Microwave Oven Bandwidth / Co-existence with Wi-Fi Networks

[Halcyon]

I'm in the process of designing a wireless network which unfortunately requires use of the 2.4 GHz band (as congested as it is). I know microwave ovens have a centre frequency of 2450 MHz which puts it within the range in between channels 8 and 9 on the 2.4 GHz Wi-Fi band.

My question is, how wide is the signal being emitted from the magnetron? Does it spew radiation in a wide band on either side of 2450 MHz or is it fairly narrow?

Does anyone have a spectrum analyser and can possibly find out?

[mikeselectricstuff]

There should be some product standards that will specify the limits of radiation from microwave ovens

[Halcyon]

There should be some product standards that will specify the limits of radiation from microwave ovens

I would think so but I'm having trouble finding exact or reliable information. Perhaps if someone could heat a cup of water and wave their analyser probe around for a few seconds and report back findings? :-)

By the way, this is the chart I made which I'm using to plan/visualise channels, power levels etc... (Specific to Australia), because I like colours and Microsoft Excel.

[Gyro]

I suspect that one of the problems is that the magnetrons have a mechanical manufacturing spread, which will affect the exact output frequency - maybe the tolerance of individual anode cavities causes frequency spread. It may be that thermal expansion during operation also shifts the frequency?

[richard.cs]

As I understand it they are fairly narrowband, but are weakly voltage dependent, so what you actually get due to the PSU design is a narrowband tone that is swept up and down (or down and up) the band for half a mains cycle, and then goes off again for the other half cycle. They're not tightly toleranced in frequency because it depends on the mechanical dimensions, and spec-wise they just have to stay inside the ISM band. Essentially I expect you to get a narrowband tone swept over perhaps 20 MHz, anywhere inside or even slightly outside 2.4 to 2.5 GHz, and for a given oven off for just over half a mains cycle, but not necessarily the same half cycle as another oven.

I've not played with the inverter ovens, but unless they specifically try and recreate the half-wave gap they will be continuous and narowband with likely minimal sweep, but still anywhere in the ISM band.

In a crowded urban area you should expect interference from possibly multiple ovens at once, with a number of loud narrowband carriers, most sweeping around by 20 MHz or so and on around half the time, a few CW but even those with significant thermal drift, likely multiple MHz per minute. Any design assumption that assumes microwave ovens stay only in WiFi channels 8 and 9 is likely to be seriously flawed.

[dmills]

Even the inverter ones probably use the maggie as a rectifier so you might see 20KHz or so of 50% duty riding on the swept carrier.

Worse, if I was tasked with designing the worlds cheapest and nastiest 1kW @ 2500V inverter driver it would actually track the envelope of the mains (because why not, it saves me a bulk cap AND makes my power factor better).

About all you can really say is that a microwave will leak energy in the 2.4 - 2.5GHz band, what spectrum and centre frequency of any given unit is going to be more then a bit random. 

[Halcyon]

I knew this wasn't going to have a simple answer :-)

I'll experiment some more and run some RF environment tests with it running, see if the APs even care too much about it.

[Ampera]

Unless you can't upgrade all your devices to 802.11N 5Ghz or higher, I highly suggest you rely primarily on 5Ghz. I live in a slightly crowded area in terms of 2.4Ghz RFI, and using a 2.4Ghz network rendered all of my devices completely and totally useless.

If you're really concerned, do what the US radio quiet zone does and build a Faraday cage around your microwave. 

I use 2.4Ghz networking purely for legacy purposes. Unless you have all your laptops only with Mini-PCI, and not mini PCI-E for WLAN, I would bite the bullet and get a big bag full of 802.11ac WLAN NICs

Although, not sure if you live in a congested region or not, if you don't have a lot of RFI from other people, it might not be that bad. Personally my microwave never screwed with my network, and it's normally never even on long enough to make it much of an issue. If you think of it, microwaves are already shielded a bit. If you want to know how strong it is, get a mobile device and load up a 2.4Ghz band tester, a frozen chicken parm (or a bowl of water if you don't want dysentery, never run microwaves without something in it), and scan for what you might get. I'm not sure if those signal sniffers only respond to some sort of regulated 2.4ghz or not, I don't see why they would have to.

I mean, not to insult you, normal people wouldn't care, which is a strange thing to say to someone on the EEVBlog, as nobody here is normal. 

[jadew]

I made a little test with my oven.

The green trace is the peak power over 5 minutes. The blue trace is the same thing + 30 seconds of turned on oven.

It looks like most of the power is concentrated in a bandwidth of less than 10 MHz (on this oven at least), which is just outside of the band used by wifi (2.488 GHz), but it seems to move around and I think the extra power in the wifi channels is a result of the oven disturbing the wifi networks.

This is just a singular data point, but based on it, if I thought microwave ovens would be operating a lot around my network, I'd probably pick one of the lower channels.

You should also get one of those wifi analyzer apps for your phone, they're great for finding the quietest channels at a particular location.


Edit: Scratch that. I made another measurement with live update and it's all over the place, it's just that in the max-hold mode, the trace is masked by the regular wifi activity.

Edit 2: It does seem tho that it concentrates its power around specific frequencies and hops between them, which is repeatable, so if you had a spectrum analyzer you could at least be able to dodge a particular oven.

[Mr. Scram]

Is OP designing an internet of things microwave oven? 

[chris_leyson]

Bandwidth is quite wide maybe 100MHz and they drift with temperature, not the nicest of rf sources.

Is OP designing an internet of things microwave oven? 

Bosch and maybe Siemens already sell IOT ovens.

[Cerebus]

Unless you can't upgrade all your devices to 802.11N 5Ghz or higher, I highly suggest you rely primarily on 5Ghz.

You're suffering from "Didn't read the OP's post syndrome".

I'm in the process of designing a wireless network which unfortunately requires use of the 2.4 GHz band (as congested as it is).

[CopperCone]

I want microwave ovens to migrate to 5ghz. Wifi is getting off too easy now.

[Mr. Scram]

You're suffering from "Didn't read the OP's post syndrome".

It's sound advice. Without any additional clarification on the how or why, relying on a passing mention is a terrible idea. People suggest or even explicitly mention specific limitations or requirements all the time that when asked turn out to be completely arbitrary or irrelevant. What people say and what they're really asking for is generally worlds apart.

Let's omit the pedantry and accept suggestions that strays close or far with grace. If OP already considered the suggestion there's no harm done, and it might on some occasions lead to new insights.

[Ampera]

Likewise, I did read the original post, and decided to suggest 5Ghz over all other options, while suggesting ways it could be more practically done. I had consistent and constant issues with 802.11 2.4Ghz, which is why I am advocating against it.

[Cerebus]

Likewise, I did read the original post, and decided to suggest 5Ghz over all other options, while suggesting ways it could be more practically done. I had consistent and constant issues with 802.11 2.4Ghz, which is why I am advocating against it.

It was just meant as a "I think you've missed this" but obviously you hadn't. The poor old OP is saddled with it as a requirement and made it clear that he knows it's sub-optimal, so it's just rubbing salt into the wound to tell him, yet again, what he already knows.

[Mr. Scram]

It was just meant as a "I think you've missed this" but obviously you hadn't. The poor old OP is saddled with it as a requirement and made it clear that he knows it's sub-optimal, so it's just rubbing salt into the wound to tell him, yet again, what he already knows.

Again, you seem to assume a lot more than is actually stated. What you say may very well be the case, though the statement can also mean a number of other things. Especially with no reason given. Meanwhile, TwoOfFive provided new information that helps to shed light on possible issues if this route is pursued. Construing that as malice seems to take some creativity.

Can we now stable our high horses? It's better to have a thread with fifty too many suggestions than a thread with one too few.

[Cerebus]

You're suffering from "Didn't read the OP's post syndrome".

It's sound advice. Without any additional clarification on the how or why, relying on a passing mention is a terrible idea. People suggest or even explicitly mention specific limitations or requirements all the time that when asked turn out to be completely arbitrary or irrelevant. What people say and what they're really asking for is generally worlds apart.

Let's omit the pedantry and accept suggestions that strays close or far with grace. If OP already considered the suggestion there's no harm done, and it might on some occasions lead to new insights.

It's not pedantry, it was just meant to be a "heads up". The OP made it clear that he had a requirement and understood the drawbacks with that, vis:

... which unfortunately requires use of the 2.4 GHz band (as congested as it is).

It is, frankly, a bit insulting to the OP to ignore what he said and offer him advice he clearly doesn't need. I could have made quite a pointed remark to that effect but I chose something a bit shorter and less judgemental because I know that TwoOfFive is a smart cookie and I actually thought that his sin was merely missing the fine detail of the original point. We've all sometimes read a bit too fast and respond a bit too quickly. 

TwoOfFive has also demonstrated that he's quite big enough and smart enough to take me down a peg or two if he thinks he needs to, without your meagre help. You seem to have developed a habit recently of pouncing on anything I say, with the most negative interpretation possible, at every opportunity. I wish you wouldn't as it's not helpful, big or clever. If you want a slanging match with me you ought to have had the decency to PM me rather than forcing everybody else to endure the rather dull and predictable stream of needling and counter needling.

[Mr. Scram]

You're suffering from "Didn't read the OP's post syndrome".

It's sound advice. Without any additional clarification on the how or why, relying on a passing mention is a terrible idea. People suggest or even explicitly mention specific limitations or requirements all the time that when asked turn out to be completely arbitrary or irrelevant. What people say and what they're really asking for is generally worlds apart.

Let's omit the pedantry and accept suggestions that strays close or far with grace. If OP already considered the suggestion there's no harm done, and it might on some occasions lead to new insights.

It's not pedantry, it was just meant to be a "heads up". The OP made it clear that he had a requirement and understood the drawbacks with that, vis:

... which unfortunately requires use of the 2.4 GHz band (as congested as it is).

It is, frankly, a bit insulting to the OP to ignore what he said and offer him advice he clearly doesn't need. I could have made quite a pointed remark to that effect but I chose something a bit shorter and less judgemental because I know that TwoOfFive is a smart cookie and I actually thought that his sin was merely missing the fine detail of the original point. We've all sometimes read a bit too fast and respond a bit too quickly. 

TwoOfFive has also demonstrated that he's quite big enough and smart enough to take me down a peg or two if he thinks he needs to, without your meagre help. You seem to have developed a habit recently of pouncing on anything I say, with the most negative interpretation possible, at every opportunity. I wish you wouldn't as it's not helpful, big or clever. If you want a slanging match with me you ought to have had the decency to PM me rather than forcing everybody else to endure the rather dull and predictable stream of needling and counter needling.

You say that it's clear what Cerebus meant. I say it isn't. The fact that we're having this conversation tells me the latter might be true. There are thousands of scenarios that could lead to the statement being made and you seem to be sure which one it is. I'm merely stating that the information provided is too meagre to say anything conclusive. Your assessment isn't unfair, just a bit premature.

Please don't make this into a personal thing. I have no bone to pick with you, and never had in the past. If you perceive this differently I'd be more than willing to look into your concerns. I have no interest in a feud.

[gregariz]

My question is, how wide is the signal being emitted from the magnetron? Does it spew radiation in a wide band on either side of 2450 MHz or is it fairly narrow?

Does anyone have a spectrum analyser and can possibly find out?

I've made the measurement for work over an extended period, so unfortunately I can't post it here. Basically it spews it over the entire 2400-2480 band and is indeterminate at any point in time. This is partly due to the types of magnetrons being made for ovens are very cheap and secondly their supplies are half wave rectifiers so are very noisy. The top of the ISM band is now reserved for other things.

There are regs all over this of course as Mike suggests to limit human exposure - but its still recommended you don't stand in front of it with your eyeballs glued to the door. ISO regs will cost a couple hundred per piece.

[Cerebus]

You're suffering from "Didn't read the OP's post syndrome".

It's sound advice. Without any additional clarification on the how or why, relying on a passing mention is a terrible idea. People suggest or even explicitly mention specific limitations or requirements all the time that when asked turn out to be completely arbitrary or irrelevant. What people say and what they're really asking for is generally worlds apart.

Let's omit the pedantry and accept suggestions that strays close or far with grace. If OP already considered the suggestion there's no harm done, and it might on some occasions lead to new insights.

It's not pedantry, it was just meant to be a "heads up". The OP made it clear that he had a requirement and understood the drawbacks with that, vis:

... which unfortunately requires use of the 2.4 GHz band (as congested as it is).

It is, frankly, a bit insulting to the OP to ignore what he said and offer him advice he clearly doesn't need. I could have made quite a pointed remark to that effect but I chose something a bit shorter and less judgemental because I know that TwoOfFive is a smart cookie and I actually thought that his sin was merely missing the fine detail of the original point. We've all sometimes read a bit too fast and respond a bit too quickly. 

TwoOfFive has also demonstrated that he's quite big enough and smart enough to take me down a peg or two if he thinks he needs to, without your meagre help. You seem to have developed a habit recently of pouncing on anything I say, with the most negative interpretation possible, at every opportunity. I wish you wouldn't as it's not helpful, big or clever. If you want a slanging match with me you ought to have had the decency to PM me rather than forcing everybody else to endure the rather dull and predictable stream of needling and counter needling.

You say that it's clear what Cerebus meant. I say it isn't. The fact that we're having this conversation tells me the latter might be true. There are thousands of scenarios that could lead to the statement being made and you seem to be sure which one it is. I'm merely stating that the information provided is too meagre to say anything conclusive. Your assessment isn't unfair, just a bit premature.

I think that you're the one demonstrating, quite clearly, that you don't know who said what. I don't think the OP's wording needs careful textural analysis to determine his meaning.

Please don't make this into a personal thing. I have no bone to pick with you, and never had in the past.

Experience would demonstrate differently.

If you perceive this differently I'd be more than willing to look into your concerns. I have no interest in a feud.

Now you sound like a carefully prepared corporate statement. In a conversation between two individuals that is a very, very odd tone to take. I can't think of a tone more likely to make a bloke think that there was something 'sus' going on.

Anyway, I tire of this, as I'm sure that any poor, bewildered onlookers do too. I'm off to cook my supper and, probably, adjust my ignore list. Life's too short...

[Ampera]

This sounds like a case of two people pissed at each other, without knowing exactly why.

I do think it's a good idea, and I have urged him to reconsider, with more ideas to back it up because I've had so many problems with 2.4Ghz in the past. I also stated reasons why that might be the case, in an attempt to acknowledge the fact that I might not know all the details, and that there are genuine reasons to rely on a 2.4Ghz network.

Halcyon's probably gonna come back and wonder wtf happened/is wrong with us, so I suggest we let whatever stupidity this is cease, brush off hard feelings, and I recommend that you two figure out what's up between the two of you, but that's none of my business. 

I think anybody who tries to use wireless communication with any real throughput should understand and expect the limitations. Wireless communications will most likely never be as fast as a direct connection.

I personally use a Ubiquiti Uni-Fi AP-AC-LR, which was I believe cheaper than the pro version at the time. Some people have gone on to say the LR stuff is a bit wank, but despite having no real comparisons, besides a 2011 AP/Router that only does 802.11n 2.4ghz, the range is quite positively impressive. I'm not sure if it maintains 5Ghz the entire way, it might step down to 2.4Ghz, but that is actually another point. Most 5Ghz APs that I have seen are dual band, and will work in 2.4Ghz and 5Ghz at the same time, and in theory should "seamlessly" switch to 2.4Ghz when the 5Ghz signal drops out, and will maintain compatibility with legacy devices.

I actually just installed an 802.11n card into my ~14 year old Prescott Celeron laptop to replace it's 802.11g card that I didn't want to use for fear of bottlenecking other devices (which can happen on some, if not all APs, at least to some extent). I don't know if it's a 5Ghz card, but it hooks up to my AP, and works flawlessly on Arch Linux.

As stated, I'd honestly really only be concerned if I am either running crucial applications (which shouldn't be on wireless anyways), or if the microwave REALLY gets used a lot to the point where severe downtime can be expected.

[Halcyon]

Guys, no need to fight. Allow me to clarify.

My "requirement" to use a 2.4 GHz network is simply to maintain legacy gear which does not have the ability to be upgraded to 5 GHz cards. For the most part, I only use 5 GHz Wi-Fi networks with enterprise authentication, but there are a small number of devices which don't support this so I'm forced to maintain the older slice of the network. Where possible I use an Ethernet cable, but that isn't always possible either.

That being said, the 2.4 GHz part of the network is isolated so it only has access to very limited domains on the internet and sits on its own VLAN.

I'm fully aware that the 2.4 GHz ISM band is congested with not only microwave ovens, but Bluetooth, cordless phones etc... I just figured if I can at least avoid one of the largest radiators from being an issue then that's one less thing to worry about. But it seems that based on the measurements and data provided, this is not entirely possible.

I do use Wi-Fi Analyser on my Android, however it only looks for Wi-Fi, It doesn't recognise other non-Wi-Fi sources of interference whereas the UBNT RF analyser does.

Thanks everyone for their comments. I've decided to use channel 13 on the 2.4 GHz band (which is legal in Australia). It seems as though none of the visible devices around here want to use that channel at all (for whatever reason) and they are all clustered around the common channels 1, 6 and 11.

[Gyro]

Thanks everyone for their comments. I've decided to use channel 13 on the 2.4 GHz band (which is legal in Australia). It seems as though none of the visible devices around here want to use that channel at all (for whatever reason) and they are all clustered around the common channels 1, 6 and 11.

That sounds like a good decision - I use channel 13 on my router too (also have legacy stuff). Most routers seem to default to the bottom of the band and the auto switching ones seem to be a bit deterred from picking 11 when they find something up at 13, that one might just be my imagination though! Either way 'taking the high ground' seems to get me as far from interfering sources as possible.

If you've got a laptop, try installing inSSIDer https://www.metageek.com/products/inssider/ and wandering around, it gives a real-time rolling graph or different SSID strengths as well as the normal level by channel number display. Metageek also do products that sense other, non wifi sources but I think those rely on their WiSpy plug is spectrum analyser.

P.S. They seem to be charging for inSSIDer now, though there's still a demo version available. You might want to look around for a download of one of their previous freeware versions, either V2.1 or V3.

[radar_macgyver]

Here are three spectrum analyzer plots. First is in swept mode from 2.35 to 2.55 GHz, with a small directional antenna connected to an RSA306. You can see my home router at 2.457 GHz (channel 10), and the rest of the neighborhood camping on channel 6. Couple of narrow-band signals at the edge of the ISM band as well.

Next, same settings but with my microwave switched on about 5 meters away. I had a glass of water as a load. I suspected saturation, so I re-did this with attenuation, it scaled linearly. I think I'm going to switch my router to channel 1 instead...

Finally, I have a 40 MHz span with real-time mode. You can see some interesting behavior, with the magnetron frequency sweeping around (see the spectrogram, with a time scale of 0.5 s/div). I wonder if this is due to the 'mode stirrer' inside the cavity. Wish I had one of the newer Tek RTSAs that can do a wider bandwidth in real-time mode.

Keep in mind the RSA306 (I have the original, not the -B version) is not known for good out-of-band filtering, so some of the spurious tones may be internally generated. Given the power levels, though, I suspect that they are real.

In general, magnetrons used in ovens are all over the place. The ones used in radars use a coaxial cavity to stabilize the operating frequency. They still drift ~10-15 MHz due to temperature variations from when they're cold to when they've been running for many hours. They also have a much more stable and matched load than the cavity of an oven, so load pulling won't affect their operating frequency as much.