USB2.0 questions

[cupoftea]

Hi,
We have a USB Type 2.0 connection coming on to a board, where it connects to a PIC micro. The connector is TE Connectivity 2173157.
Do you believe that all USB2.0 cables on the market, that connect with the above connector, will be shielded cables?

2173157 USB 2.0 connector

https://www.mouser.co.uk/datasheet/2/418/7/ENG_CD_2173157_A1_2173157_C_revA1-2028883.pdf

Also, would you agree that all USB2.0 connections coming on to a board from a cable, will need a common mode choke to be fitted at the cable entrance point, on the PCB? (in the D+ AND D- lines)

...___...___
Also, the USB2.0 has five connections, D+, D-, VBUS,ID and 0V.
We use a common mode choke for D+ AND D-.
And for VBUS, ID and 0V, would you agree that these need a ferrite bead to be placed in their tracks, right by the connector?

--- Updated Jul 21, 2021 ---

....___...___..---...---
This connector......

https://www.mouser.co.uk/datasheet/2/418/7/ENG_CD_2173157_A1_2173157_C_revA1-2028883.pdf

...is for USB 2.0. It has a metal shell, which is supposed to be connected to circuit 0V. But do you think its best to connect it to circuit 0V via a ferrite bead? (ie, for the purposes of radiated emissions suppression , for passing EN55032 EMC testing)
...---...___...---...___

 

[betwixt]

All USB cables SHOULD be shielded but how well it is achieved depends on the manufacturer, I must add that I don't think I've ever seen un-shielded one although it may be in the form of metallised plastic. The shield should ideally be bonded to a low impedance ground. I'm not sure adding any impedance in line with it is a good idea. A USB cable fitted with a ferrite ring around it would seem a better option for stopping EMC.

Brian

 

[cupoftea]

Thanks,
I did wonder whether these USB2.0 common mode chokes were just cashing in on the radiated EMC fears of those such as myself?....

https://www.farnell.com/datasheets/2724119.pdf

 

[betwixt]

I wouldn't say they are cashing in because there are legitimate needs for such devices but equally, I wouldn't suggest they should be fitted 'just in case'. Do consider that common-mode or not, the USB signal frequencies extend into the 10s of MHz so they will inevitably degrade the waveform integrity so some degree. In most instances that wouldn't be noticeable but with long cables or poor quality ones it could cause data failures. They are probably very effective at stopping external RF reaching the USB interface inside a product but not so useful at preventing radiation from USB signal leaking out.

Brian.

 

[emresel]

If you have chance, I suggest to use common mode filters.
In my experience in past, i faced emi issue @960Mhz which is 2nd harmonic of USB signal (480Mhz) and those filters helped to improve, since then i get to used place it in designs.
Keep in mind, USB2.0 is bidirectional either host or device can apply signal on bus/cable. Both side can be root cause of any emi issue.

 

[cupoftea]

In my experience in past, i faced emi issue @960Mhz which is 2nd harmonic of USB signal (480Mhz) and those filters helped to improve, since then i get to used place it in designs.

Thanks, so i guess it must have been the "stray" diff mode inductance of the common mode choke, and the stray line capacitance that did the filtering for you?

 

[tmraugmap]

If you have chance, I suggest to use common mode filters.
In my experience in past, i faced emi issue @960Mhz which is 2nd harmonic of USB signal (480Mhz) and those filters helped to improve, since then i get to used place it in designs.
Keep in mind, USB2.0 is bidirectional either host or device can apply signal on bus/cable. Both side can be root cause of any emi issue.

2nd harmonic is 480MHz, USB 2.0 is NRZ and strongest harmonic content will be at 3rd harmonic @ 720MHz

USB chokes will have significant effect on signal integrity in 480mbit (Hi-speed) mode, if you are dealing with a weak (slow edge rate) transceiver you may encounter significant signal integrity issues with a choke. The slow edge transceivers are much more EMI friendly and concerns can often be mitigated by following good layout practice and using a ferrite as mentioned above. Choke will be more suitable for fast edge transceivers.

Be aware that many modern USB transcievers from SMSC(now Microchip), transceivers in the iMX processors, etc have tunable USB transceivers which can help in mitigating both signal integrity and EMI concerns.

Low cost/china cables are notorious for very poor shielding, any good USB cable will have braided wire shielding + a drain wire running with the shielding in addition to the USB ground wire. Differential impedance of USB cables varies widely among low cost manufacturers, in the past I characterized many cables on a VNA and settled on the "cables to go" brand, but this was years ago and I do not know if they have changed manufacturing practices since then. Found that some manufacturers were 20% or more outside of specified 90-ohm differential impedance which will have a significant effect on signal integrity.

Keep in mind when dealing with USB 2.0 that the common mode range of your host PHY is of key concern and if your USB peripheral is sinking a notable amount of current, and your USB cable has poor GND impedance, this current can push the peripheral/host USB signaling out of each other's common mode range. One way to mitigate this is to tie the USB GND wire and shield together, but depending on the type of downstream device (peripheral, or another part of the system) this can introduce another world of possible complications.

Good luck