Jump to content

Measure RFI


spokenward

Recommended Posts

  • Members

I just resolved an apparent RFI (I'm talking radio frequency interference here) disturbance that was affecting the performance of a little SOHO style Gigabit switch. Is there anything like an app or application to measure RFI?

 

More audio plumbing seems to be headed to RJ45 so this might be a growth industry.

Link to comment
Share on other sites

  • CMS Author

Tune an AM radio in between stations. What you hear is RFI. (don't use an AM radio app for your iPhone ;) ) That used to work real well with computers back when their clocks were a few megahertz but isn't so effective with modern computers and FCC and CE standards.

 

A while back, some owners of the PreSonus StudioLive mixer were finding that they got interference with wireless mic receivers when the receiver was placed too close to the mixer. They were telling each other to test their mixers with an FM radio tuned to a vacant spot near the low end of the band. (PreSonus had a fix for this that required a factory modification).

 

The real tool for the job is a spectrum analyzer which is essentially a calibrated wide band radio receiver. This has traditionally been a lab instrument but there are some less expensive (a couple of grand) instruments that connect to a PC and show you what's out there which are getting some use where there are a lot of wireless mics in use and you need to find out about problems (including interference from other wireless mics in the area) before the show starts. You'll find some useful information on Spectran analyzer kits here. I've seen these at the NAB show and I think I recall seeing one (or one from a similar manufacturer) at the NAMM show last year or the year before.

Link to comment
Share on other sites

  • Members

No theres no programs you can plug in and just test. RF is an analog wave and a switch isnt going to

have the intelegence to sence that no less measure it for you.

 

Interference like that is measured with normal electronics test tools. You would use an RF generator and measuer how much the

circuits resist the signal being generated. You couldnt do it any other way because some areas where the equipment may be

used may be closer to a radio wave source than another. Someone may live next door to a Ham operator or CB radio transmitter

and be overwhelmed EMF. Other areas may be free from it so the only way to test it is under lab conditions and using a variable

RF generator and induction coil and wave the coil in close proxcimity while measuring any signal being picked up by the circuit

in various stages with a scope and analizers.

 

I doubt this is a problem with most switches or routers. Most failure with those is just a dirty signal coming in

or signal echo created on fiber optic lines, repeaters and signal amplifiers, low band wiidth and any number of other issues coming

from the data line. Most of those switching devices devices are UL approved. Those kinds of diagnoses fall into

the catagory of not knowing where the real problem lies and blaming it on something else. It couldnt possibly be the

IT provider has crappy servers or has too many customers for the hardware so the bandwidth sucks.

For all they know it can be sun spots as the cause communication failures.

 

Devices are rarely intermititant too so they dont usually fail one day and fix themselves another.

They do vary in quality though as far as being able to filter out noise.

Chances are its a software conflict or internet provider problem and not the switch, unless you have changed out the switch and

found it was truely the faulty. Then you would have to proove it isnt a weak component or connection on the switch,

and RF difinately is the cause of the problem.

 

Thats a big hassle not worth persuing. Switches are cheap and just changing one out is the quickest and cheapest solution.

 

Digital components arent as susceptable to an analog waveform like RF as they are to ac noise on the line or a degraded signal quality

which has spikes that can trigger a false impulse on a chip or just confuse it and cause a gate to open or close.

Even then the software protocol continually scans the conditions of the gates and prevents false events like that to occur.

The best you can do to prevent damage to a circuit from AC noise is to put ferrite chokes on all the

wires coming from an to a unit and use a good quality Noise supressor on your AC line to block RF.

 

Other then that, the Cable or DSL provider has tools to test your signal strength and quality. Problem is with that

it need to be run over a few days of use with a historogram running to detect when the peak usage occurs and how

low the bandwidth gets. Chances are they already know what areas of their grid needs updating by the number of

complaints they get in an area. My work computer DSL line has major slow downs and cutoffs due to peak usage during the day.

As a matter of fact, Today we are switching over to cable because the situation has gotten so bad its hurting our productivity.

Link to comment
Share on other sites

  • Members

thanks for the thoughts!

 

Now I will give the backstory.

 

I agree with WRGKMC that RFI was not the most likely source of the problem I had with the switch. In this case though I believe that it may have been the case. There was a "housekeeping" sort of problem with a tangle under a rack. It may have been intermittent because it was aggravated by different computers being on or off. It was presenting with an XP computer complaining that "a network cable is unplugged." A Vista laptop would not find it's network connection. It seems to be stable now.

 

Mike's AM radio example is pretty close to my thoughts. I thought about antennas and realized that we live in an antenna-rich environment. My problem seemed to become irresolvable after I swapped in a new wireless router when my old router abruptly failed. So like Smokey Robinson wrote, I found myself thinking "The Hunter Gets Captured by the Game". We are walking around with antennas on a lot of portable devices. I was hoping that someone had reverse engineered the classic antenna and analyzer solution.

 

I thought that I was onto something when I found this:

 

 

You can measure RFI in your home or community, and contribute to increased global awareness of radio astronomy! It

Link to comment
Share on other sites

  • Members

Yes switches will take a jolt before a computer will.

You would think a modem being before a switch will go first, but their circuits usually have more

filtering because they are connected directly to the line and will see more spikes. The chips in

newer switches likely cost a few cents and you cant expect quality and low cost being the norm with all brands.

 

My company uses Cisco equipment which seems to be pretty good. I've gotten 5 years out of this last switch

since the last upgrade. heat sources do attract dust so just blowing the things out with compressed air can make all the difference in the world.

Wire connectors are the other biggie. The wires are just crimped into connectors that pierce the wire to make a connection. its not

like they are soldered so temp causes the contacts to flex and become intermittant. You may not even recognise its a wire problem

because you dig around and move the cables and it starts working again. A good IT tech can ping the circuit and see if which device in a chain is failing.

If its a switch its obvious.

 

Some switches can also be accessed remotely and have the switches changed remotely. And sometimes an IT provider will do upgrades and changes the

devices priority level andf it doesnt get recognized the same as it did when it was first installed. Youi can usually deinstall the old drivers, remove registry entries then reinstall

the drivers and the thing takes off and runs like new again. i had this happen on a wireless router recently.

 

I had upgraded from DSL to cable recently and the rireless quit function well. Some times it worked, sometimes it didnt.

I went and reinstalled the thing with the disk, and changed my password and security (very important because most hackers know the default password is admin and will hack your network) and the thing runs like a champ now.

 

Theres a ton of other possibilities for your problem from what I've read, On a list of 100 possible causes, RF would be like 99 on the list.

The strength of even a strong RF signal is really weak in comparison to the 3~ 5V DC signal used to switch a gate.

You do have to rule out what the problem is not and localize it down to your possibilities like you have done by switching things out. If it holds up like it is

I'd chalk it up to a few more likely possibilities like dust and dirt in the unit, connectors, heat which can change component tolerances, loosen connectors,

plug in chips have oxidation on the chip legs, weak components beginning to fail, power supply fatigue, stuff like that.

 

Jumping straight to the bottom of the lsit is something a tech may do to a customer when he tries to explain an unexplainable failure after checking

all other normal causes. It not like thay can say "Our company makes Cheap Cheezy Crap like our compeditors do in the orient with workers being paid

$1 a day and with low grade parts so we can suck more profit out of your wallet" but he would be telling you something closer to the truth if he did instead of

blaming it on crop circles, UFO's Sun Spots or The loud Radio DJ on the radio this morning. Its kind of like blaming it on Bush when it gets to that level.

i mean its possible, but so unlikely.

Link to comment
Share on other sites

  • Members

I've worked in this area in the past, testing aerospace & avionics gear for RFI susceptibility.

 

There are a few categories of RFI - - Conducted, Radiated and Induction. Before analysing an RFI problem you need to determine which one(s) are involved.

 

The ways to mitigate the problem will also depend on which one(s) are contributing - - For example, if the RFI is being carried into the box on a wire, it won't help to put the box into a shielded enclosure.

 

If you want to measure RF field strength, Google for 'Software Radios' - - there are some direct-to-baseband converter kits out there that wire into a PC audio card, with software available to let you scan the RF spectrum up into a few GHz, using different Nyquist modes on the A/D. You'd also need some sort of calibrated broadband directional antenna, which is not a trivial task... Each RF band normally requires a separate antenna; you can't normally use the same one for VHF as you do for C or L band microwaves.

 

A simple test would be to unplug all the lines into and out of the unit (except power) and turn it on in it's normal environment. If it fails, it's likely failing from radiated RFI or possibly overheating. If you then build a screen box (no gaps larger than 1/4") and put the unit in it and it stops failing, you almost definitely have a radiated RFI failure issue.

 

If it doesn't, it's probable that the issue is induced or conducted RFI. A hard (persistent) failure on an ethernet switch or router due to voltage transients ('spikes') is fairly uncommon, since the ethernet spec calls for tolerating transients up to 2000 peak volts, and pretty much every ethernet PHY (physical layer interface) on the market is tested for compliance before it hits the market. It is possible that radiated RF being picked up by wiring is causing the problem, but that is also pretty uncommon, unless you are using the wrong class of cabling (like using CAT5 in an environment that requires CAT6).

 

But try the simplest thing first: Test for overheating by positioning a muffin fan or two so cool air is directed over (or into) the unit. If the problem goes away, it's more likely a heat issue than an RFI issue...

Link to comment
Share on other sites

  • Members

thanks, Philbo! That's a lot of great information. I will explore the overheating theory next. Thanks to all.

 

In the interest of science I nursed that switch along for another day. I became convinced that it had more than a single problem. It would drop to lower speed and resume correct behavior after it was power cycled and offline for awhile. So the heat theory is probably a good one.

 

I have replaced that Netgear with a D-Link and now my little network is end-to-end D-Link. These are the "greener" (cooler, lower power) switches so I can feel better about that anyway.

 

I will add that the intermittent "Network cable is unplugged" error has resolved with this switch replacement. In all of my googling I don't think I ran into that as a solution. Thanks everybody!

 

Link to comment
Share on other sites

Archived

This topic is now archived and is closed to further replies.

×
×
  • Create New...