FEDEVEL
Platform forum

User , 01-07-2025, 05:56 PM

Hello all, do you have tips to find USB cable with a good shield, better than just a 2cm drain wire soldered to the connector shield? Like a 360° connected shield

QDrives , 01-07-2025, 10:51 PM

Make one yourself?

User , 01-08-2025, 07:46 AM

For commercial use, that is not an option,I need to find one already well shielded but on the paper they are all the same

Robert Feranec , 01-08-2025, 07:58 AM

I will be publishing a video about cable shielding and USB is very interesting. Is your second USB device powered from the USB port or it has an external power adapter? Because it may influence how effective shielding will be

User , 01-08-2025, 09:39 AM

Hi Robert, on the other side I have a USB thumb drive, powered through the cable

User , 01-08-2025, 09:40 AM

Do you have some tips already? 😅

Robert Feranec , 01-08-2025, 11:58 AM

I am not expert for EMC, but if your second device is not connected to ground, shielding may be tricky ... so the question is, why do you need the good shield? problems with EMC?

User , 01-08-2025, 12:45 PM

Yes, the cable shows huge H field at 110mhz while probing with near field probe

QDrives , 01-08-2025, 10:37 PM

1) Do you see ringing on the USB3 signals?
2) Is it terminated properly?
3) What about far field?

Robert Feranec , 01-09-2025, 06:15 AM

the 110MHz is probably coming out from somewhere, I would maybe try to find the source

User , 01-09-2025, 09:33 AM

It is only usb 2.0, we haven't tested far field yet
To be honest, the noise probably come from an isolater placed right after, these things creates a lot of CM noise, placing a Snap on Choke on the cable helps a lot

User , 01-09-2025, 09:34 AM

The source is an isolator placed right before the connector, For now, the PCB can't be redesigned so I'd like to focus on the cable shielding if possible.

Robert Feranec , 01-09-2025, 09:40 AM

I will try to publish the video next week, so it can help you to understand a little bit what is happening

User , 01-09-2025, 12:34 PM

Thanks Robert!

QDrives , 01-09-2025, 02:41 PM

Isolator? Do you have a capacitor between both grounds?

User , 01-09-2025, 05:57 PM

Yes, isolator for user's safety, needed by regulations. I have studied the subject from an EMC perspective and know what to do to ameliorate the layout/schematic, can clearly be done better in the future (not my design actually), but for now we don't have time to redesign the board and we are betting on a good shielding from the USB cable.

QDrives , 01-09-2025, 08:23 PM

I am not asking to redesign the board, I am asking **if** you have a capacitor between the 2 grounds?

User , 01-09-2025, 08:58 PM

Yes I do

QDrives , 01-09-2025, 09:05 PM

"Yes I do"? - Do you mean "Yes, there is"?
And 110MHz or !08MHz? 108MHz is the 9th harmonic of the 12MHz USB2 (1.1) clock.
Does the ferrite core 'filter' enough to pass?

User , 01-09-2025, 09:44 PM

You said: "I am asking if you have a capacitor between the 2 grounds"
I answered: "Yes I do"
Isn't that proper english? I am not a native english speaker, but it sounds right to me, am I wrong?

User , 01-09-2025, 09:47 PM

The near field emissions are rather broadband, it reaches its pic at around 110 MHz

User , 01-09-2025, 09:48 PM

Green: Noise floor

QDrives , 01-09-2025, 09:52 PM

So you got a peak of 38.1dBuV.
Problem is, the near field probe is not calibrated to a far field level.
I would also suggest to do a measurement from 30MHz to 230MHz. In EU that is a frequency range with a limit value. It also gives a lot more detail in that range as it is now just about a single division.
With this range, you cannot see if it is broadband or narrow band peaks.

User , 01-09-2025, 10:29 PM

I don't intend to predict far field with these measurements, It's just to know which part of our system might be the cause of problematic radiation in the far field region during pre-compliance tests.

User , 01-09-2025, 10:31 PM

From 30mhz to 230mhz with a RBW of 100, what more would I see ? Broadband and narrowband are notions dependent on the RBW anyway

QDrives , 01-09-2025, 11:14 PM

"*Broadband and narrowband are notions dependent on the RBW anyway*" -- Yes and no.
With the plot you show now, it is (almost) impossible to tell what it might be.
Just putting shieling on a cable that is (almost) impossible to obtain is not the right approach.

User , 01-10-2025, 07:26 AM

What would your approach be?

QDrives , 01-10-2025, 05:00 PM

I always start with conducted emission. The easiest, the fastest and most important, the lower frequencies.
Lets say you have a 8MHz oscillator (not just a crystal) on your board. Every odd harmonic up to 99 is emitted due to bad design.
In the 30MHz to 2GHz, the first one to pop-up is 40MHz. The other thing is that you see many narrow spikes every 16MHz. However, you also see the 8MHz fundamental one!
If you 'stop' this 8MHz from emitting, you have eliminated also the 13th harmonic too (104MHz).
The next frequency range is 30...230MHz.
I cannot recall ever to have seen a interesting frequency above that then was not already known at the lower frequency.