Device side cable shield connection

Post
lbodnar
Super Member
2007/07/09 05:31:33
I used to terminate the USB-B connector housing to the system ground but now I am not sure... We have this:
6.8 USB Grounding
...The user selected grounding scheme for USB devices, and cables must be consistent with accepted industry practices and regulatory agency standards for safety and EMI/ESD/RFI.

Which suggests nothing useful really.  I have also run across the impressive EMC study   http://www.ce-mag.com/archive/1999/julyaug/usb.html which suggests
The USB specification permits the connection of the USB shield to chassis ground at the host end only. We made no attempt to use discrete capacitors for shield termination, a technique sometimes used to connect signal ground to chassis ground at high frequencies but maintain low-frequency isolation.


So who does what these days?

1) leave the shield open on the device
2) connect it to device system ground
3) terminate to system ground with a resistor in case host side shield connection is floating
4) terminate to system ground with a capacitor for RF grounding
post edited by lbodnar - 2007/07/09 05:33:29
DarioG
Scheisse Menschen
RE: Device side cable shield connection 2007/07/09 08:11:41
ORIGINAL: lbodnar
2) connect it to device system ground


This one, rather basicly Smile
Pacer
Super Member
RE: Device side cable shield connection 2007/07/09 08:26:46
Both 3 and 4.
DarioG
Scheisse Menschen
RE: Device side cable shield connection 2007/07/09 08:32:23
Real professionals do show their knowledge, in these circumstances SmileSmile
lbodnar
Super Member
RE: Device side cable shield connection 2007/08/26 16:20:22
Apparently tying cable shield to the device USB Gnd is one of the deadly USB sins: http://uk.farnell.com/images/en/ede/pdf/usb_dev_mistake.pdf (last page)

The proper EMI termination of the shield on the device side is via parallel connection of 1MOhm resistor and 4.7nF 250V capacitor to the system ground (USB ground).
SugarCreek
Senior Member
RE: Device side cable shield connection 2007/08/26 22:21:47
Thanks for sharing some specifics on those R/C values!  However, 1M is higher than I would have expected; I wonder on what they're basing that recommendation...
 
Any other designers want to weigh in on actual values? [looking across the Atlantic... wink]
bob_barr
Super Member
RE: Device side cable shield connection 2007/08/26 22:28:11
Any other designers want to weigh in on actual values? [looking across the Atlantic... wink]

Unfortunately, it's just a bit after 5 AM over in Britain. I'm sure that Pacer will be checking in on this thread sometime today.Smile
DarioG
Scheisse Menschen
RE: Device side cable shield connection 2007/08/27 00:57:12
ORIGINAL: SugarCreek
Thanks for sharing some specifics on those R/C values!  However, 1M is higher than I would have expected; I wonder on what they're basing that recommendation...


Well, 1M is rather common in that use (EMI termination), also on Ethernet side IIRC.
I think that's ok.
lbodnar
Super Member
RE: Device side cable shield connection 2007/08/27 03:01:52
ORIGINAL: SugarCreek
Thanks for sharing some specifics on those R/C values!  However, 1M is higher than I would have expected; I wonder on what they're basing that recommendation...

I believe the resistor's purpose is to dissipate static charge in an unlikely event that either cable shield was damaged or the host fails to terminate it at all.

P.S. Today is a bank holiday in the UK.
xiaofan
Super Member
RE: Device side cable shield connection 2007/08/27 05:13:16
http://www.cygnal.org/ubb/Forum9/HTML/000586.html
 
Apparently there are different recommendations from different vendors. I think it all depends.
 
EMC is really a very difficult issue and often complicated by the space constraints. I was and still am struggling with EMC requirements and often the troubleshooting time is very long -- months or years...
DarioG
Scheisse Menschen
RE: Device side cable shield connection 2007/08/27 07:26:23
ORIGINAL: lbodnar
P.S. Today is a bank holiday in the UK.

 
I.e. you don't have to use BANKSEL ... ?? Smile
(sorry, could not resist!)
SugarCreek
Senior Member
RE: Device side cable shield connection 2007/08/27 09:20:03
ORIGINAL: lbodnar

I believe the resistor's purpose is to dissipate static charge in an unlikely event that either cable shield was damaged or the host fails to terminate it at all.

Ah, 1M makes sense from a static-discharge perspective, indeed.  I was expecting something else from my experience with RS-422/485, where I've seen substantially less (but still non-zero) resistance to separate different grounds.
 
However, if the entire exercise is to avoid coupling local ground noise into the shield of the USB cable (as per the Cypress document), 4700pF coupling sounds high to me.  With that as the stated objective, it seems to me that the right approach would be to make a hard connection between the shield of the USB cable and the ground-plane (and shielding) of the USB device, then isolate that combined shielding ground from the power-supply/IC ground via 1M and a smaller capacitor.
lbodnar
Super Member
RE: Device side cable shield connection 2007/08/27 09:47:16

- Tie them together (Intel)
- Connect them with a resistor (Cypress 1)
- AC coupling through capacitor (SMSC)
- Use both - resistor and capacitor (Cypress 2)

Oh, well... [&:]