shielded usb cable
I think this the right place for my issue
Ok, i got a 2.4GHz Receiving RF unit with:
- an RF part
- a microcontroller part
- an UART to USB circuit.
I use an ALUBOS enclosure. See http://www.bopla.de/en/product-catal.../alubos-1.html for more info on it.
The PCB board (2 layers with GND plane) is slided into the enclosure.
I have an SMA connector that is screwed onto one side of the enclosure (cable ground is thus short-circuited to enclosure).
On the other side i have a cable "pass-through" for a USB cable.
The circuit board is powered through the USB. The communication with the microcontroller goes through the USB as well.
Now the problematic:
I got transmitters around the receiving unit. When they're "far away" (10 meters typical), I receive them THROUGH the antenna (as one can expect). But when they're close to the unit (<2 meters typical), I both receive them through the antenna and through the USB cable. I simply tested that by putting a 50 Ohm termination on the SMA.
I'm not that surprised with those results as the USB cable acts as every cables as an antenna. And most probably, the RF signal coupled through USB cable flow back to the receiving circuit of the unit.
My question is how can I remove those RF signal coupled from the USB cables at board level. The USB cable has got 4 signals connected to a USB2UART circuit (Prolific chip):
- 5V
- GND
- D+
- D-
Shall I add ferrite somewhere as well as decoupling caps here and there?
Or do you guys have got another way to interface the USB cable with the enclosure?
My goal is to be sure that a signal received was received THROUGH THE ANTENNA and not coupled by the USB cable.
Let me know if you need more info on the board / enclosure / cable /etc.
Thanks for your help on this puzzling issue
Yep, it is impossible to stop the signal on the usb cable, since its shield (assuming it even has one) is not electrically connected to the housing metal. The cable acts as an excellent antenna, and delivers the RF energy right to the ground plane on your pcb.
There are a couple ways to try to stop it.
You can buy a couple of round ferrite shields and clamp them at various places along the cable. That might reduce the signal level by perhaps 12 db.
You can use a shielded usb cable, and use a metal shelled usb connector that is electrically bonded to the metal housing. Might be good for 25 db shielding.
IF you transmitters are all at around the same frequency, you could try to form a choke joint where the insulated cable goes thru the housing. You would pass the cable through some sort of metal braid that is attached to the housing, and group the braid close to the cable in at least two places along the length.
Thanks for your quick answer. I understand the 2 first options you propose, i.e. 12dB and 25dB (this could be already enough... Have to test). But the "choke joint" solution is less clear to me. Maybe you could point me to some reference weblink or picture or book, etc.?
I had to pass an FCC EMI test once on a fully designed unit. There was no low cost way to change the design to use a shielded cable and a metal-shelled connector on the unit. So I fooled around and got this to work. I got some braided shield wire and connected it to the metal housing. I then passed the Cable through the housing wall, thru the braid, and then along its merry way to the indoor unit 50 feet away. By crimping the braid tight to the cable in two places, I was able to make a "choke Joint" type of arrangement, so that RF energy on the cable was returned to the metal housing and not allowed to travel down the cable. The two places where the tie wrap crimps the braid in are high shunt capacitance points, and the sections in between are high impedance (series inductance) sections. So this forms a lowpass filter. I think the whole braided section was around 6 inches long. It has to be bonded to the metal housing for it to work well.
From my understanding you placed the radio inside of the enclosure and its antenna outside of the enclosure. This kind of test is good to verify the performances of the radio when the antenna is far away from the radio.
But in real life is possible the antenna to be placed relative close to the radio. In this case the radio gets some desense interferers generated from the digital part of the radio, that goes direct to the antenna.
Placing only the radio inside of the enclosure you will never get this situation.
Myself I used a bigger enclosure (900mm x 900mm) very well shielded (120dB)
http://www.ets-lindgren.com/page/?i=5240
and I placed inside, the radio and its receiving antenna.
The transmit antenna is also placed inside of the enclosure (through a coax cable), at some distance from the RX antenna (ideally, outside of the near field).
Why is it that sometimes un-important threads, like this one, get stuck at the top of the forum?
THANKS!
By hazard I used such an approach and it actually worked pretty fine. What I did too was to solder the USB cable shielding braid to the PCB ground inside the enclosure.
Added after 4 minutes:
I've redesigned the board using 4 layers PCB process to be compliant with this document:
http://www.semtech.com/pc/downloadDocument.do?id=2550
I'll let you know about the results.