Microwave oven interfering with WiFi on the 2.4GHz band
As I understand it, microwave oven magnetrons operate at 2.45GHz, which is an unlicensed band in most of the world. When 2.4GHz technologies such as WiFi (802.11), Bluetooth and Zigbee were launched, there were concerns that these would collide with microwave oven frequences. I remember a very early (year 2000-ish) Bluetooth evaluation project where I amused myself by placing two Bluetooth modules on each side of a running microwave oven and couldn't get any form of Bluetooth traffic working.
But are microwave ovens really allowed to cause interference at longer distances, such as jamming out WiFi located in the same room or building? My take is that they should count as non-specific SRD (short range device). I'm located in Europe, so in that case the 2400-2450MHz unlicenced, harmonised band should apply - this is standardized by the European Radiocommunications Committee within the whole of EU. If so the SRD standard EN/ETSI 300 440 for non-specific SRD (sorting under the new RED directive) should apply. Is this correct so far? Or do microwave ovens have their own specific technical standard?
ETSI standards are conveniently available for free, so I checked v.2.1.1 of EN/ETSI 300 440. Chapter 4.2.4.4 regarding limits of unwanted emissions in the spurious domain are limited to 1uW for frequencies > 1 GHz. I'm assuming this is 1uW E.R.P, so roughly -30dBm E.R.P. The spurious domain is defined in complex ways by this standard depending on bandwidth from carrier. I don't think a device that disturbs with -30dBm E.R.P. would affect WiFi or any other radio technology much at all. Or would it? Would the "listen before talk" RSSI limit of WiFi cause any problems here?
Could any radiated emission from a microwave be regarded as an intentional radiator? Does it have a designated carrier and occupied band, where it may act just as any other SRD? I'm assuming that the microwave oven occupied band is very wide, several 100MHz.
If so, then what happens when the microwave occupied band (+/- x MHz from carrier) collides with the WiFi occupied band? There are a couple of (European) WiFi channels that overlap with 2.45GHz specifically. Since WiFi (and Bluetooth) use spread spectrum technologies, would that mean that WiFi skips certain channels, resulting in reduced bandwidth?
2 answers
You are accessing this answer with a direct link, so it's being shown above all other answers regardless of its score. You can return to the normal view.
Microwave Materials, Spectrum then EMI effects on SNR of WiFi, BT.
There are a lot of measurable parameters in microwave dielectrics which affect;
- Dielectric Constant , Dk between conductors and thus Impedance , Zo
- scattering parameters in a transmission path from mismatch and losses
- skin depth: compare wavelength of IR vs uWave
- Longer λ is better for skin effects but worse for loss tangent
- Microwave Ovens:
- Consumer: f= 2.45GHz +/-100~120Hz sidebands λ=12.2cm, P=1.8kW max
- Commercial: 915 MHz +/-100~120Hz sidebands λ=32.8 cm, P=3kW max+3kW IR(opt.) The modulation sidebands depends on power line frequency (doubled with many harmonics) Then there are SIGNIFICANT 1/4 λ sidebands typ 300MHz due to the geometry of standing waves in the chamber but again modulated with power line harmonics and not high bit rates. The spectrum will show quasi-peaks and not modulation and the modulation is dispersed at low f by the rotating turntable and contents reflecting off the oven walls.
- compare this with WiFi: including but not limited to 2.4 GHz, 5 GHz, 6 GHz, and 60 GHz frequency bands with frequency hopping within a BW of 10 MHz, 20, 40 , 80MHz etc which is rather immune to 2.45GHz in a narrow band unless the front end is compressing with WiFi signal <= -78 dB and Rician Fading loss of 10dB
Test the SNR on a Laptop with a strong signal -70dBm and a weak signal-80dBm which is borderline on 54Mbps but OK on 11Mbps. Then put near a microwave oven and look for LOS (loss of signal). There are many windows tools to convert RSSI (received signal Indicator to dBm) for any WiFi signal on a Laptop or Mobile.
Contamination is more critical for interference, e.g. acid rain, snow on Sat dish, body reflecting/blocking antenna.
- Loss Tangent is the critical factor in a dielectric that absorbs power (heat) and the high dielectric constant (Dk=80) lowers the impedance by a factor of 80 but does not mean it is more lossy but can conduct more current in series with the cycling waves.
Oil, plastic and fat have a Dk around 4. The presence of contaminants and additives in FR4 makes it a poor dielectric for microwave with the exception of special blends of epoxy/fiberglass such as GETEK and others use more polyamide which affects cost a bit but reduces the loss tangent enough for use in the ISM (928MHz) and 2.4GHz bands at a 10% cost increase of the PCB.
Experiment DIY
This means de-ionized, sodium-free pure water does not heat up quickly in a microwave oven.
The presence of salt in any dielectric is an ion that makes it significantly more conductive. Test a glass of clean water and one with a tbsp of salt after 60 secs for temp rise. BIG difference in temp rise ('C only not 'F no exceptions for Engineers even in the Excited States).
That is Loss Tangent loss due to ionic impurities.
Disclaimer
I have measured RSSI accurately to 0.1dB for loss of signal and BER on WiFI extensively in the past, but not now. When 54Mbps fails, 11Mbps works and near threshold 1dB rise in SNR is huge ~2 orders of magnitude difference in BER vs 1dB rise SNR. The Laptop antenna motion of <1mm in rotational angle aiming for a reflection, can easily change 1 to 10 dB from multipath when the range is far.
I do have an E/H Field Strength meter but it is not frequency selective up to 5GHz so I could not test this and my 10yr old laptop power connector is dead (solder joint) so I could not make measurements, but my own experience is Rician Fading is alot more serious (>-10dB) on weak signals from body / wall reflections out of phase than the effect of a microwave oven affecting the Wifi/BT radio if well designed.
The Physics of water like any dielectric don't have a microwave resonant frequency The absorptive effects have been characterized from 10cm to 10um rise with lower wavelength but this is enhanced by salt ions.
@Olin's assertion to resonant frequency is totally bogus. Water has absorptive properties over many decades of f without this resonance.
Conclusion
http://www1.lsbu.ac.uk/water/microwave_water.html#loss This is a good reference but fails to discuss the ionic behaviors of contaminants but clearly agrees with my assertion that there is no interference.
Post Mortem
I take everything back. I pulled out an old Panasonic 1800 W microwave oven and it interfered with our Wifi. Unlike the Samsung 1kW.
Microwave ovens work on the frequency they do because that's one of the main resonant frequencies of water molecules. The radiation excites water molecules, which then transfer some of their extra energy to other surrounding molecules.
No, microwave ovens are not intentional radiators, at least not in the legal sense here in the US. Yes, they use radiation internally, but there is not purpose or intent of that radiation getting outside the unit. So unless there is a specific exemption for microwave ovens in the law (haven't checked), they have to comply with the same maximum radiation limits any other device does. I don't have time right now, but maybe later I'll look the applicable FCC part 15 rules and see what that radiation limit is and exactly how it is defined.
Added
From a very cursory look, it seems like a home microwave oven in the United States is not allowed to produce external radiation exceeding 500 µV/m at a distance of 3 m at the microwave frequency.
1 comment thread