Computer RFI – ferrite clip-ons

[pnoelw]

I’m about to tackle a clean-up operation on the computer leads – USB, video, etc.

I can see that I have interference from the computer right from hf through to a few hundred MHz. I see that the ferrite materials in the clip-ons are designed to peak at specific frequency ranges.

The computer has an i7-8700 processor which works at up to 4.6GHz. The video processor runs at around 1.7GHz. I’m sure there’s plenty else inside the case which will also add lots of hash.

Where should I be focusing? In the 10MHz region or 100MHz or maybe even 300MHz or higher?

 

[Ubbe]

You should probably focus on the frequencies that bother you the most. But first disconnect the device that are on the other side of a lead. Is it the device or the actual lead that radiate interference? Disconnect the computer screen at the screen end, disconnect whatever are on the other end of the USB cable, the network cable, and so on.

/Ubbe

 

[pnoelw]

Thanks Ubbe

In the case of the screen and by removing the signal lead as you suggested as well as powering on and off, it would appear that both the unit and the lead are causing issues. The Dell 24" monitor is mains powered and I suspect that it's internal converter may be creating noise in addition to the video signal lead.

Of course, one notices (sees) this RFI more on the SDR side of things for which it is necessary to have the screen running in order to make the diagnosis!

It will obvioulsy need to be an iterative process.

 

[RFI-EMI-GUY]

Focus on the frequency bands you want to receive.

Too start with, I would make sure the guts are inside a real metal case to begin with. Like Lian Li models. There are some aluminum cases that are pretty RF tight. It astounds me to see cases with huge plastic windows cut in and flashing doodads.

Then use aluminum foil duct tape to seal unused drive openings.If you have dummy inserts that are metal, use them but use foil tape to bond them to the frame. The foil might not make a great electrical connection but the overlap will create capacitance as well as a waveguide beyond cutoff protection.

I have seen some nice steel CPU cases, but they have heavy paint spray on the mating surfaces so no guarantee they will be electrically bonded on each corner.

Every wire and cable leaving the enclosure is an antenna. Try to minimize them.

 

[pnoelw]

RFI_EMI_GUY. Thanks for the pointers. It's a Dell 8930 tower so the case integrity is quite good but i'll review the joins and work on all the cables.

 

[Ubbe]

Of course, one notices (sees) this RFI more on the SDR side of things for which it is necessary to have the screen running in order to make the diagnosis!

Use your AR8000 to tune to one of the interfering signals you see from the SDR and fine tune using the smallest step size with open squelch and use as a monitoring tool when the SDR screen are gone.

/Ubbe

 

[jim202]

Most of the noise problems I have run up against are due to the LAN cables. Use a receiver to listen to the noise your having problems with. A portable radio would work the best. That way you can use it to track down the source. But I find that when I unplug the LAN cables from the computer, it drops like a rock. So put your effort into putting the ferrite on the LAN cables right at the back of the computer. It may take as many as 3 or 4 of them right next to each other to drop the noise down.

Jim

 

[prcguy]

The most effective way to use ferrite is to run multiple turns of the wire through the core, snapping just one bead over a cable will do nothing at HF. When you add ferrite beads in series the inductance goes up by the number of beads you use, but when wrapping turns around a single bead, the inductance goes up 4X when you double the number of turns.

So for a ferrite bead that has 150 ohms impedance at say 10MHz, stick four of them in series and you now have about 600 ohms at 10Mhz. Take a single ferrite bead and wrap two turns and you also have 600 ohms for just that one bead. You need at least several thousand ohms impedance to make any difference to RFI riding on the cable and about 6 turns of wire through a 31 mix snap on ferrite bead will get you in the ball park at HF. Read that last sentence again, its important!

Here are some effective snap on ferrite beads that will handle six or more turns of Ethernet cable or medium size power cable and this guy has the best retail prices on the planet. This first one has a .722" ID hole and the second is the largest available with a 1" ID hole. https://proaudioeng.com/fair-rite-0-722″id-ferrite-snap-core/
https://proaudioeng.com/fair-rite-1-01″-id-ferrite-snap-core/ Other vendors are charging upwards of $19 for the 1" ID core!

I had a recent problem with some remote powered over LAN security cameras that were destroying my HF reception with 10dB over S9 interference in the 40m amateur band when the cameras were on. Testing proved the noise was coming from the camera end and not the video server. Several of the LAN cables run right under my HF antenna and parallel with it. Due to the large diameter shielded direct burial LAN cable to the cameras I had to use the large 1" ID ferrite snap on bead at each camera with 6 turns of cable and it completely snuffed out the camera RFI. My HF reception is back to normal.

Most of the noise problems I have run up against are due to the LAN cables. Use a receiver to listen to the noise your having problems with. A portable radio would work the best. That way you can use it to track down the source. But I find that when I unplug the LAN cables from the computer, it drops like a rock. So put your effort into putting the ferrite on the LAN cables right at the back of the computer. It may take as many as 3 or 4 of them right next to each other to drop the noise down.

Jim

 

[prcguy]

Edit, had a duplicate post.

 

[pnoelw]

Thanks all. A few good pointers to start working with. Luckily, I don't have any LAN cables. It's going to be an iterative process which will probably take a few weeks to sort.
prcguy - your recommendation in respect of multi turns rather than multi clip-ons is illuminating. Thank you for pointing that out.

 

[Ubbe]

I got 9 cameras and a ip-telephone router, two RaspberryPi devices and of course routers and WiFi extenders and networked sat receivers, videogames and television sets. There are terrible interferencies from most devices when trying to use LAN cables.

The only ones I still use are CAT7 cables from the WAN router to the first LAN WiFi router and to my computer and have everything else on WiFi. Only snag is that neighbours have their WiFi routers channels set to auto find and their routers use different channels from one day to the next and sometimes they interfere and the cameras furthest away from my WiFi routers then drop their frame rate.

/Ubbe

 

[pnoelw]

When dealing with signal cables (video and usb) - is there a risk that using clip-ons and multiple turns that you actually degrade the signal that the cable is intended to carry to the monitor or device?

 

[prcguy]

No. The signals between devices are traveling inside the cable and the ferrite absorbs RF traveling on the outside of the cable. Its like wrapping a water hose around something, the water still comes out the same.

When dealing with signal cables (video and usb) - is there a risk that using clip-ons and multiple turns that you actually degrade the signal that the cable is intended to carry to the monitor or device?

 

[pnoelw]

Are you saying that the RFI travels along the cable screening and can we assume that all computer signal cables are screened?

 

[prcguy]

The problem you would be trying to fix is common mode RF currents flowing on the outside of the screen or flowing on all the cables of an unscreened bundle. Intended computer or audio signals will be flowing in equal and opposite directions between conductors in the cable and will cancel out any radiation from the wires acting like an antenna.

Common mode currents flow with no equal opposite counterpart and will radiate like the RF currents flowing on an antenna element. In fact, any computer or audio cable with common mode currents is an antenna element.

Placing a lossy material like ferrite in the path of common mode currents will absorb the energy and you will get the most absorption when you have the right ferrite mix and amount of inductance determined by the mix and number of wire turns through the ferrite core.

Are you saying that the RFI travels along the cable screening and can we assume that all computer signal cables are screened?

 

[pnoelw]

Thank you for all your suggestions and advice. I'm getting a much better understanding of the problems and the route I need to follow.