Some people experience noise when using their front panel jacks, especially when using USB ports. Sometimes it's due to how the PCB dongle in front of the PC Case is wired. Some of these PCBs have the Signal GND of the audio, bridged to the Ground for the USB devices, and then bridged to the Chassis GND.
While the USB ports will probably have no issue sharing the ground with the chassis, they are all meant for "powered" devices. Signal Ground is not exactly the same thing. Now I've seen cases where Signal ground is meant to be connected to the chassis ground, and cases wheres it's not. And sometimes it is connected, but first passes thru a filter or other circuitry.
In my PC Case I have two separate dongles, audio and USB on different sides. But the Audio dongle is still bridged to the chassis via the screws that hold the PCB in.
I don't know what's the official stance of INTEL on their specification, I tried googling, didn't find nothing official. So this is my question here, Does anybody know about this? or something related?
I've seen various guides on hacking the PCB to remove where it's bridged to ground. I did an easy test by unscrewing the PCB and leaving it in mid air, and testing my headphones. With the PCB not connected to the CHASSIS the front audio seemed perfect. But IDK what implications this may have in the long run. IDK if to proceed with modifying the PCB and make it permanent. In theory the sound card would handle the grounding on its end.
Should be fine. I'm a little surprised it makes such a difference though, I had assumed that signal ground (through a decoupling capacitor) was meant to be connected to case ground).
I'm surprised of the difference it makes as well.
After removing a little bit of copper from the PCB, and screwing/connecting it back in. I decided to verify the work with a continuity test. I used my multimeter to check resistance. It was pretty much saying it was shorted, and just read 0 ohms.
I thought I didn't scrape enough, but when re-checked the PCB on it's own, it was how I intended. So I remembered my assumption that it's likely bridged to ground on the sound card end.
With the header disconnected from the front PCB, I tested from the header's ground pin to the chassis, and indeed it was bridged to ground (0 ohm). And I tested from the signal ground of the 3.5mm jack to the same point on the chassis, and it was disconnected, as expected by the change.
Therefore, it is still grounded, but it's now all handled from the motherboard. (this is motherboard integrated sound, Realtek)
I like this statement I found on Wikipedia - Ground loop (electricity):
Generally the analog and digital parts of the circuit are in separate areas of the PCB, with their own ground planes, and these are tied together at a carefully chosen star point. Where analog to digital converters (ADCs) are in use, the star point may have to be at or very close to the ground terminals of the ADC(s).
Good technique calls for all of the assorted grounds to be combined at a single point, thus avoiding "ground loops". This is not so important for digital circuits, but can make a big difference in low-level analog circuits. I would not classify headphone jack circuits as low level.
It is slightly more complicated, the DAC's ground isn't earth ground (since it has no negative rail), so usually there is a decoupling capacitor to prevent any DC ground loop from existing. There can still be a loop at higher frequencies however.
Looking at a random datasheet, the recommendation is to put a large resistance to ground between both the positive and negative terminals of the headphone jack, I suppose so that any fluctuations in the DAC's ground relative to earth ground couple equally into both ends of headphone jack and thus cancel out. It sounds like here they just left out the resistors.