Isolating Digital Noise, need help.

Hi Lowrider,

I’ve just read through this thread. First, let me summarize some of the key points I’ve perceived:

1)The problem does not occur in any of the following cases:

(a)The CDP’s stock power cord is used instead of the Audience power cord.
(b)The CDP is connected to a more distant outlet. When you did this, how was the connection implemented (e.g., with a long extension cable attached to the Audience cord)?
(c)A different CDP is used.
(d)A different preamp (the prior one) is used.

2)The CDP’s stock power cord does not provide the open soundstaging and depth of the Audience power cord.

3)You will be demo’ing an SR power cord.

4)No ground loops are present, because:
(a)There is a Jensen transformer between the CDP and the preamp.
(b)There is a cheater plug on the preamp’s power plug, which eliminates the possibility of a ground loop between the preamp and the power amp.

5)Ideally you would prefer a solution that in addition to resolving the issue isolates the AC power provided to the digital components from the AC power provided to the analog components, because you anticipate purchasing a DAC.

Questions and suggestions:

1)If you haven’t already, it might be worthwhile trying the CDP’s stock power cord on the preamp. A reasonable speculation would be that the CDP’s stock power cord, as well as the extension cord I presume you used when you connected the CDP to a different outlet, resulted in higher series inductance than when using the Audience cord into the normal outlet, which in turn provided some degree of filtering of high frequency digital noise, that is causing the problem by coupling into the preamp through its power cord. So perhaps putting the CDP’s stock power cord on the preamp would resolve the problem, while not causing the sonic degradation you observed when using that cord on the CDP.

2)Are you sure that when you changed between the Audience and stock power cords on the CDP, and when you tried connecting the CDP to a distant outlet, that the physical proximity between those power cords and the preamp was not changed significantly? And that in all of those cases the CDP’s power cord has not been positioned close to the preamp or to the preamp’s power cord or interconnect cables?

3)For that matter, how close is the CDP to the preamp, and if they are close together have you tried moving them further apart? Although it seems likely that the noise is coupling through the power wiring, I don’t feel 100% certain of that at this point, and I wouldn’t rule out the possibility that radiated interference is contributing.

4)Regarding power isolation or filtering between digital and analog components in anticipation of obtaining a DAC, my feeling is that perhaps you should cross that bridge when you come to it, and not necessarily as part of the process of resolving this problem. For one thing, the optimal approach at that time may be DAC-dependent, requiring greater or lesser degrees of isolation or filtering depending on the particular DAC. And who knows, it may not even be necessary to provide any.

Good luck. Best regards,
-- Al

Lowrider, I had previously suggested trying the CDP’s stock power cord on the preamp. However, given that it is now evident that the CDP’s stock power cord doesn’t fully eliminate the problem when used on the CDP, what I would suggest is that you try connecting the CDP with the Audience power cord in series with the long extension cable (which eliminated the problem when plugged into a distant outlet), plugged into the **SAME** AC socket you normally use for the CDP.

And if that resolves the problem but also results in any adverse sonic effects, try putting the extension cord in series with the preamp’s power cord instead, plugging it into the same place as usual.

Keep in mind that when you eliminated the problem by connecting the CDP to the distant outlet you changed two things at once, namely using the distant outlet, and putting the extension cable and its undoubtedly significant inductance in series with the CDP’s power cord. At this point we don’t know which of those variables eliminated the problem, but it very conceivably could be either of them (or both).

Best regards,
-- Al

I believe this proves that the ARC CDP is introducing noise into the shared dedicated line and needs to be isolated.

Yes, but per my previous post perhaps the extension cord itself would provide sufficient isolation, when plugged into the outlet that is normally used. Seems worth trying, on both the CDP (the source of the noise), and the preamp (which is probably where the noise is causing the trouble).

The amp is now known to have a grounding issue, but is it possible that that this low-level hum could be due to the fact that the second line’s grounding is unknown?

That is very conceivable. Jim (Jea48) has previously provided us with a link to a paper by Bill Whitlock (the Jensen Transformer person, although he may be retired now) in which it is explained that ground loop hum is usually caused or contributed to by coupling of EMI between the hot & neutral wires and the safety ground wire in the house wiring. Which occurs very minimally with Romex, due to its symmetrical design, but occurs to a much greater degree with wires in conduit. Is there any way to determine what kind of wiring is used between that outlet and the breaker panel?

Also, Jim (Jea48) makes a good point about jumpers. Given that there is no balanced/unbalanced switch on the amp, and that it has an unbalanced internal signal path, I suspect that jumpering pins 1 and 3 together on each XLR connector might be helpful, with respect to hum and/or sonics. Check for continuity between the center pin of each RCA connector and pin 2 of the corresponding XLR connector. Also check for continuity between the ground sleeve of each RCA connector and pin 1 of the corresponding XLR connector. I suspect you’ll measure close to zero ohms in both cases. If so, I would suggest that at some point you obtain some solid core wire of a gauge that would provide a snug fit and use it to jumper pins 1 and 3 on each XLR connector.

BTW, "OL" is most likely derived from either "out-of-limits" or "overload," and when you are measuring resistance signifies that the resistance is greater than the 20 million ohm upper limit of what the particular meter can measure. In other words, an open circuit.

Best regards,
-- Al

Thanks, Jim (Jea48). Note, though, that the transformer coupled outputs shown on the schematic you linked to are the digital outputs, not the analog outputs which Lowrider is using.  Although that schematic is for the CD2, rather than the CD3 MkII he is using.

Regarding overload, the UV-1 doesn't appear to have an input overload point specified, but I'd be surprised if it couldn't handle 2.7 volts.  And even more so given that it provides a 6 dB input attenuator.  And in any event, if it were overloading I'm sure Lowrider, who is clearly a very perceptive listener, would have noticed the distortion that would have resulted.

Regarding Mr. Crump's statement, I would add that in the case of most power amplifiers, at least, one certainly does NOT want to attenuate everything above 60 Hz. As has been pointed out in past threads by Ralph/Atmasphere among several other members, since most power amplifiers draw current in narrow high amplitude spikes, rather than continuously throughout each period of the 60 Hz waveform, their AC supply needs to be provided in a manner that is not bandwidth limited below something on the order of tens of kHz. That requirement being one of the reasons upgraded power cords can often provide superior performance compared to stock power cords, for power amplifiers at least, because many of them provide significantly greater bandwidth than most stock power cords.

Also, I just Googled "inductance of Romex" and found a not necessarily authoritative post in a thread elsewhere stating that it is about 1.45 uH (micro-Henries) per foot.  That is indeed high, but for run lengths that are not particularly long I suppose is probably low enough to be compatible with the kinds of bandwidths Ralph refers to, perhaps depending on the particular amplifier.  While at the same time being high enough to provide significant attenuation of digital noise, which typically has a good deal of its frequency content in the MHz range and higher.  

Best regards,
-- Al
 

Thanks Jim. As with the CD2 only the digital outputs are transformer coupled. And I don’t see any issues with the unbalanced analog outputs Lowrider is using, or with using a Jensen transformer between those outputs and the preamp (assuming cable capacitance at the output of the transformer is kept low, which I believe he is doing).

Best regards,
-- Al

Thanks, Jim. Regarding ...

UL says the equipment ground must be connected, bonded, directly to the chassis doesn’t it?

... Shortly after submitting my post, and shortly before you responded, I revised the paragraph you quoted from as follows:

In any event, it is strange indeed that the Sunfire amp connects signal ground and safety ground together but not to chassis. If it happens to be connecting the grounds to chassis via a capacitor, though, it would explain why the meter reads OL between the grounds and chassis. But it still would not explain the safety issue that would presumably result from the lack of a direct connection between safety ground and chassis.

Best regards,
-- Al

Jim (Jea48), you provided an excerpt and a link to Mr. Hansen’s statement in the last post on this page. However, I would not interpret his statement to mean that signal ground should be connected **directly** to the chassis, as that can be conducive to ground loop issues. Preferably signal ground should be connected to the chassis through a low value resistor (e.g., 10 to 100 ohms). As you’ve probably seen, for example, a lot of ARC designs use 10 ohms.

My VAC amp, btw, has a three position switch controlling that connection. One position connects signal ground and chassis directly; one position is described as connecting them at audio and higher frequencies but not at DC (presumably via a high value capacitor, perhaps in combination with a resistor); and one position is described as connecting them only at RF frequencies (presumably via a low value capacitor, perhaps in combination with a resistor). The AF and RF positions seem to sound slightly better in my system than the DC position, and I’ve settled on the AF position (which the manual recommends as the one to start with).

In any event, it is strange indeed that the Sunfire amp connects signal ground and safety ground together but not to chassis. If it happens to be connecting the grounds to chassis via a capacitor, though, it would explain why the meter reads OL between the grounds and chassis. But it still would not explain the safety issue that would presumably result from the lack of a direct connection between safety ground and chassis.

continuity of each RCA center pin and pin 2 of XLR... OL

Check for continuity between the RCA center pin and pin 3 of the XLR connector. If that measures near zero ohms, you would want to put the jumper between pins 1 and 2 of the XLR connector.

If that also reads OL, however, it would probably mean that the XLR and RCA inputs are received by separate receiver stages, in which case it would be ideal (although perhaps neither necessary nor practical) to jumper both pins 2 and 3 to pin 1. Or, in that situation (both pin 2 and pin 3 read OL relative to the RCA center pin) you could try jumpering each possible combination of two pins together (1 to 2; 1 to 3; 2 to 3), assessing the three possibilities one at a time.

Best regards,
-- Al

Good comments by Jim (Jea48).

As far as the readings I'm getting, do I need a better multi-meter?

Nothing that has been said so far suggests that to me.  See my comments from yesterday about how to try the jumpers if XLR pins 2 and 3 both measure open with respect to the RCA center pin, which now appears to be the case.

Best regards,
-- Al