PCIe USB card with "clean" power

[Dilettante]

While trying to figure out the issues with noise on USB bus, I have realised it might be due to the widely reported fucked-up bad implementation of USB on X570/B550 chipsets.

After a quick study of what's currently available from various factories in CN, I have narrowed it down to the following options.

#1

Based on the D720201 host controller, no external power needed.

#2

Similar, but with the distinctly different layout. Question is which one is likely to provide "cleanest" power supply to an interface.
This one has holes for the radiator, despite the very same chip.

#3

And the last one, with the external power supply via 4-pin MOLEX. Note two different host adapters.

---

An interesting quote from a seller at AliExpress:

Question 1: Why is power-free better than external power supply?

Answer: External power supply uses wires to connect external 5V to the interface. Voltage loss will occur
through wires. The larger the load, the greater the loss, the more unstable there will be.

The power-free board takes 12V. After voltage reduction, interference removal and voltage stabilization,
the output is stable 5V, and loss compensation is added, so there will be no shortage of power supply.

Now the ultimate question: which one will be best for the given purpose? Expert opinions are very welcome.

 

[Lambda]

I have realised it might be due to the widely reported fucked-up bad implementation of USB on X570/B550 chipsets.

This has nothing to do with Chipset.
It's a EMI RFI issue caused by bad ground (Manibord/conector to PC case. or badly shielded pc case or a combination of this.

Get the Topping Isolator.

How exaclt are you even recording this? what is connected to the interface as "antenna"

question: which one will be best for the given purpose?

impossible to say from the information we have.
Maybe the one with the best grounding from USB GND to PCI bracket grounding.

Is the bracket of a graphics card grounded?

I would guess so, but I want to make sure. Edit: After measuring with a multimeter, I suspect that brackets (of most graphics cards) are grounded to the graphics card's ground-plane. I measured 0 ohm

electronics.stackexchange.com

but we don't know how they have implemented it and we don't know waht cases your using and if it even has Proper grounding to mainboard/power supply or if everything is Painted or plastic

 

[Count Arthur]

I have had two X570 motherboards and haven't had any USB DAC sound issues with either of them, so I don't know how widespread the problem with the "widely reported bad implementation of USB on X570/B550 chipsets" really is.

I did stumble across this: https://hifigo.com/collections/othe...face-expansion-card-crystek-femtosecond-clock, but I thought it seemed rather expensive.

 

[somebodyelse]

This has nothing to do with Chipset.
It's a EMI RFI issue caused by bad ground (Manibord/conector to PC case. or badly shielded pc case or a combination of this.

Get the Topping Isolator.

How exaclt are you even recording this? what is connected to the interface as "antenna"


impossible to say from the information we have.
Maybe the one with the best grounding from USB GND to PCI bracket grounding.

Is the bracket of a graphics card grounded?

I would guess so, but I want to make sure. Edit: After measuring with a multimeter, I suspect that brackets (of most graphics cards) are grounded to the graphics card's ground-plane. I measured 0 ohm

electronics.stackexchange.com

but we don't know how they have implemented it and we don't know waht cases your using and if it even has Proper grounding to mainboard/power supply or if everything is Painted or plastic

It could also be USB 0V being modulated by variable current demand somewhere on the mobo between the USB port and the PSU. In this case a board with a power connector might help. The isolator seems like a more reliable option though.

Is this for the Audient iD4? Balanced outputs usually avoid USB-related noise issues.

 

[AnalogSteph]

Be warned that old USB 3.0 controllers on modern hardware may be a can of worms. The ASMedia ASM1142 on an Asus Xonar SE not only did not like PCIe ASPM but also is buggy and claims to support 64-bit addressing when in fact it doesn't, resulting in bizarre system behavior. This would be entirely incompatible with the Resizeable BAR feature for modern GPUs.

The trusty NEC/Renesas µPD720201 should be pretty solid assuming you have the last firmware on it, which is 2.0.2.6 from 2013 (still a potential issue as recently as 2020). Some fiddling with ASPM states may still be required if it refuses to wake up properly. It's a 10-year-old USB 3.0 controller, so yeah.

To my understanding the Ryzen USB issues seem to have been largely solved with recent AGESA versions. After all the hoopla I didn't quite trust matters though and went with Team Blue. My old hardware was literally ancient (socket 775 with DDR2), so even the ill-reputed 11th gen meant a decade's worth of progress.

 

[Lambda]

It could also be USB 0V being modulated by variable current demand somewhere on the mobo between the USB port and the PSU.

Yes, that's what i meant and said to OP before.

In this case a board with a power connector might help.

Don't think the long cable is low enough impedance work reliably for this kind of issue.
also it might be another ground loop inside the PC to pick up interference.
That's why Low impedance ground paths trough the cases and "IO shield" are important against this kind of EMI.

 

[Dilettante]

This has nothing to do with Chipset.
It's a EMI RFI issue caused by bad ground (Manibord/conector to PC case. or badly shielded pc case or a combination of this.

We might be referring to different things. AMD issues were widely reported, at some point acknowledges by AMD themselves and somewhat fixed with the recent AGESA versions. Not completely though.

How exaclt are you even recording this? what is connected to the interface as "antenna"

Assuming you are talking about the USB bus noise topic: DBX 286 powered off, TRS jack into ID4 mk. II, gain setting on the interface is irrelevant as issue can be reproduced regardless of it being at minimum (line input) or maximum, just the noise floor changes.

Proper grounding to mainboard/power supply or if everything is Painted or plastic

Bog standard ATX board and black case... Painted black case made of (presumably) aluminium or some other alloy. The pointing is thorough, I see no exposed metal.

Get the Topping Isolator.

So... No chance of running ID4 II on 12Mbps? Annoying to say the least, but I appreciate the suggestion.

I have had two X570 motherboards and haven't had any USB DAC sound issues with either of them, so I don't know how widespread the problem with the "widely reported bad implementation of USB on X570/B550 chipsets" really is.

Thousands of messages across all the Internet. In a nutshell:

1. USB 2 ports don't wake-up after an exit from S3 suspend.
   This happens on motherboards where USB2 are arranged in a block of 4.
2. Issues with the on-board and external sound when running PCIe 4.0 devices.
   I use 4.0 NVMe drives and GPU. AGESA updates has resolved keyboard/mouse drops, but not the audio.

It could also be USB 0V being modulated by variable current demand somewhere on the mobo between the USB port and the PSU. In this case a board with a power connector might help. The isolator seems like a more reliable option though.

Hmm.... Given that the situation improves by a lot when CPU is under lot, will this assumption be still valid? If so, I might actually go with the powered card and draw some from SATA cable.

Is this for the Audient iD4? Balanced outputs usually avoid USB-related noise issues.

2nd version of ID4, yes.

Be warned that old USB 3.0 controllers on modern hardware may be a can of worms. The ASMedia ASM1142 on an Asus Xonar SE not only did not like PCIe ASPM but also is buggy and claims to support 64-bit addressing when in fact it doesn't, resulting in bizarre system behavior. This would be entirely incompatible with the Resizeable BAR feature for modern GPUs.

I actually do have ASPM L1/L2 and BAR enabled. Might try with both disabled... Stabbing in the dark here.

To my understanding the Ryzen USB issues seem to have been largely solved with recent AGESA versions.

I wish, v1.2.0.5 is still problematic. Waiting for v1.2.0.7 to stabilise and be released under non-beta BIOS.

Some fiddling with ASPM states may still be required if it refuses to wake up properly

Could you elaborate?

 

[Dunring]

I got the Febsmart 2 port USB card on Amazon for 15 bucks and it works excellent, completely quiet noise floor. If your case has vents on the bottom, flip the power supply over and put the card in the bottom slot up against the shielded side away from the video card. I glued rubber door stops under it as feet to raise the case up for better airflow with loc-tight gel.

 

[Katji]

I have had two X570 motherboards and haven't had any USB DAC sound issues with either of them, so I don't know how widespread the problem with the "widely reported bad implementation of USB on X570/B550 chipsets" really is.

What brand of motherboard is it?

"widely-reported" might be audiophile forums, or otherwise unrelated to audio.

This has nothing to do with Chipset.
It's a EMI RFI issue caused by bad ground (Manibord/conector to PC case. or badly shielded pc case or a combination of this.

What about isolation/shielding of the USB port/circuit on the motherboard?


..."cards" a thing of the past for me. ...............................and my solution is to buy better laptops, not low-budget/cheapos.

 

[Dilettante]

What brand of motherboard is it?

Gigabyte, model name is `Aorus Pro Wifi rev. v1.1`

"widely-reported" might be audiophile forums, or otherwise unrelated to audio.

By "widely" I mean this: https://www.google.com/search?client=firefox-b-e&q=amd+usb+issues

What about isolation/shielding of the USB port/circuit on the motherboard?

Mmm... What am I looking for?

..."cards" a thing of the past for me. ...............................and my solution is to buy better laptops, not low-budget/cheapos.

I have connected the interface to a battery-powered laptop and the issue isn't present there. It's just about the PC.

 

[Katji]

To my understanding the Ryzen USB issues seem to have been largely solved with recent AGESA versions. After all the hoopla I didn't quite trust matters though and went with Team Blue.

hhh, Problem issues with AMD started around 1990 /early 90s. Maybe it was the BIOS that I'm remembering.

 

[Lambda]

We might be referring to different things. AMD issues were widely reported, at some point acknowledges by AMD themselves and somewhat fixed with the recent AGESA versions. Not completely though.

The Noise has nothing todo with the digital part.
The analog part is not made by AMD.

The pointing is thorough, I see no exposed metal.

Maybe the paint is avoiding a proper GND connection?

So... No chance of running ID4 II on 12Mbps? Annoying to say the least, but I appreciate the suggestion.

Don't think it will work but maybe.

I actually do have ASPM L1/L2 and BAR enabled. Might try with both disabled... Stabbing in the dark here.

Its not like some Bits are flipped all the digital stuff seams to work why would you think this will change things?
What ever software / driver or bios issues there might be with your USB they have nothing todo with electrical common mode noise on the USB port.

What about isolation/shielding of the USB port/circuit on the motherboard?

this is why is say to get the USB isolator from topping or any other USB isolator that is known to work with "hi-speed" usb2

 

[Dilettante]

Its not like some Bits are flipped all the digital stuff seams to work

I thought there is some CRC harnessing in USB, no?... So it's either negotiates / works or doesn't and retransmits.

What ever software / driver or bios issues there might be with your USB they have nothing todo with electrical common mode noise on the USB port.

Perhaps you are correct. Thanks for bringing some sanity here.

 

[Katji]

I have connected the interface to a battery-powered laptop and the issue isn't present there. It's just about the PC

Ok, that's another "data point." My widely reported/iow audiophile forums is that most USB noise problems are with desktop computers not laptops.

But I saw that^^ yesterday, wanted to ask whether it still applies when laptop is connected to charger (which is what i do most all the time.) Maybe a silly question.

 

[Lambda]

So it's either negotiates / works or doesn't and retransmits.

And it seams to work for you.... why do you think its an "amd" or chipset problem if USB communication works fine

By "widely" I mean this: https://www.google.com/search?client=firefox-b-e&q=amd+usb+issues

https://trends.google.com/trends/explore?date=today%205-y&q=intel%20usb%20problem,amd%20usb%20problem,intel%20usb%20issues,amd%20usb%20issues

more search results if i look for "Intel usb issues" ... an example of estimation / conformation bias.?

 

[Dilettante]

And it seams to work for you.... why do you think its an "amd" or chipset problem if USB communication works fine

Your comment has changed my opinion Now I am hunting solely for an electrical issue.

 

[Count Arthur]

I wonder if the issue is with the power supply and not the motherboard.

They vary quite a bit, from cheap, nasty and noisy, to very good. This is the one I currently have: https://www.tomshardware.com/reviews/corsair-rm750x-v2-psu,5585.html

 

[somebodyelse]

Hmm.... Given that the situation improves by a lot when CPU is under lot, will this assumption be still valid? If so, I might actually go with the powered card and draw some from SATA cable.

CPU under heavy load may have less variation in current as it's less likely to be switching clock speed than when lightly loaded. But...

Don't think the long cable is low enough impedance work reliably for this kind of issue.
also it might be another ground loop inside the PC to pick up interference.
That's why Low impedance ground paths trough the cases and "IO shield" are important against this kind of EMI

It could work that way too - maybe I should have been explicit with the 'or might not' part implied by 'might'. That's why I suggested an isolator rather than a USB card. I can't help on whether the Audient will work with the cheap, slow isolators or whether you need the faster ones.

 

[Katji]

I thought there is some CRC harnessing in USB, no?... So it's either negotiates / works or doesn't and retransmits.

This is a very good place to start - at XMOS website: https://www.xmos.ai/file/fundamentals-of-usb-audio?version=latest

 

[Dilettante]

I have a bit of an update here: https://www.audiosciencereview.com/forum/index.php?threads/usb-bus-noise.34694/#post-1211917

PSU/PC were probably innocent. Narrowed-down the suspects.