Why are my USB audio devices flaky?

[amano]

Before buying anything new, I would first check BIOS settings regarding USB, and make sure the mainboard drivers are up-to-date. USB Audio is rather timing critical, so it’s easily disturbed if the system is misconfigured.

I enabled "ErP ready" which shouldn't affect USB ports while the computer is turned on. I didn't find anything else that seemed relevant to USB timing.

 

[amano]

Another thing I suspect is unstable physical connection between port and plug because the connection didn't seem particularly sturdy.

After vacuuming my room, I tended to notice USB disconnect issues especially for USB audio devices because my vacuum cleaner can pull usb cables slightly sometimes. Perhaps, I should change my vacuum technique.

The type-A connection felt loose and flimsy regardless of my vacuum technique.

Perhaps, plugging USB audio devices into the ports on the top side of the case instead of the ports on the back of the case can improve stability of the physical connection.

The physical connection is flimsy, or the mainboard's internal USB hubs don't actually deliver 0.5A consistently to all ports. If the mainboard's design is poor, then a new power supply will not help. Raspberry Pi was known for not delivering enough power to all USB ports.

 

[amano]

After vacuuming dust in my computer case, my USB audio devices stopped being sensitive to nudges on USB plugs.

Dust buildup was the real issue? I cancelled the order on powered USB hub.

I think USB audio devices may be more sensitive to electrical issues than usb keyboard and usb mouse.

 

[kyuu]

After vacuuming dust in my computer case, my USB audio devices stopped being sensitive to nudges on USB plugs.

Dust buildup was the real issue? I cancelled the order on powered USB hub.

... How much dust was in the case?

 

[amano]

... How much dust was in the case?

Not a huge amount, but it was there. Vacuuming dust eliminated RAM errors before. Vacuuming dust around cables seems to improve system stability around especially USB.

Eliminating dust in CPU heatsink is a trouble because I don't want to disassemble it and reassemble it.

 

[voodooless]

Make sure you have positive pressure in your case, so have more air blowing in, than out. That will slightly help you with the dust accumulation in your case.

 

[Philbo King]

Another thing I suspect is unstable physical connection between port and plug because the connection didn't seem particularly sturdy.

After vacuuming my room, I tended to notice USB disconnect issues especially for USB audio devices because my vacuum cleaner can pull usb cables slightly sometimes. Perhaps, I should change my vacuum technique.

The type-A connection felt loose and flimsy regardless of my vacuum technique.

Perhaps, plugging USB audio devices into the ports on the top side of the case instead of the ports on the back of the case can improve stability of the physical connection.

The physical connection is flimsy, or the mainboard's internal USB hubs don't actually deliver 0.5A consistently to all ports. If the mainboard's design is poor, then a new power supply will not help. Raspberry Pi was known for not delivering enough power to all USB ports.

I've heard of some PCs that have a total of 0.5A (USB1) for the whole computer. So in some instances a powered hub is necessary.

On my own 15 year old PC, I have these USB devices:
- 2 hard drives,
- a single RF dongle for USB mouse/keyboard/trackball,
- a 4-port midi interface,
- a Presonus 18 in/out audio interface,
- a couple thumb drives,
- 2 cords devoted to USB charging,
- maybe a couple more I don't recall.

So I added a USB3 powered hub, mostly for amperage capability, since my elderly PC is USB2 only. It works great except sometimes the (cheap Chinese) midi interface locks up till I do a disconnect/reconnect, after a reboot.

 

[amano]

It works great except sometimes the (cheap Chinese) midi interface locks up till I do a disconnect/reconnect, after a reboot.

It's time to buy a better midi interface.

 

[boxerfan88]

I entered UEFI settings. The voltage level of 5V rail was 4.928 volt. The voltage level of 3.3V rail was 2.9~3 which was in red color.

Not good at all. Clear warning from your BIOS. That's more than 10% off. Clearly something is not working well inside PSU.

I don't know whether UEFI sensors are wrong or I need to buy a new power supply.

If your PC PSU is >6/7/8 years old, and those smallish 400W type, probably you should consider replacing.
(IME those >= 650W type tend to be more robust and last much longer)

 

[amano]

I avoid ETA/80PLUS-certified activePFC power supplies which tend to inject a lot of EMI into the house wires that affect my amplifier.

Non-certified activePFC power supplies inject less.

Power supplies without activePFC inject even less, but you may pay more for electricity due to lower power factor if power company charges you for lower power factor. However, unless you are running a bitcoin mining farm, you aren't going to worry about electricity bills.

Clean power in the house wires can help with audio quality.

 

[AnalogSteph]

I have no words. So effectively your choice of power supplies is between "15+ years old" and "real POS" (potentially both)? That is, let's say, less than ideal. More so when you can't recap the darn things yourself. (Around these parts, new supplies like that haven't even been legal for about a decade.)

May I suggest sorting out the ground loop issue(s) in your system the proper way? You're just putting a massive band-aid on the problem with this ridiculousness. Under no circumstances should hoop-jumping of the magnitude be necessary. It is a clear indication that your setup has a problem with common-mode voltage diffferences being made audible. Being able to eavesdrop on your power supply currents should not normally be a feature.

Best make a topological map of your devices and their connections (including mains). Your trouble spots should be revealing themselves upon close inspection. For starters, if your PC has a USB connection to a DAC with unbalanced output and the audio setup is earthed somewhere, I would look into either using a Toslink connection or a USB isolator (e.g. Topping HS02 or similar).

 

[Philbo King]

It's time to buy a better midi interface.

It's just a minor alteration of my startup sequence.

 

[amano]

So effectively your choice of power supplies is between "15+ years old" and "real POS" (potentially both)?

ActivePFC doesn't mean quality. 80PLUS and ETA certifications exist to certify a certain level of efficiency, not quality.

Power efficiency doesn't mean quality. Manufacturers are still releasing power supplies without ActivePFC.

 

[mcdn]

ActivePFC doesn't mean quality. 80PLUS and ETA certifications exist to certify a certain level of efficiency, not quality.

Power efficiency doesn't mean quality. Manufacturers are still releasing power supplies without ActivePFC.

You should listen to @AnalogSteph . So many of these threads start the same way with a "simple" problem. And then after some help and clarification it turns out we have dust/poor connections/clear warnings on PSU difficulties in the BIOS/non-mainstream OS/ancient & deliberately out of spec PSUs/audible mains interference with an amp/whatever. I mean it would have been nice to know all of that up front because none of it is normal.

 

[amano]

audible mains interference with an amp

I didn't say it was audible. But, music studios generally clean their power with an expensive power conditioner. I clean my power more cheaply by avoiding dirty power supplies. I am not a professional music studio, and there are cheaper ways to clean the power.

it would have been nice to know all of that up front because none of it is normal.

Diagnosing all of that would require specific knowledge. I found dust problem through trials and errors. I couldn't have known the dust problem without trials and errors.

ancient & deliberately out of spec PSUs

It isn't ancient or out of spec? I bought a cheap PSU that lived beyond its 3-year warrantee. This time, I'm thinking of buying a more expensive power supply with 5-year warrantee and without ActivePFC. According to my measurement, ActivePFC circuit creates high-frequency noise. Instead of ActivePFC, a power supply can use other technologies that achieve reasonable levels of power efficiency. Instead of trying a bunch of ActivePFC power supplies and finding one that sufficiently filters out ActivePFC noise, I buy a reasonably priced power supply with reasonable power efficiency and much less EMI. ActivePFC isn't a perfect technology that doesn't have trade-offs. It increases power efficiency, but it doesn't care about EMI because customers generally don't care about EMI.

I try to clean my power even if dirty power is not audible if cleaning power is cheap and doesn't take much time.

 

[amano]

You should listen to @AnalogSteph .

I listen to myself first even if I'm wrong because I'm the only one who can iterate fast enough to learn fast enough. I couldn't pay him enough to care enough about my situation and be with me 24/7. I might listen to him closely if he was in my team and worked full-time with me and actually cared about me.

Everyone is forced to listen to themselves first. How do you know who to trust? You have to trust your judgment on who to trust. So, you don't escape the problem of your own discernment. I trust myself enough because I spent quite a bit of time on my audio issue. No one else is going to sit with me on "my" audio issue. I have to be the expert on "my" problems.

There is no external authority. All authority stems from oneself. You merely delegate or outsource your own authority to others. If you don't trust yourself and thus your own judgment, you can't trust others, either.

AnalogSteph couldn't have known about the dust problem because he doesn't have physical access. He wouldn't spend hours with my computer and go through trials and errors to figure out the dust problem. I am the only one who will do that.

 

[amano]

(Around these parts, new supplies like that haven't even been legal for about a decade.)

ActivePFC is not legally required everywhere on earth. I don't live in your country. I suspect it is not strictly required in the legal sense in most places.

 

[mcdn]

Please read @AnalogSteph's post and work through the recommendations

 

[amano]

Please read @AnalogSteph's post and work through the recommendations

I already addressed concerns mentioned in his comment a long time ago. There is no ground loop in my audio system.

Everything is grounded through my computer's grounded power supply. Anything other than my computer doesn't have a grounded power supply. Apart from USB audio devices dropping in and out from time to time, I didn't have any other issue.

My little obsession with clean power exists to get a little peace of mind regardless of audibility. There are japanese audiophiles who install their own power utility poles to get clean power. I don't do that, but I just buy cleaner power supplies whenever I get to buy a new power supply. There are amplifiers that don't have great SINAD scores, but people don't really hear distortions from lower SINAD amplifiers. Noise might be audible as hiss in near-field listening scenarios. Hiss is not audible at my near-field listening position.

I'd also say that a non-activePFC power supply isn't automatically a piece of shit. At least, in my country, there are manufacturers that release non-activePFC power supplies at various price points.

 

[voodooless]

I didn't say it was audible

Then why the:

I avoid ETA/80PLUS-certified activePFC power supplies which tend to inject a lot of EMI into the house wires that affect my amplifier.

If it’s not audible, it doesn’t affect your amplifier..