28 Dec
2016
28 Dec
'16
11:11 p.m.
Hi,
if I connect an active USB hub and an USB audio card by a 1.5 m USB
cable, there are no xruns. If I connect the devices using a 0.5 m cable
with a ferrite core [1] there's one audible xrun after the other.
Technically a ferrite core affects the quality of a digital
signal, but I wonder if it could be that much, that it could cause
xruns. Maybe just the cable or the jacks are bad. Does anybody know
more about USB cables with ferrite core and possible impact on the USB
signal or other reasons that make a differences for USB cables or USB
jacks?
[1]
https://www.reichelt.de/USB-Kabel/DELOCK-82057/3/index.html?ACTION=3&GR…
Btw. the iPad now could connect to the PreSonus and shows all I/Os, but
neither MIDI nor audio I/Os work, IOW I noticed the above when I tested
the PreSonus in combination with the USB hub using Linux.
I'll return the PreSonus and try to find another card for usage with
Linux as well as with iOS.
Regards,
Ralf
29 Dec
29 Dec
2:56 a.m.
1) Perhaps it is not so much the issue of whether the cable has a ferrite RFI suppressor
but the quality of the cable that is important. I recommend that you use a quality USB 2.0
spec'd cable. Better yet, use a USB 3.0 cable.
2) Unfortunately, USB is inherently noisy. Using a ferrite to reduce common-mode currents
in the cable is, in principle, a good idea. However, there are potential issues.
Either or both the sound card and your computer's USB port may not properly isolate
the signal and power grounds, and there may then be noise issues if the two grounds are
not electrically the same. When you install a common mode choke (the ferrite), you are
effectively isolating the two grounds. This should be fine, but if either sound card or
USB port are not properly engineered, there could be a problem. You could try running a
wire (16 gauge or larger) from the chassis of the sound card to the chassis of the
computer. Do not put a ferrite around this ground wire.
3) Another potential issue is the sound card power supply. Unfortunately, almost all power
packs sold these days in the U.S. are Chinese switching power supplies which produce
TREMENDOUS RFI. You could try a linear power supply instead, although they are getting
harder and harder to find.
Jim
________________________________________
From: Linux-audio-user [linux-audio-user-bounces(a)lists.linuxaudio.org] on behalf of Ralf
Mardorf [ralf.mardorf(a)alice-dsl.net]
Sent: Wednesday, December 28, 2016 3:11 PM
To: LAU
Subject: [LAU] USB cable with ferrite core
Hi,
if I connect an active USB hub and an USB audio card by a 1.5 m USB
cable, there are no xruns. If I connect the devices using a 0.5 m cable
with a ferrite core [1] there's one audible xrun after the other.
Technically a ferrite core affects the quality of a digital
signal, but I wonder if it could be that much, that it could cause
xruns. Maybe just the cable or the jacks are bad. Does anybody know
more about USB cables with ferrite core and possible impact on the USB
signal or other reasons that make a differences for USB cables or USB
jacks?
[1]
https://www.reichelt.de/USB-Kabel/DELOCK-82057/3/index.html?ACTION=3&GR…
Btw. the iPad now could connect to the PreSonus and shows all I/Os, but
neither MIDI nor audio I/Os work, IOW I noticed the above when I tested
the PreSonus in combination with the USB hub using Linux.
I'll return the PreSonus and try to find another card for usage with
Linux as well as with iOS.
Regards,
Ralf
_______________________________________________
Linux-audio-user mailing list
Linux-audio-user(a)lists.linuxaudio.org
http://lists.linuxaudio.org/listinfo/linux-audio-user
12:33 p.m.
On Thu, 29 Dec 2016 01:56:37 +0000, Schatzman, James wrote:
1) Perhaps it is not so much the issue of whether the
cable has a
ferrite RFI suppressor but the quality of the cable that is important.
Saw + pincers = ferrite core removed = no xruns anymore
I guess it isn't fortuitousness and the culprit indeed was the ferrite
core.
I recommend that you use a quality USB 2.0 spec'd
cable. Better yet,
use a USB 3.0 cable.
You are aware that the audio interface has got a type B socket?
http://blog.macsales.com/wp-content/uploads/2012/07/USB-BTYPES.jpg
2) Unfortunately, USB is inherently noisy. Using a
ferrite to reduce
common-mode currents in the cable is, in principle, a good idea.
However, there are potential issues.
Either or both the sound card and your computer's USB port may not
properly isolate the signal and power grounds, and there may then be
noise issues if the two grounds are not electrically the same. When
you install a common mode choke (the ferrite), you are effectively
isolating the two grounds. This should be fine, but if either sound
card or USB port are not properly engineered, there could be a
problem. You could try running a wire (16 gauge or larger) from the
chassis of the sound card to the chassis of the computer. Do not put
a ferrite around this ground wire.
There were xruns, not just analog noise issues.
3) Another potential issue is the sound card power
supply.
Unfortunately, almost all power packs sold these days in the U.S. are
Chinese switching power supplies which produce TREMENDOUS RFI. You
could try a linear power supply instead, although they are getting
harder and harder to find.
Sure that's why I just in case purchased a cable with a ferrite core.
Actually there are no issues without ferrite core.
Regards,
Ralf
OT: Voodoo, the PreSonus inputs and MIDI in and out work with my iPad,
but all output channels are silent. Here is a list of devices, known to
work with iOS devices, the PreSonus is one of them, but actually it
doesn't work here:
http://www.wavemachinelabs.com/Support/auria-audio-interfaces
30 Dec
30 Dec
7:12 a.m.
Ralf Mardorf wrote:
Saw + pincers = ferrite core removed = no xruns anymore
Thank you for doing the experiment.
>2) Unfortunately, USB is inherently noisy. Using a
ferrite to reduce
>common-mode currents in the cable is, in principle, a good idea.
Using a common-mode choke works only for differential signals.
But USB also uses common-mode signalling for certain states.
Ferrites on a USB cable should be one per data line, and sized so that
only frequencies higher than the USB signalling rate are filtered out.
I would guess that that cable was (originally) intended for low speed only.
Regards,
Clemens
12:01 p.m.
On Fri, Dec 30, 2016 at 07:12:34AM +0100, Clemens Ladisch wrote:
Using a common-mode choke works only for differential
signals.
True, but any signal that has both the forward and return path
in the cable is differential (even if not truly balanced). That
remains true even if the return path is the screen instead of
signal ground.
But USB also uses common-mode signalling for certain
states.
If you mean signalling that depends on a ground path outside
the cable I doubt very much this is true.
Ciao,
--
FA
A world of exhaustive, reliable metadata would be an utopia.
It's also a pipe-dream, founded on self-delusion, nerd hubris
and hysterically inflated market opportunities. (Cory Doctorow)
1:11 p.m.
A few minutes ago I ordered new headphone cables, jacks etc. so I've
taken the opportunity to speak to a technician from Reichelt, since I
ordered the USB cable there, too. He doubts that the ferrite core dose
cause issues. His guess is a "Monatgsproduktion" (bad produced cable),
but actually removing the ferrite core solved the issues with the
PreSonus. Btw. I'll return the Presonus 1818 VSL. I already ordered a
Focusrite Scarlett 18i20 2nd Gen, less expensive at the moment
https://www.thomann.de/de/focusrite_scarlett_18i20_2nd_gen.htm . Maybe
it will work with Linux, as well as with an iPad.
31 Dec
31 Dec
7:10 a.m.
Fons Adriaensen wrote:
On Fri, Dec 30, 2016 at 07:12:34AM +0100, Clemens
Ladisch wrote:
USB does not use currents but voltage levels for signalling.
The two data lines (D+ and D-) have controlled impedances and drive
levels so that the effective currents are symmetrical, but when both
lines are pulled low to signal EOP, there is not necessarily an equivalent
return current over the ground line.
(That depends on whether the device has an independent power supply.)
Regards,
Clemens
Using a common-mode choke works only for
differential signals.
True, but any signal that has both the forward and return path
in the cable is differential (even if not truly balanced). That
remains true even if the return path is the screen instead of
signal ground.
But USB also uses common-mode signalling for
certain states.
If you mean signalling that depends on a ground path outside
the cable I doubt very much this is true.
12:43 p.m.
On Sat, Dec 31, 2016 at 07:10:16AM +0100, Clemens Ladisch wrote:
USB does not use currents but voltage levels for signalling.
The two data lines (D+ and D-) have controlled impedances and drive
levels so that the effective currents are symmetrical, but when both
lines are pulled low to signal EOP, there is not necessarily an equivalent
return current over the ground line.
There must be. The net sum of the currents flowing through he cable
must be zero, unless some current takes a path (e.g. via protective
ground) not going via the cable.
If two devices are connected only by the USB cable there is no
such path. And if there is one (e.g. protective ground), and a
signal current takes that path rather than via the cable, then
there's something seriously wrong with the design of at least
one of the two devices.
Ciao,
--
FA
A world of exhaustive, reliable metadata would be an utopia.
It's also a pipe-dream, founded on self-delusion, nerd hubris
and hysterically inflated market opportunities. (Cory Doctorow)
But USB also uses common-mode signalling for certain
states.
If you mean signalling that depends on a ground path outside
the cable I doubt very much this is true.
1:40 p.m.
Fons Adriaensen wrote:
On Sat, Dec 31, 2016 at 07:10:16AM +0100, Clemens
Ladisch wrote:
Indeed. But if the device is _not_ bus powered, there is such a path.
Regards,
Clemens
The two data lines (D+ and D-) have controlled
impedances and drive
levels so that the effective currents are symmetrical, but when both
lines are pulled low to signal EOP, there is not necessarily an equivalent
return current over the ground line.
There must be. The net sum of the currents flowing through he cable
must be zero, unless some current takes a path (e.g. via protective
ground) not going via the cable.
If two devices are connected only by the USB cable there is no
such path.
2680
days inactive
2683
days old
linux-audio-user@lists.linuxaudio.org
8 comments
4 participants
participants (4)
-
Clemens Ladisch -
Fons Adriaensen -
Ralf Mardorf -
Schatzman, James