Using a computer to detect EMI?

[Sorade]

Hi all,

I was wondering if a computer could be use to monitor the EMI on a house power line ? I have read a few papers on the subject, all of which proposing all sorts of devices, but I was wondering if it could be done directly by prompting a computer to monitor the electric flow through itself ?

I don't know much about the subject but I'd like some tips on where to start looking.

Thanks

 

[anorlunda]

EMI refers to radio waves in the air, not the current flowing through a device. Detecting it needs some kind of radio receiver.

 

[Sorade]

Thanks for the reply.

The paper's abstract I got the information from seems to refer to noise measurable along electrical wiring:

"In this paper, we describe LightWave, a sensing approach
that turns ordinary compact fluorescent light (CFL) bulbs
into sensors of human proximity. Unmodified CFL bulbs
are shown to be sensitive proximity transducers when they
are illuminated. This approach utilizes predictable
variations in electromagnetic noise resulting from the
change in impedance due to the proximity of a human body
to the bulb. The electromagnetic noise can be sensed from
any point along a home’s electrical wiring. This allows
users to perform gestures near any CFL lighting fixture,
even when multiple lamps are operational. Gestures can be
sensed using a single interface device plugged into any
electrical outlet. We experimentally show that we can
reliably detect hover gestures (waving a hand close to a
lamp), touches on lampshades, and touches on the glass part
of the bulb itself. Additionally, we show that touches
anywhere along the body of a metal lamp can be detected.
These basic detectable signals can then be combined to
form complex gesture sequences for a variety of
applications. We also show that CFLs can function as more
general-purpose sensors for distributed human motion
detection and ambient temperature sensing."

I'm a bit confused.

 

[anorlunda]

EMI stands for electromagnetic interference; such as if your computer intefered with cell phone reception.

If it is proximity sensors that interest you, I suggest starting a new thread with a more appropriate title.

 

[tech99]

Thanks for the reply.

The paper's abstract I got the information from seems to refer to noise measurable along electrical wiring:

"In this paper, we describe LightWave, a sensing approach
that turns ordinary compact fluorescent light (CFL) bulbs
into sensors of human proximity. Unmodified CFL bulbs
are shown to be sensitive proximity transducers when they
are illuminated. This approach utilizes predictable
variations in electromagnetic noise resulting from the
change in impedance due to the proximity of a human body
to the bulb. The electromagnetic noise can be sensed from
any point along a home’s electrical wiring. This allows
users to perform gestures near any CFL lighting fixture,
even when multiple lamps are operational. Gestures can be
sensed using a single interface device plugged into any
electrical outlet. We experimentally show that we can
reliably detect hover gestures (waving a hand close to a
lamp), touches on lampshades, and touches on the glass part
of the bulb itself. Additionally, we show that touches
anywhere along the body of a metal lamp can be detected.
These basic detectable signals can then be combined to
form complex gesture sequences for a variety of
applications. We also show that CFLs can function as more
general-purpose sensors for distributed human motion
detection and ambient temperature sensing."

I'm a bit confused.

EMI means electro magnetic interference, and can be either "radiated" or "conducted". For the present application, I would guess that a radio receiver is coupled to the house wiring to monitor the noise characteristics from the CFL. To use a computer, it may be possible to couple to the wiring using a toroidal ferrite transformer, either on one wire or both, or to use a small coupling capacitor in conjunction with protective circuitry. A computer can be used as a radio receiver for low frequencies directly by using the sound card with spectrum analyser software, or by using a dongle to allow operation right up to microwaves. Bear in mind that if you use the power lead to the computer, it is likely to be very noisy due to the EMI from the computer itself. In fact, the domestic environment anywhere is now very polluted with EMI. Bear in mind that CFLs are being replaced by LEDs.
.

 

[meBigGuy]

I agree that there is "radiated" and "conducted" EMI.

There is nothing in a standard PC-like computer that will monitor, measure, or otherwise quantify powerline EMI. You need a conducted-EMI sensor of some sort connected to the computer in some way. You've researched that, I think, so I won't.

 

[Sorade]

Thank you for your comments. I was hoping I could use some of the computer's components to monitor conducted EMI but I think I will have to build the following:

To measure the EMI on the power line, we used the same
hardware used in LightWave [4]. An analog high-pass filter
(HPF) with a 3 dB corner frequency of 5.3 kHz is used to
reject the strong 60 Hz component. The output of the HPF
is sampled at 1 MS/s using a 12-bit analog-to-digital (ADC)
converter in the USRP (Universal Software Radio
Peripheral) followed by transforming the signal into
frequency domain using a 32,768-point fast Fourier
transform (FFT), yielding a frequency resolution (or bin
size) of 30.5 Hz. The signal from the USRP is then fed into
the computer for data analysis. It should be noted that a
USRP was used in this prototype simply for convenience.
Since the required hardware is very simple (a 1st order HPF
and an ADC), it can easily be integrated into a small plug-in
unit, which can be installed anywhere in a home or office.
Chen, Cohn, Gupta, Patel, 2013. uTouch: Sensing Touch Gestures on Unmodified LCDs.

 

[the_emi_guy]

EMI refers to radio waves in the air, not the current flowing through a device. Detecting it needs some kind of radio receiver.

EMI means electro magnetic interference, and can be either "radiated" or "conducted".

I agree that there is "radiated" and "conducted" EMI.

We do test our products to regulatory limits for both radiated and conducted emissions. The FCC's intent in both cases is to limit harmful interference with radio services (EMI). At the low frequencies of conducted emissions testing, the EUT itself is seldom large enough to contain a suitable antenna (low frequencies require long antennas). Instead, the conducted test ensures that the EUT does not launch signals onto power lines which can be harmful radiating antennas at low frequencies due to their length.

The low end, 450KHz, is intended to protect the AM radio intermediate frequency of 455KHz. Historically, radiation at that frequency would bleed through old AM radios since only the antenna tuner had any rejection at that freq.

In other words this is a shameful nit-pick on language: EMI is always radiated, but one of the tests for it involves measuring RF conducted onto power lines.

 

[sophiecentaur]

EMI is always radiated,

So do we ignore the interference that creeps its way into equipment down feed lines? I don't think that vulnerable equipment is able to distinguish between the two types (they are both Electromagnetic). 'Guided' interference may be a bit harder to detect and specify but it can be just as relevant. (It's the last one that people can manage to eliminate when it really matters.)

 

[the_emi_guy]

You are right, we would absolutely care if noise creeps between pieces of equipment via feed lines and causes problems. We would not want to design such equipment or purchase it as a consumer. However, regulatory EMI testing does not address this.

For example, a commercial stereo or TV may have many cables/wire connected up to it: RCA cables, USB cables, headphone cables, speaker wires etc. Conducted emissions testing is required on the AC mains cables and telecom cables (such as DSL), but not on the other cables. In other words, if a poorly designed CD player produces 540KHz noise on its audio-out RCA jack, and this interferes with AM reception within the attached stereo due to conducted coupling across the RCA cable, you would not be in violation of any regulatory requirements. This would be a product quality issue that the FCC would not care about any more than a poor quality RCA plug that broke apart after a few uses and prevented the system from working at all.

The FCC is concerned with managing the electromagnetic spectrum. 540KHz on a 1 meter audio cable will not radiate, but launch it on a 5Km long overhead power line and now they care.

 

[sophiecentaur]

OK. I get it now. An Admin Anomaly. Lots of those in life.

 

[anorlunda]

If you make the definition of EMI too broad it becomes meaningless.

As a power engineer, the worst types of interference I saw were:

1. Electric arc funaces used to melt scrap metal. Interference was in the 60-360 Hertz range. One of them in Mexico caused compensating capacitor banks and transformers on the nearby grid to explode repeatedly and spectacularly.
   
   
2. An open pit mine drag bucket mining machine in Australia. That bucket's load was 40% of the total capacity of the local power grid. It cycled about once every180 seconds, i.e. 0.005 Hertz. It interfered violently with the water levels in the boiler drums in the power plant.http://www.touring-ohio.com/southeast/art/big-muskie.jpg

But in the normal use of the word EMI, neither of these qualifies. They are not RF frequencies.