Noise with Computers, USB DACs, and Powerline Ethernet Kits

[Paul Birkeland]

There are some vintage components that just use a 10 ohm resistor instead of the antiphase diodes.  I prefer the diodes a bit because the peak current capabilities are better specified.

[cddc]

https://en.wikipedia.org/wiki/Ground_loop_(electricity)

The Ground loop (electricity) article on Wikipedia is terribly written, but I think I roughly got the idea of 'ground loop'.

So the RCA cables that connect Crack and DAC form a ground loop (ground of Crack -> Crack chassis -> RCA grounding wires -> DAC chassis -> ground of DAC -> ground of Crack), and the loop acts as a single-turn secondary winding of a transformer. So through induction it will generate some small noise current from ambient magnetic fields - magnetic fields that come from the main power line (60Hz), router, computer, etc.

So the 'ground breaker' diodes from PB will break the ground loop, preventing it from generating noise current thru induction. Am I correct?

[Paul Birkeland]

It's more about creating a break in the blue line in this image, and mostly breaking the line between the PC and the Crack.  What we don't want is for any nasties riding on the earth connection from the PC to travel through the signal cables, into the Crack, then to earth through the Crack.  A DAC which has its own power supply and does not rely on USB for power will tend to do this on its own.  We used to suggest using a DAC with an optical cable to do this, or SPDIF with a proper input transformer, but the antiphase diodes create an opening in the loop for far less money.

[cddc]

Thanks a lot, PB, for the image.

I think my understanding of 'ground loop' from the terribly written self-conflicting Wikipedia article is exactly the blue line loop in your image.

My computer and DAC both come with SPDIF. I once tried connecting them via an optical cable, but couldn't hear any difference vs the USB output to DAC. 

Then I did some google search, and it seems some people think the SPDIF output from computers is not optimized for HiFi audio, they are designed for data transfer, so the jittering from SPDIF on computers will be very high due to poor clocking. Not too sure about these technical stuff 

[cddc]

It just occurred to me that there is relatively easy solution for dealing with lots of shielding that needs wiring to ground: install your two-diode mod, but isolate the shielding from chassis ground and wire the shield drains directly to the earth ground lug of the iec inlet.

You can do that or move the diodes to between the audio ground and the chassis ground.

That's an even better idea! Thanks PB.

Just intrigued by the talk, but couldn't fully understand it...

I can understand @Deke609's method to wire his Faraday shielding cage to the earth lug of the IEC power socket directly in order to bypass the diode 'ground breaker' and drain the EMI noises captured by the Faraday cage to the ground.

But I couldn't quite understand PB's approach. Where are the audio ground and chassis ground?

I just learned today that the Crack's "ground bus" is grounded thru the 3U/L tab, so should we break the grounding connection there and insert the "ground breaker" there?

[Paul Birkeland]

Sure, you can go ahead and isolate terminal 3 and install the antiphase diodes between terminal 3 and the chassis. 

[cddc]

Cool, thanks a lot for pointing out the "ground breaker" location for Cracks with Faraday cages, PB!

[Geebs]

Hi there - I'm a newbee for comments to the forum.  I built my BHC with Speedball last June; it works well until recently when I notice the following.

I notice/hear/feel a 'static electricity' type shock in the right channel of my new (2 mos) ZMF headphone.  It is winter, dry and lots of static all over my gear.  I rolled a new tube and thought it went away.  Rolled another one and it's still there.  Zach at ZMF has not experienced before.

Has anyone had this happen to them?   Would the fix that Paul started with this thread solve this?

Thanks - and if I'm doing anything wrong in the forum comments (style, quotes, etc.) let me know that as well.

[Paul Birkeland]

Is this noise present with nothing plugged into the RCA jacks on the Crack and the volume pot turned all the way down?

[Geebs]

Thanks Paul - I will try to replicate it under that condition.  I will also try another headphone/cable to see what the result might be.

[Geebs]

So I pulled my last comment - I simply cannot replicate my static shock/energy jolt in the right channel of my headphone.

I tried Paul's suggestion of unplugging the inputs from my Crack with volume at zero; no static shock resulted.

I also tried to replicate with different tubes, on different days (even other days when static was present in the day) and could get no shock through the headphone, even though there would have otherwise been a static electric charge when I touched other parts of my equipment.

Must have been the tubes that was in place on the day.    :-(

[[email protected]]

For the purpose of this mod, you can think of a diode as erecting a barrier that is approx. 700 mV tall: @700 mV and below, no current flows, but everything above 700 mV passes over the barrier. (The "height" of the barrier, called the diode's "forward voltage" -- the minimum voltage required to turn the diode "on" and permit voltage and current to pass, varies a bit by type of diode). This applies to a "forward biased" diode - i.e., one where the anode is more positively charged than the cathode (banded end).  Because the mod uses two "reverse paralleled" diodes, there is a forward-biased diode in each direction: to earth ground and from earth ground -- so it blocks sub-700 mV currents in both directions.

In reality, the "barrier" is sloped and some small amount of current makes it through when the applied voltage is less than 700 mV.  See attached graph, and specifically the region to the right of the X,Y origin.

cheers, Derek

And there it is. An explanation simple enough for me to understand!. thanks PB for the brilliant Mod. and thanks to Derek for lifting the veil!

[sl-15]

Hi, I am currently trying this mod on my Extended Foreplay 3. I am using 1N4004 diodes.
There is a drain wire from the power switch cable connecting the IEC power entry to the first ground lug and then it continues to the ground lug from the power transformer.
When I do the mod I am assuming I need to clip the drain wire between the IEC and the first ground lug but leave a connection between the first ground lug and the transformer ground lug. This is how it is done in the Crack.
I am wondering if it would be better to route the drain wire from the power switch directly to the transformer ground lug and not to the IEC ground? The photo just shows my layout nothing soldered yet and no clipped leads yet. If somebody can confirm that I would appreciative it. Thanks, Best, Stefan

[Paul Birkeland]

That should work the way you have it setup.

[sl-15]

Thanks Paul, this mod works wonders did it on my Crack as well.