signal and chassis ground wiring with RCA and XLR

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Need to go back to basics here. I've been reading up on the correct and proper way for wiring and grounding. In particular, there are a couple of Rane notes that are a good resource. Number 151 and 110.

151 - Grounding and Shielding Audio Devices
110 - Sound System Interconnection


So chassis needs to be connected to signal ground for best performance, unexpectedly (see note 151). This needs to be done at one point only, such as at the star ground, so that noise and interference that is picked up is not carried through a signal ground wire. Makes sense so far.

RCA seems simple enough. You have the signal pin and you have the signal ground (which is also the signal return). And yet, the 'correct' way to wire a RCA to RCA cable is with a coax cable with the shield connecting the signal ground. How can this be a good idea to use a signal ground as if it were a shield? It contradicts what was already described as the best way to do it.

XLR is more complicated. Is the cold pin the signal return? Will balanced cables still work with just hot and cold and no shield or ground at all? That is the way rane notes describe it as they say pin 1 of the XLR should go directly to chassis. i.e. it is not needed for signal ground as you have the hot and cold pins. I thought you had hot, cold (for common mode rejection only), and the obligatory signal ground which I guess was also for the hot and cold signal return as well. Is this not true?

Please explain :confused:
 
That sure can be a confusing matter, but those Rane notes are very good. However, the application notes mostly concern large audio systems like studios, concert sound systems and so on where it is impossible or at least very hard to avoid ground loops. In a small hifi system unbalanced interconnections are not as problematic as those application notes make them seem.

As for your question about RCA cables, there are not many other sensible ways to make the cables. It is an inherent drawback of unbalanced interconnections that signal return and ground are not separated but one can try doing one's best. To minimize noise from any current that may flow between units, for example coming from Y capacitors or flowing in a ground loop, resistance needs to be minimized. You also want to screen the signal lead from electrostatic pick up of noise. Using the shield as signal return gives both low resistance and screening from electrostatic pick up, so it's usually done that way.

As for balanced interconnections, ideally the signal is extracted as the difference in voltage between hot and cold. The ground of the cable is primarily used to screen the cable from electrostatic pick up of noise and RF ingress. It is an extension of the equipment cases, as the application notes put it. It also serves to equalize the ground potential of the two pieces of equipment so that the common mode range of the receiver may not be exceeded, something that would be likely to happen if there were not a ground connection between them.

The main point of the Rane application notes is that if the "pin 1 problem" is designed away, so that shield currents cannot flow through audio paths, one should be able to connect all the equipment in a large audio system with standard XLR-XLR cables without hum problems. There will be a lot of "ground loops" but that won't be a problem. A great relief to installers of large audio systems, not having to worry about these matters.

For equipment that has the "pin 1 problem", cutting the ground connection on one end is a common solution to ground loop hum problems. This works because the ground lead of the cable isn't needed by the balanced signal, but RF ingress immunity will suffer a bit as a hole is made in the previously continuous shielding. For unbalanced signals cutting the ground does not work at all however because it's also the signal return so you'll have to resort to either reducing the resistance of the interconnection ground or breaking the ground loop outside of the signal-return loop.

While not optimal from an RF immunity perspective, a common remedy to be able to use unbalanced interconnects with grounded equipment, while avoiding ground loops, is not to have a direct connection between signal ground and chassis ground. If this is done, input and output grounds are insulated from chassis and connected to signal ground instead.

With meticulous attention to layout details this may not be needed however. Transformer stray fields can be avoided by smart layout of the interconnection connectors, mains entry and ground attachment...
 
Last edited:
Nice reply, thx.

Going back to just hi-fi and not pro audio, how about using 2 core wire for unbalanced RCA instead of using coax? Just a twisted pair for example.

I know you said that using a screen as signal return will help with screening from electrostatic pick up. But how does that work? If the return signal is flowing directly through the screen, and so is the noise pick up, then why doesn't it reduce performance?
 
A noise current injected onto the signal lead will, besides the voltage drop from cable resistance, drop voltage over the output impedance of the source equipment (parallelled with input impedance of the receiver if one wants to be picky) while a current injected onto the shield will not. Thus the signal lead is much more sensitive to capacitively coupled noise, especially if the impedances are high.
 
I was searching around and found this thread, which seems to relate quite well to my present quandry... I am currently assembling a classD amp for my subwoofers and have the option of grounding the power supply and amp module grounds to the chassis, or not. The manufacturer told me there is no need to ground them to the chassis so I'm inclined not to, but I'm not sure which is best. On top of this, I have balanced and unbalanced connections to consider.

The signal chain is: receiver preamp (unbalanced; signal grounded shield) > dcx2496 (balanced; chassis grounded shield) > amps (unbalanced).

So, assuming I don't connect the chassis & signal grounds on the amps, would the best interconnection method be #17 for the preamp>dcx and #4 for the dcx>amps, or should the latter be #6 (which seems to be the usual recommendation). Here are the diagrams for those options:

An externally hosted image should be here but it was not working when we last tested it.


- Doug
 
So, assuming I don't connect the chassis & signal grounds on the amps

They will have to join somewhere for safety.
I will assume all your gear has 3 prong power (chassis are grounded thru power cord). Then use #6 conection the #4 causes a ground loop: power gnd (chassis) of DCX thru shield back to power grnd thru the amp chassis.

When they started using balanced interfaces they where all transformers, which where perfect isolators and made interface life simple.(the studio I worked in had a $500k Neve mixer that no matter what we did with the connectors or grounding, still had inputs that hummed. We ended up installing input transformers (48 of them at $100 a pop) and this completely solved the ground hum) The move to solid state has seen the development of many different balancing I/O circuits all with there own advantages and disadvantages (some are even called psuedo balanced). Without studying the actual schematics its hard to be sure exactly of how one type of output will behave with a certain input (especially mixing balanced and un).

Sorry this dosnt really help, I guess my point is interconects and grounding get complicated fast in a balanced/unbalanced system.
 
They will have to join somewhere for safety.

That's what I thought, but apparently the mfg doesn't think so.

I will assume all your gear has 3 prong power (chassis are grounded thru power cord). Then use #6 conection the #4 causes a ground loop: power gnd (chassis) of DCX thru shield back to power grnd thru the amp chassis.

Yes, all 3 prongs, all chassis grounded to power ground. One advantage of not connecting the signal ground to power ground in the amp is that it would avoid ground loops like the one you pointed out. However if I use an interconnect like #4 (or unbalanced direct from the preamp), the signal ground will still become connected to power ground through the interconnects and the DCX.

- Doug
 
Unfortunately, it is hard to find unbalanced equipment of this type. Many people have used the DCX in the same way as I am to control multiple subs. I have read one or two complaints about the extra gain that the DCX adds, but most seem to find that it works fine. This shouldn't be an issue for me since I can control the sub out level on the receiver as well as output levels on the DCX. There is a moderate impedance mismatch between my receiver sub output (1.2k ohms) and the DCX analog inputs (22k ohms). So, using a transformer on the input might not be such a bad idea. The commercial ones are a bit pricey, though, and I'm a bit concerned about low frequency losses with those. The output impedance of the DCX (160 ohms) and input impedance of the amps (rated only as "up to 600 ohms") are in the same ball park.

I have pretty much decided to keep the signal ground separate form the chassis ground on the amps at this point. If I run into issues I can always connect them, and rewire the interconnects, if necessary.

- Doug
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.