For reference see Jensen Transformers'
AN007 ["covering the causes of noise pickup and practical ways to address them in detail.
Huge thanks for your response and the a great reference document from Bill Whitlock!!![/INDENT]
In additional to the AN007 document (aka "Understanding, Finding, & Eliminating Ground Loops in Audio & Video Systems", Bill Whitlock, 2005), I also came across a slide deck that he presented at an AES conference in 2021:
The Myth and Mystery of Analog Signal Interfaces
That presentation helped me more fully understanding the whitepaper. That said, the whitepaper (AN007) on its own is pure gold.
Ground related noise (loops, USB noise etc.) is a system problem that depends on the details of all the connected equipment, the mains wiring and how everything is connected.
In a home setting the biggest potential problem is ground related noise (loops, USB etc.) although it's far from a certainty. [...]
In the home you probably don't have too much risk of noise pickup from electrical or magnetic sources. [...]. Moving the cable further away from the interference source is usually enough to make it inaudible, even with single ended interconnects. For this reason I wouldn't get too worried about the CMRR part.
I'd go with Whitlock first, and mod the cable only if I encountered a ground related noise issue.
Putting your response (combined with what I understood from reading Whitlock's paper) into my original framing context.
On whether to connect the ground (Whitlock vs. Rane):
Whitlock—who has done extensive research on ground noise and is a leading audio industry Fellow—recommends always connecting the output device's chassis ground with the interconnect shield, regardless of whether the chassis ground on input device is also connected to the shield wire. Further, though Whitlock is well aware of Rane Notes, he does not cite any situations where the RCA sleeve on the unbalanced output device should
not be connected to the shield wire.
Thus:
- Whitlock's wiring (i.e., connecting RCA sleeve & interconnect shield) should be recommended for home studio / home stereo.
- That connection should only be removed after (a) following the outlined diagnostic steps to determine the real source of the issue, and (b) there is no other effective way to solve the ground noise issue.
- Only at that point would the Rane wiring be recommended.
- Whitlock would recommend using an input transformer (with no gain increase) as a better solution to adding a resistor. But that is more expensive (and out of scope of my focus here). Also, if the issue is only CMRR and not ground noise, it is not clear to me whether use of a transformer is any better than using resistors to achieve impedance matching between the Pin 2 and Pin 3 conductors... I'd love comments from others on this aspect!
On maintaining CMRR benefit:
The most common type of RCA to XLR cable (i.e., using a 1-conductor Coax) will work to connect an Unbalanced device to a Balanced device. However, it will provide 0 dB CMRR benefit,
which may (or many not) affect how well this interconnect works in your specific context. For many people, that will be a non-issue.
Using a 2-conductor STP cable (as described by Bill Whitlock or Rane Notes) will automatically provide a CMRR (common mode rejection ratio) of 30 dB; and perhaps even more depending on how the balanced input circuit (i.e., IC chip or transformer) is implemented within the Balanced device.
Achieving the maximum potential CMRR available from the Balanced device's input circuit requires precisely matching the impedance (Z) between the two balanced wires at the Unbalanced device's side. As such, adding a resistor between the RCA Sleeve and the interconnect wire (Pin 3) to more closely match the driver's output impedance (i.e., on the RCA Tip side) will incrementally increase the CMRR. However the actual CMRR increase will depend on a number of factors, including the design of the input balanced circuit to start with, how closely the resistor matched the impedance, the input device's impedance, output device's impedance, etc.
Whitlock does not directly mention adding a resistor to the RCA-end of the interconnect cable to match the impedances. However, he does imply the potential value based on (a) commentary on the benefit of exactly matching impedances, and (b) reference to the value of the Jensen input transformer or their special designed IC chip for balanced input circuits.
Thus:
- If you aren't having any issue with common mode noise being injected into the interconnect in the first place, any properly wired Coax or STP is fine
- That said, using a 2-conductor STP interconnect is always preferred since it provides an automatic 30 dB CMRR (and perhaps even more). In most home audio situations, this should be sufficient assuming proper management and isolation of signal lines vs. AC power lines.
- If common mode noise is still a problem in the 2-conductor interconnect, then adding a resistor to more closely match impedances would increase the CMRR. E.g., cases where the interconnect is much longer (e.g., out to subwoofers) or a shorter interconnect can't be moved away from very noisy RF/EM sources
- Finally, consider use of an input transformer if you are facing significant common mode noise issues (or replace your Unbalanced device with a Balanced device)
On when RCA-side impedance matching could be pragmatically (i.e., potentially noticeably) beneficial:
Probably not for most people, but definitely for some people:
- Start with 2-condcutor STP (per Whitlock or Rane) as best practice
- Add resistors only after determining there is a common noise problem that can't addressed by isolating from the EM/RF source
- Or, go ahead and include a resistor proactively if you have resistors lying around, you like doing such soldering, you know the impedance of the Unbalanced output driver, and you have nothing better to do with your time)
In a hifi setting ground related noise issues are rare enough that the industry has stuck with RCAs but common enough that we keep getting threads from people asking how to solve them.
From what I read, Bill Whitlock considers that a
serious failure on the part of the HiFi industry. From his view, CMRR is an increasingly pressing issue as we have moved from AM Radio and 78 RPM records to modern audio equipment that is now pushing >120 dB SINAD across all components in the audio chain (excluding the speaker itself).
Hifi systems change relatively rarely, so you aren't constantly having to sort such problems and it's harder to justify the small extra expense for balanced.
Fully agree that in a professional setting, there are huge benefits in using balanced audio. That is why it has been the professional standard for a very long time. And more recently, also within a home studio now that the cost of balanced equipment has dropped to very affordable prices.
BUT... (IMHO) I don't believe the lack of frequent audio component changes is a reason not to use balanced audio in a home stereo setting. Especially in view of the increasingly high SINAD ratings of the even the cheapest audio components. And even more so, now that modern DAC and DSP chips themselves are being internally based on balanced audio architectures.
Of course, as always, I would greatly appreciate any additional insights and corrections that others can add. My end goal is to have a good understanding of how the various options (and their trade-offs) apply within the context of a home stereo / home studio setting.
gnd-tech.com
