output XLRs connected to the ground plane of my Ono. I have the barrel of the input RCA connected to amp ground as well. The turntable ground as well. The umbilical connects the ground of the power supply to the ground of the amplifier section. All jacks are isolated from the chassis, only chassis connection is the earth connection of the IEC power. I have absolutely no hum whatsoever.

Next time I'm up your way Wayne (or vise-versa), we should compare Onos! I live in Austin

[wayne325]

Hi Dunhill,

Yes absolutely we should get together some time and compare
notes.

My amp isn't finished yet. It's a bit sick in the MC section still,
I think I've killed the cascode BJTs, and I think I also killed one
of the ZTX450s when I accidentally shorted V+ to sig ground
when I was probing around. I hate it when that happens.
Dang alligator clips. I also have to solder in new bias resistors
to get the main amp section current sources corrected.
So I'm waiting for parts.

From your description about grounding, I can't figure out
exactly what you mean - when you say "ground" - what is
"signal ground" and what is "chassis ground" ? Maybe a picture
will help.

I'm very willing for another scheme where the XLR1 gets
connected to the chassis produces no hum - because that will
mean no RF noise being injected into the amps. I just haven't
found it yet. It's hard to find time for all this also, and reading
all the papers, etc.

[Audio_se]

Hi

I belive the Rane papers saying Pin 1 should be connected to chassis is correct.
A ground loop is supposed go in the chassis where it won't do any harm for the electronics inside.

What more parts do you have in your system?
Which of them have a power ground?
Are all your equipment connected to the same wall outlet?

A gound loop must be a closed loop in order to exist. The attached picture don't show the complete system.

You could try to run one device at the time without power ground (be careful) and figure out where the loop is.


/J

[KSTR]

Well, I'd say if Wayne fixed the problem with the way he connected all his stuff, then it is the correct way (perhaps not the only one, though) for him -- whatever works is fine.

Still a big question is: what is the root cause? I can't imagine "ground loop currents" to be the reason IF the gear has no internal reference to pwr gnd and does the "subtraction thing" right. A ground loop could have only two causes:
- loop area, with an AC magnetic field going through that loop area, inducing AC current
- voltage drops/differences between different mains outlets because somewhere current is induced into the safety gnd conductor (a problem with some mains filters, ie).

Of course I don't know the details of your components but I still speculate there is a GND reference problem somewhere INSIDE a device...

- Klaus

[spzzzzkt]

It looks like you are treating the setup as two separate devices. I'd decouple the chassis ground in the preamp with a thermistor and reference all gnds to the chassis ground in the psu. I'd also make the two gnd points in the chassis a star gnd. It looks like tying the xlrs to chassis would create big gnd loops as configured now.

[Speedskater]

Sorry about my abrupt reply yesterday.
My thoughts are:
a) All balanced interconnects should have the shield connected at both ends.
a1) Most pros allow the cable shells to float.
b) Chassis mount XLR connectors must have pin 1 connected to the metal chassis. Not to audio ground.
b1) In a balanced circuit, the shield is not part of the audio signal.
c) All RCA un-balanced connectors should be connected to the audio ground.

Not sure but I think that the "Turntable Chassis ground wire" should go to metal chassis ground, and the "Drain wire of phono cable" should go to audio ground.

[wayne325]

Originally posted by Speedskater
Sorry about my abrupt reply yesterday.
My thoughts are:
a) All balanced interconnects should have the shield connected at both ends.
a1) Most pros allow the cable shells to float.
b) Chassis mount XLR connectors must have pin 1 connected to the metal chassis. Not to audio ground.
b1) In a balanced circuit, the shield is not part of the audio signal.
c) All RCA un-balanced connectors should be connected to the audio ground.

Not sure but I think that the "Turntable Chassis ground wire" should go to metal chassis ground, and the "Drain wire of phono cable" should go to audio ground.

Skater,

No offense taken.... I can't look at this now as I am too busy at
work until next weekend. I'm hopeful that my preamp has a
chassis ground problem and that is interacting with the phono
pre output. But I won't figure that out for sure until I can get
in there and see how it's connected, draw some pictures,
and see if I can correlate my observations last weekend with
what all the connections are, as alluded to by Audio_se.
(That is, there's more to it than the phono pre - it also connects
somewhere else) I agree with what you write - my problem is
my best attempt to do that resulted in a really bad situation.
If I may be so bold I'd want to connect the phono cable drain
wire also to the chassis so that all the RF it contains is sent
to the chassis rather than the electronics inside.

While all this is going on, I'm trying to read the Rane papers
again, and Ott's book to educate myself about what exactly
to do. It's easy to say "connect XLR1 to chassis ground" but
there are lots of other connections in a 2-chassis amp - plus
the amp it connects to.

Thanks for all the input gents.

W.