From Newsgroup: alt.comp.os.windows-11
On Sat, 10/10/2026 10:24 AM, sticks wrote:
On 10/10/2026 3:26 AM, Andy Burns wrote:
sticks wrote:
I came across a suggestion of adding a USB DAC Amp and External Sound Card. This completely takes the sound away from the motherboard audio and was supposed to eliminate the Realtek popping speaker problem. I decided to give it a go and bought this one for only $17:
<https://us.creative.com/p/sound-blaster/sound-blaster-play-3>
Finally, the popping is gone and never to return. It also gives the speakers more power so it does get louder. Super easy to install and use. Cheap. Solved the problem!
I bought a cheapish USB DAC a couple of years ago, as the motherboard audio was picking up all the chirps, pops and whines from within the computer, it worked ... I no longer use that DAC (wonder where it is?) because I ended-up buying a USB "audio interface" basically a two stereo DACs and ADCs and a pre-amp with mixer and headphone output all in a neat little box.
Glad you also found a working solution. It surprised me that other than you I've not heard other complaints here on either the popping I got or the also common box noises you experienced with using on-board motherboard sound. It drove me crazy and I don't know how you could continue with it happening.
When you're picking up "mouse noises" on a motherboard, that's
a layout issue for the analog circuitry of the HDAudio sound chip.
It was "cured" to some extent, by placing a "moat" around the analog island. This can be done with via-stitching to the ground plan. Sometimes
a graphic is applied to the silk screen, to highlight the design has
a moat.
At one time, motherboards seemed to be designed, without
regard to checking SNR on the audio. Boards go through
three design cycles, and a motherboard is "released" on
the 3rd cycle and is supposed to be "modification wire free".
Any mods discovered in the 2nd cycle, are supposed to be put
into the 3rd and final layout cycle.
While a number of motherboard subsystems can be "class-designed
by an expert" and then given to each motherboard designer as
a finished good (VCore) in a sense, the audio performance is not guaranteed when you use an old audio design. The electrical connections to the
island can incite noise problems, and these really need to be
checked and dealt with before the third cycle.
In addition to using a moat, analog designs can also use
a "metal can" and on a plug in PCI card, this could buy
you an extra 10dB or so on noise floor. That's how a
Sound Blaster noise can be -113dBm while a more cheesy
sound circuit (pretty well no precautions at all),
might be -70 to -80 or so. With a -113 number, you can
listen to classical music, and on quiet passages, not
have to worry about computer digital noise overriding
the concert hall noise figure.
It's because the sound chip runs off 0V and +5V, that
there has to be a potential for a "pop" at startup.
If the sound chip was bipolar powered (+5V and -5V),
then the center of the range is 0V. And that would not
pop at startup. Otherwise, you have to "leak out"
the transition to 2.5V (DC resting point), and as long
as the edge rate is slow enough, the speaker cone will
follow it.
Paul
--- Synchronet 3.21d-Linux NewsLink 1.2