Grounding anomaly
Dec 29, 2007 at 8:33 AM Thread Starter Post #1 of 26

FallenAngel

Headphoneus Supremus
Joined
Mar 7, 2006
Posts
9,635
Likes
82
Location
Toronto
Hey guys,

I just "finished" my big amp - CKK-III PreAmp, GainClone amp, AlienDAC and source selector all in a single chassis with the transformers.

After dealing for a long time with noise issues from the AlienDAC - specifically, the speakers make noise when I work on the computer. For example, I open a web browser which takes a quarter of a second and I will hear a little buzzing noise from the speakers.

I have the chassis earth grounded. All of the components (even the AlienDAC) are also grounded to the chassis through their own ground loop breakers, all tied together into a star. The ground loop breakers consists of a 10 Ohm / 5W wirewound resistor in parallel with a 0.1uF film X-type capacitor rated at 250VAC as amb suggested in his site under "Wiring & ground".

Still all this grounding did nothing to remove the noise from the speakers.

Then I did something that worked which I can't explain. I removed the star ground from the chassis and the noise went away! Why is this so? I can definitely say I'm happy with it, but I just don't understand why this happened. Any ideas?

There is still a slight hum from the speakers when nothing is playing, but it is so quiet that I have to have my ear right up against the speaker to hear it. I'm fine with it, but of course will be playing with the grounding more to figure out what I can do to fix this last bit. Probably I'll start connecting/disconnecting components one by one in hopes that it gets resolved.

So, any ideas why disconnecting the grounds from the chassis fixed my AlienDAC noise? Oh, and just to notify, I had this AlienDAC feeding many headphone amps without noticing any noise, it's only with this monstrous all-in-one-box thing that I've noticed any noise.

Thanks
 
Dec 29, 2007 at 9:44 AM Post #4 of 26
Since everything is in a common case, you only need one ground loop breaker. Having multiple ones create small ground loops because you now have multiple ground paths to the case (even though they're through 10 ohm resistors).

As for the low-level hum, how far is the power transformer from all the audio circuitry?
 
Dec 29, 2007 at 10:27 AM Post #5 of 26
Quote:

Originally Posted by amb /img/forum/go_quote.gif
Since everything is in a common case, you only need one ground loop breaker. Having multiple ones create small ground loops because you now have multiple ground paths to the case (even though they're through 10 ohm resistors).

As for the low-level hum, how far is the power transformer from all the audio circuitry?



But is it safe to connect all of the grounds together? That will mean that the ground from the power amps (+/- 31V) will be connected to the ground from the CKK-III (+/- 15V) and the AlienDAC ground. Do excuse if that is an obvious answer.

The transformer is this far
smily_headphones1.gif

 
Dec 29, 2007 at 11:25 AM Post #6 of 26
Quote:

Originally Posted by FallenAngel /img/forum/go_quote.gif
But is it safe to connect all of the grounds together? That will mean that the ground from the power amps (+/- 31V) will be connected to the ground from the CKK-III (+/- 15V) and the AlienDAC ground. Do excuse if that is an obvious answer.


The signal grounds should absolutely be connected together. It should, afterall, serve as the reference for all voltages in the amp.

Quote:

The transformer is this far
smily_headphones1.gif



That's a substantial transformer and will have a relatively strong magnetic field around it. The distance between it and the CK2III board, the Alien DAC board, and the myriad low-signal wires are probably no more than 6 inches or so? You're probably still getting some interference. You might want to experiment with shielding the transformer with something from EMF Safety Superstore- "Electromagnetic Field Detection & Protection". .
 
Dec 29, 2007 at 11:54 AM Post #7 of 26
Quote:

Originally Posted by amb /img/forum/go_quote.gif
The signal grounds should absolutely be connected together. It should, afterall, serve as the reference for all voltages in the amp.

That's a substantial transformer and will have a relatively strong magnetic field around it. The distance between it and the CK2III board, the Alien DAC board, and the myriad low-signal wires are probably no more than 6 inches or so? You're probably still getting some interference. You might want to experiment with shielding the transformer with something from EMF Safety Superstore- "Electromagnetic Field Detection & Protection". .



Just to double check, how would you recommend connecting all the signal grounds? I have quite a wiring diagram right now and it's a little complicated.



The 4 input grounds are connected, and all go to the POT via single wire (after a small HF filter of 680 Ohm resistor in series with signals and a 500pF to ground).

From the pot, I have a ground twisted with EACH of the signals going to the CKK-III board (1 for Left, 1 for Right).

Output Ground of CKK-III goes to the switching headphone jack and muting delay. When there are no headphones connected (switching headphone jack is closed), I have a ground wire going to each GainClone.

Current ground loop breaker scheme: the AlienDAC, CKK-III and both GainClones are connected to the star using the 0.1uF/10Ohm breakers.

If the ground scheme looks ok, I guess all I have to do is join all the ground loop breaker inputs and have only one breaker to the case.

As for transformer interference, yes, it's around 6 inches away from the AlienDAC and CKK-III, and about 10 inches from the input wires.
 
Dec 29, 2007 at 2:36 PM Post #10 of 26
definitely shield that transformer, hook your setup to a soundcard and test with RMAA, and I bet you will see a huge 60hz noise hump. Had the same problem when my STEPS was in an unshielded enclosure.
 
Dec 29, 2007 at 7:17 PM Post #11 of 26
FallenAngel, your description doesn't really match your diagram. I assume the description is more accurate. At the headphone jack, are you switching the ground? Also, you have multiple power supplies in the box. How are the grounds from them connected?
 
Dec 29, 2007 at 9:47 PM Post #12 of 26
In addition to the other suggestions, you might try loosening the transformer bolt and rotating the transformer while listening to the hum... I've found that with toroids, the rotational position can reduce (or enhance) hum. Usually they emit the most EMI from the area where the wires come out, and I see you have these pointed away from the amp pretty much. I've also seen some put the toroid on edge, putting the axis of its EMI radiation off plane with the amp (I guess...).
 
Dec 29, 2007 at 10:24 PM Post #13 of 26
Pars
I'll try that, thanks

amb
The description is correct, the diagram is what I had in mind.
The GrainClone power supply creates a ground which is tied to the GainClone boards. Those in turn are tied to the star via breaker. CKK-III has output ground tied to the star via breaker as well.
Yes, at the headphone jack, the ground is switched as well. Is that a bad thing? (I have to find an innocent looking smiley for this).

LawnGnome
I'll try the RMAA as soon as I figure out how it works and how to connect the damn thing
smily_headphones1.gif


d-cee & TzeYang
I have another MUCH shorter cable that I will try. If there is any improvement, I'll cut that one and make it 6 inches long.
 
Dec 30, 2007 at 12:45 AM Post #14 of 26
Updates:

Pars
Unfortunately, I couldn't tell a difference when I twisted the transformer around.

d-cee & TzeYang
I replaced the cable with a VERY short one and still no change. I also noticed that there is no more hum when I'm using USB vs any RCA input which makes me believe that the AlienDAC is not the problem.

amb
I removed all but one circuit breaker and connected the grounds together. Although the hum is still there, I do not get more nor less when I ground everything to the case, so I have the grounds now connected.

There is one thing that throws me off and I'm not sure about. I have a ground wire going from the headphone jack to each of the GainClones along with a signal. If I attach a signal ground to the star from each GainClone, and also attach the PreAmp (Output ground attached to headphone jack), doesn't that create a ground loop at the star? I tried both and still, I do not notice a difference in amount of hum out of the speakers (there is actually slightly more hum from the right speaker. It has a longer Pot-PreAmp wire and the GainClone is closer to the transformer).

LawnGnome
I'll try RMAA as soon as I figure out how to connect everything together, for now I'm clueless
frown.gif
 
Dec 30, 2007 at 1:49 AM Post #15 of 26
I am unsure about how you actually connected the headphone jack ground, so a diagram would be helpful.

The input side grounds as you describe are probably OK (Input jacks/AlienDAC -> pot -> CK2III).

The power supply grounds should all be tied to a single point near the PSUs (a global star point, but not necessarily connected to the case), and both the gainclones' ground and CK2III's board star point should also be connected to here. The headphone jack and speaker output ground returns should ideally also go here via separate wires.

You can then decide whether to connect this global star point to the case directly or via a ground loop breaker.
 

Users who are viewing this thread

Back
Top