I'm using KRK Rokits as computer speakers. There's a plethora of social media posts stretching back 12+ years where people are struggling with CPU/GPU-induced electrical noise and have tried everything including DACs with balanced outputs with no satisfactory result.
I myself have had CPU noise for years, across two houses, two different computers (one with two different power supplies), and even with different powered monitors (JBL LSRs in addition to these Rokits). Last night I finally decided to deep dive and make a plan to fix this damned problem. Some findings and observations:
1. https://www.ranecommercial.com/legacy/note110.html : "Here is where special test cable assemblies, shown in Figure 3, really come in handy. These assemblies allow you to connect the shield to chassis ground at the point of entry, or to pin 1, or to lift one end of the shield. The task becomes more difficult when the unit you've isolated has multiple inputs and outputs. On a suspect unit with multiple cables, try various configurations on each connection to find out if special cable assemblies are needed at more than one point."
2. https://www.prosoundweb.com/balanced-and-unbalanced-connections/5/ : "One method of creating a balanced output is so simple it seems almost like cheating, and indeed, some condemn it as such. (It’s a popular comment in online forum “reviews.”) But it’s not cheating at all, and it works quite well. The output signal is connected only between pins 1 (ground) and 2. A resistor equal to the output impedance of the op-amp driving pin 2 is connected between pins 1 and 3. This satisfies the balanced output requirement of equal source impedance for pins 2 and 3. How well it works is a function of how accurately the “dummy” impedance on pin 3 matches the active source impedance, but that’s easy to do with inexpensive 1% tolerance resistors, and most manufacturers who use this output configuration do it right."
3. https://www.sweetwater.com/insync/great-balance-unbalanced-line/ : "At the end where it plugs into the keyboard connect a precision potentiometer pin 3 (cold, or “-“) of your XLR and ground/shield (pin 1). On a TRS this would be between Ring and Sleeve. Connect pin 2 (or Hot) as normal. On a TRS this would be the tip. Often it is easiest to just make up a special cable you can connect in line with the pot on it. Turn the equipment on and turn up the gain on your mixer until you hear a good amount of hum and line noise coming in from the keyboard line. Turn the potentiometer (which will change the resistance between pin 3 (“-“) and ground) and you will hear the hum and noise level change. With some careful work you should be able to find a “sweet spot” where the hum is minimized."
4. https://www.soundonsound.com/sound-advice/q-there-better-balanced-unbalanced-cabling-solution : "The practical problem with this approach lies in trying to squeeze a resistor and capacitor into the very limited amount of space inside a standard TS plug without everything shorting out. It's not impossible if you have decent soldering skills, small components and some patience, but for those who don't, help is at hand in the form of our very own custom-made SOS pseudo-balanced cables. These are available from the SOS shop to UK customers at £19.99 (or £16.99 to SOS subscribers)."
That was 10 years ago and there's no longer any sign of these special diagnostic cables.
-----
Given that this topic keeps resurfacing over and over again; the actual solution varies depending on the architecture at either end; isolation transformers suck unless you spend an arm and a leg; and balanced DACs can be hit-or-miss in these setups... why doesn't someone design and sell a "FAFO box" that can be inserted in-line to test these strategies without messy breadboarding and/or a rat's nest of alligator jumpers?
I can't even find a TRS to 2x TS insert that deviates from the usual practice of dumping each unbalanced channel into TS -- the 4th URL above has some verbiage to suggest that opening ground at one end plus driving TR instead of TS might be enough to fix the problem. But I can't buy a cable or adapter that does this.
Am I barking at the moon here?
I myself have had CPU noise for years, across two houses, two different computers (one with two different power supplies), and even with different powered monitors (JBL LSRs in addition to these Rokits). Last night I finally decided to deep dive and make a plan to fix this damned problem. Some findings and observations:
1. https://www.ranecommercial.com/legacy/note110.html : "Here is where special test cable assemblies, shown in Figure 3, really come in handy. These assemblies allow you to connect the shield to chassis ground at the point of entry, or to pin 1, or to lift one end of the shield. The task becomes more difficult when the unit you've isolated has multiple inputs and outputs. On a suspect unit with multiple cables, try various configurations on each connection to find out if special cable assemblies are needed at more than one point."
2. https://www.prosoundweb.com/balanced-and-unbalanced-connections/5/ : "One method of creating a balanced output is so simple it seems almost like cheating, and indeed, some condemn it as such. (It’s a popular comment in online forum “reviews.”) But it’s not cheating at all, and it works quite well. The output signal is connected only between pins 1 (ground) and 2. A resistor equal to the output impedance of the op-amp driving pin 2 is connected between pins 1 and 3. This satisfies the balanced output requirement of equal source impedance for pins 2 and 3. How well it works is a function of how accurately the “dummy” impedance on pin 3 matches the active source impedance, but that’s easy to do with inexpensive 1% tolerance resistors, and most manufacturers who use this output configuration do it right."
3. https://www.sweetwater.com/insync/great-balance-unbalanced-line/ : "At the end where it plugs into the keyboard connect a precision potentiometer pin 3 (cold, or “-“) of your XLR and ground/shield (pin 1). On a TRS this would be between Ring and Sleeve. Connect pin 2 (or Hot) as normal. On a TRS this would be the tip. Often it is easiest to just make up a special cable you can connect in line with the pot on it. Turn the equipment on and turn up the gain on your mixer until you hear a good amount of hum and line noise coming in from the keyboard line. Turn the potentiometer (which will change the resistance between pin 3 (“-“) and ground) and you will hear the hum and noise level change. With some careful work you should be able to find a “sweet spot” where the hum is minimized."
4. https://www.soundonsound.com/sound-advice/q-there-better-balanced-unbalanced-cabling-solution : "The practical problem with this approach lies in trying to squeeze a resistor and capacitor into the very limited amount of space inside a standard TS plug without everything shorting out. It's not impossible if you have decent soldering skills, small components and some patience, but for those who don't, help is at hand in the form of our very own custom-made SOS pseudo-balanced cables. These are available from the SOS shop to UK customers at £19.99 (or £16.99 to SOS subscribers)."
That was 10 years ago and there's no longer any sign of these special diagnostic cables.
-----
Given that this topic keeps resurfacing over and over again; the actual solution varies depending on the architecture at either end; isolation transformers suck unless you spend an arm and a leg; and balanced DACs can be hit-or-miss in these setups... why doesn't someone design and sell a "FAFO box" that can be inserted in-line to test these strategies without messy breadboarding and/or a rat's nest of alligator jumpers?
I can't even find a TRS to 2x TS insert that deviates from the usual practice of dumping each unbalanced channel into TS -- the 4th URL above has some verbiage to suggest that opening ground at one end plus driving TR instead of TS might be enough to fix the problem. But I can't buy a cable or adapter that does this.
Am I barking at the moon here?
