[Ghoostknight]

Hey

i know there are mainly

1. Matrix Elements
2. JCat

if we speak about audiophile pcie usb cards
but did somebody try another card that is performing great? or maybe even a simple commercial usb hub pcie card that performs well?

Also if someone compared USB stuff of any kind, im interested to hear it

I have a Ifi Isilencer, Topping HS01, and a DIY Supercap usb filter currently but i definitely heared differences with different usb ports and hubs before, so i wanna try somekind of dedicated PCIE card for usb

Open to any suggestions, thanks!

 

[dunring]

I use this one and it works great, silent noise floor located up against the shielded side of the power supply (away from the video card).
https://www.amazon.com/gp/product/B071P5C6CS/

Also I undervolted my video card for more performance but also less interference and it helped a lot. The best article I've seen was Archimago who did extensive testing.
https://archimago.blogspot.com/2015/05/measurements-corning-usb-3-optical.html

 

[Ghoostknight]

I use this one and it works great, silent noise floor located up against the shielded side of the power supply (away from the video card).
https://www.amazon.com/gp/product/B071P5C6CS/

Thanks for the link

It probably is best to test an standard usb pcie card first again, i did some tests a while back as i was still using windows and they definitely made a difference, but not one where i would say one usb port/card sounds definitely better than the other, but still they sounded different

i just recently switched from a rpi4 setup back to straight usb from pc (because its just more convenient on a PC setup...)

i found this card: https://de.aliexpress.com/item/1005005913316841.html
it does not feature a VLI or ASM usb chip but one by Renesas, which might be a interesting comparison too, since i think i never had a renesas usb hub, the usb hub chip itself probably determines noise to great degree
There was also some talk i read on audiophilestyle about the different chips, and some expensive audiophile pcie cards actually seem to use either AsMedia or Renesas
The VLI805 on the raspberry pi 4 didnt sound that bad either tho

In theory it also isnt that hard to add a external power supply slot and an external clock, but i have to see how much of a difference the pcie card still makes after the ifi isilencer and my diy filter

Also I undervolted my video card for more performance but also less interference and it helped a lot. The best article I've seen was Archimago who did extensive testing.
https://archimago.blogspot.com/2015/05/measurements-corning-usb-3-optical.html

yea this can definitely made a difference depending on your setup, i still remember in the beginning of my career i was able to hear my mouse moving trough the speakers with a straight usb wire, not sure how objectivists can argue that usb is perfect but thats another topic

i also felt like going from RX560 to GTX1070 made an improvement a few years back

Tho i think currently im at a point where the PC doesnt sound "obviously" worse than a raspberry setup, which is kinda great and i thought impossible for a while
One of the last upgrades was to go for a higher quality mainboard + Intel 12400 (and keeping my GTX1070, i dont play much games these days...) and actually putting everything into a LianLi aluminium case

 

[tjl5709]

Another idea for you.

Elfidelity AXF-100 ULTRA III USB isolator.

I use an external linear power supply to drive it. It isolates it from the internals of your PC.

 

[GoldenSound]

Hey

i know there are mainly

1. Matrix Elements
2. JCat

if we speak about audiophile pcie usb cards
but did somebody try another card that is performing great? or maybe even a simple commercial usb hub pcie card that performs well?

Also if someone compared USB stuff of any kind, im interested to hear it

I have a Ifi Isilencer, Topping HS01, and a DIY Supercap usb filter currently but i definitely heared differences with different usb ports and hubs before, so i wanna try somekind of dedicated PCIE card for usb

Open to any suggestions, thanks!

Don't get a PCIe card unless you absolutely have to.

Something such as the Intona 7055-C would be a much better choice

 

[Ghoostknight]

Oh.. i just bought the Pcie card i posted above https://de.aliexpress.com/item/1005005913316841.html , just to doublecheck whether this makes a difference

Don't get a PCIe card unless you absolutely have to.

Something such as the Intona 7055-C would be a much better choice

Hmm.. interesting opinion, do you think filtering makes the pcie card redundant?

 

[GoldenSound]

Oh.. i just bought the Pcie card i posted above https://de.aliexpress.com/item/1005005913316841.html , just to doublecheck whether this makes a difference


Hmm.. interesting opinion, do you think filtering makes the pcie card redundant?

Yes, and the Intona has actually been measured and tested to show insanely low noise whereas most other products have not.

The PCIe cards are definitely going to be cleaner than a mobo USB out, but are still a device internal to a big PC and it's going to be really difficult to keep them as clean as you could with an external solution.

 

[Ghoostknight]

Intona has actually been measured and tested to show insanely low noise whereas most other products have not.

Do you have a link here?

The PCIe cards are definitely going to be cleaner than a mobo USB out, but are still a device internal to a big PC and it's going to be really difficult to keep them as clean as you could with an external solution.

yes i get this, my main curiousity came from wondering whether "reclocking" (or clocking the usb signal "better") usb makes any difference, tho objectively it doesnt make much sense, but its hard to say whether people hear difference because of reclocking or just because the device was low noise (and i guess better reclocking also can reduce overall noise)

 

[GoldenSound]

Do you have a link here?


yes i get this, my main curiousity came from wondering whether "reclocking" (or clocking the usb signal "better") usb makes any difference, tho objectively it doesnt make much sense, but its hard to say whether people hear difference because of reclocking or just because the device was low noise (and i guess better reclocking also can reduce overall noise)

Noise up to 500khz is <9uV. For reference, johnson noise with that bandwidth (the noise literally just from the atoms in the metal being at room temperature) is about 6uV so it's basically as close to the physical limit as you can get.

As to reclocking, USB reclocking has no effect on jitter. There is no clock signal used for audio over USB and everything is buffered asynchronously. SPDIF/AES/I2S are a synchronous protocol and the clocking of the source CAN affect jitter there but not via USB. The only benefit to reclocking (ie: regeneration) is basically just to extend the signal if you have a longer USB run.
The intona does technically do 'reclocking' but that's just cause it's a necessary part of any proper galvanic isolator. Can see the two FPGAs either side of the barrier internally here:

 

[Ghoostknight]

Noise up to 500khz is <9uV. For reference, johnson noise with that bandwidth (the noise literally just from the atoms in the metal being at room temperature) is about 6uV so it's basically as close to the physical limit as you can get.

As to reclocking, USB reclocking has no effect on jitter. There is no clock signal used for audio over USB and everything is buffered asynchronously. SPDIF/AES/I2S are a synchronous protocol and the clocking of the source CAN affect jitter there but not via USB. The only benefit to reclocking (ie: regeneration) is basically just to extend the signal if you have a longer USB run.
The intona does technically do 'reclocking' but that's just cause it's a necessary part of any proper galvanic isolator. Can see the two FPGAs either side of the barrier internally here:

Ah thank you! i didnt know intona uses some kind of "proprietary" isolation via FPGA's and is this low noise

The strange thing for me was, i build a own usb filter ( https://www.diyaudio.com/community/threads/diy-usb-ground-loop-eliminator-and-filter.410706/ ) , it basicly filters every usb line in some form, even ground .... and i actually preferred this filter over using a topping HS01 before it, so im not exactly sure if i either

1. concluded things wrong here
2. or the topping HS01 makes the signal actually "worse" than the original signal in some way

i can clearly hear the difference once i switch the ground passtrough switch on the diy filter tho, so it definitely seems todo its job and a resistor between grounds is basicly the best filter you can get ... i wonder whether "isolation" can still transfer noise if not filtered well

As to reclocking, USB reclocking has no effect on jitter. There is no clock signal used for audio over USB and everything is buffered asynchronously. SPDIF/AES/I2S are a synchronous protocol and the clocking of the source CAN affect jitter there but not via USB

Ah.. makes sense i guess

 

[GoldenSound]

Ah thank you! i didnt know intona uses some kind of "proprietary" isolation via FPGA's and is this low noise

The strange thing for me was, i build a own usb filter ( https://www.diyaudio.com/community/threads/diy-usb-ground-loop-eliminator-and-filter.410706/ ) , it basicly filters every usb line in some form, even ground .... and i actually preferred this filter over using a topping HS01 before it, so im not exactly sure if i either

1. concluded things wrong here
2. or the topping HS01 makes the signal actually "worse" than the original signal in some way

i can clearly hear the difference once i switch the ground passtrough switch on the diy filter tho, so it definitely seems todo its job and a resistor between grounds is basicly the best filter you can get ... i wonder whether "isolation" can still transfer noise if not filtered well


Ah.. makes sense i guess

Topping HS01/HS02 seem quite a lot more basic internally, what the noise performance on that one is IDK unfortunately

 

[Ghoostknight]

Topping HS01/HS02 seem quite a lot more basic internally, what the noise performance on that one is IDK unfortunately

Thanks, first time i see the internals of a topping isolator
since i have only the HS01 i cant say anything specific about the HS02 ... but before i go with a HS02 its probably wiser to just invest into one of the intona's...

only thing i really notice here are: they seem to have planned 2 caps on each side and a clock on the clean side... but both options dont seem populated, the caps look like Rubycon YXJ ... not the best performing caps ESR wise, going with 1 instead of 2 caps also increases ESR, like with many chinese products you see where they cut costs i think

i should also say here i didnt test the power quality with my HS01... i kept feeding the external power supply feeding my diy filter -after- the topping HS01, just using the HS01 as a additional isolation layer mainly for datalines and it still seems to effect things negatively, i might retest this in the future just to be sure

 

[sajunky]

These Matrix and JCAT cards are very good, but a price high, close to a good DDC like DI-24. And even better with DDC in a chain.


As for these cheap isolators, it was mostly a cheat, i.e. HS01 was not isolator. HS02 is, but as seen on the PCB it is very basic, equivalent to a $50 dongle, see below. It is good to go with Intona, but hold on, as new generation of chips from AD ADuM3165 /6, ADUM4165 /6 and TI ISOUSB211 family create opportunity for many to do the same at much lower price. With 100% compatibilty.

In a case of HS02, marking on the chip is removed, but search on Alix for "USB high speed isolator" will bring couple products like:

- ~$50 dongle (smaller version of HS02). I prefer a dongle from HS02, as it will plug directly to a PC (read iDefender manual why)

- ~$60 isolator with built-in hub. I must have one of those, it has a socket for the external PSU

There is also an audiophile oriented product ~$80, refer to a link below. It also include a hub chip and author says that reclocking after isolator is beneficial which I agree.

This is my favourite choice, but there is very expensive one LHY UIP using the same chip.

Read all about few pages back from this post:

https://www.head-fi.org/threads/dou...ble-clean-digital-signal.958462/post-18096245

 

[Ghoostknight]

i.e. HS01 was not isolator.

do you have references here? i never looked inside one but my opinion on these are a bit whacky anyway, the first one i bought broke after a few months because these get incredibly warm, since i got a replacement i still use it but yea... my opinion on topping products shrunk since i started out with a topping d10 which is also gone/defect by now and im not even talking SQ here

but hold on, as new generation of chips from AD ADuM3165 /6, ADUM4165 /6 and TI ISOUSB211 family create opportunity for many to do the same at much lower price. With 100% compatibilty.

ah yea i saw a few of these, you probably mean these https://www.aliexpress.us/item/1005001927948863.html
i might check them out further later on, i feel more comfortable buying these than the next topping hs02

I prefer a dongle from HS02, as it will plug directly to a PC (read iDefender manual why)

hmm i just had a similar thought with my DIY filter...

i just tested these: https://www.wimo.com/de/iso-plus \

someone at a german used market offered "auth emi filters" which im searching since a bit and he also had these lan isolators for cheap, so i thought i also grab one for testing, these have a pretty nice effect imo... both the lan isolator and auth filters used for devices other than your hifi had a pretty noticable effect.. anyway

and i was wondering since the manufactur of this lan isolator recommends to one at each cable end if we could do the same for usb filters/cables etc ... also ferrite clamps placed on both ends of the cable seems to make a difference, so i might create a mini version of my usb filter to plug into the pc directly and use the main one short before the dac, atleast to test whether this makes a difference

 

[Ghoostknight]

did anyone played with these: https://www.amazon.de/AV-Access-Ext...lpcontext&ref_=fplfs&smid=A1QAPHQPQ85ZT6&th=1 ?

im wondering since its much easier to isolate ethernet than usb