Microwave oven PWM frequency

[rea5245]

Anyone want to jump into the middle of a domestic dispute? (*)

My wife, who uses the microwave oven a lot more than I do, says that when she puts something in the microwave oven at low power, she can tell that the microwaves are turning on and off with a cycle time of a few seconds. E.g. she can see water boil for a few seconds, then stop boiling for a few seconds, then start again.

I, while candidly admitting that I didn't know for sure, expressed surprise that it would work that way. I had assumed that the magnetron would be pulsed on and off at at least 1Hz, if not faster.

Yet she does seem to have compelling evidence. Why would the magnetron be cycled at such a low rate? Is it incapable of being turned on and off faster? Or did the designers not think it would matter?

Thanks,
   Bob

(*) Rest assured that no matter who is right, my marriage will not break apart.

[Fred27]

Your wife is right. You can even hear it turning on and off every few seconds - maybe even slower than that.

[Cliff Matthews]

FWIW, listen for fan rpm variations. I've had several since the 80's and don't recall any rate faster than about 0.5hz..

[tom66]

Your wife is correct.

The average magnetron takes about 1-1.5 seconds to warm up from cold because it's a vacuum tube and the heater needs to start boiling off electrons.  You can often hear the tone of a microwave change after the first second of operation.

So any PWM needs to be quite slow.  Also,  the inrush current on the transformer on each cycle would be likely to blow fuses if it happened so rapidly. 

Perhaps keeping the tube warm using a separate supply would be one solution, but magnetron heaters are quite power hungry.  One example I saw was 3.5Vac at 25 amps.  So the additional supply for this is probably not worth it for such a mild improvement.  It is just derived off the main transformer.

Inverter microwaves, marketed by Panasonic and possibly others, still take time to warm up but the inverter directly controls the power to the magnetron.  So the low power setting does not require PWM and the power delivery is constant.  The lowest settings (defrost and warm on my Panasonic oven) still use PWM because magnetrons are not good at producing small amounts of power.  The PWM has a longer period and higher frequency though (about 0.2-0.3Hz).

/rant:
What does annoy me is despite knowing that a magnetron takes a second to warm up - and despite being able to directly monitor magnetron current on the inverter units - no microwave seems to delay the timer if the magnetron is not generating microwaves.   If you want to nuke something for ten seconds, if the magnetron is completely cold it can take 3-4 seconds to start delivering power.  This makes timing inconsistent at the lower end of the scale. 

[Zero999]

She is right. I remember dismantling an old mechanically controlled microwave, when I was about 12 years old and looked at the power control. It consisted of a synchronous motor driving a gear system, which turned on and off a microswitch. Adjusting the power control altered the position of a cam, so the duty cycle was adjusted. The switching frequency was very low, much under 1Hz.

Now there are some inverter microwaves, which don't use a large iron cored transformer, but a switched mode power supply. I believe these either use a higher frequency PWM, or just alter the current through the magnetron, in a linear fashion.

[Cliff Matthews]

You admitted you're not sure and call this a dispute? Nah, my wife's a Cuban surgeon and if anything, she's always sure!

[coppice]

Traditional microwave ovens, with a big heavy transformer, use bursts of microwaves several seconds long. However, many of the inverter driven ovens (perhaps all) rapidly modulate the microwave power.

[rea5245]

You admitted you're not sure and call this a dispute?

I like to add drama and suspense to my dumb questions.

- Bob

[tooki]

Also, there are often two things at work:
1. the duty cycle of the microwave — usually it's 100% at maximum power. At lower power, most microwaves have a PWM frequency in the seconds. Like 10 seconds on, 10 seconds off for 50%. Inverter microwaves (like my Panasonic) seem to both vary the microwave intensity and the duty cycle.

2. The food hitting hot and cold spots in the cooking cavity. Most consumer microwaves use a turntable, and as the cup of water spins around, it hits hot and cold spots. Commercial microwaves and some high-end consumer ones instead use a "stirrer" under the floor of the cooking cavity, which is a slow-turning fan-blade kind of device that accomplishes much the same thing, but with the microwaves moving instead of the food.

[Halcyon]

As tooki mentioned, it depends on the microwave. I have a Sharp "inverter" model at home which makes no difference in sound whether at low power or maximum power. There is no audible on/off at all and liquids, once beginning to simmer/boil, will do so continually (albeit at lower intensity at lower microwave power). That's because the magnetron is always on but at variable power levels.

However more commonly (like all the microwaves we have at work) they regulate the "power" level by adjusting the duty cycle. When the magnetron is on, it's on at full power. So lets say I set the microwave to half power, it will switch on the magnetron for 10 seconds, then off for 10 seconds. You can hear a definite "clunk" as it kicks back in every time. You also see the effect your wife is talking about where the water will go from a boil, to hardly anything at all. On the "low" setting, you might have the magnetron on at full power for 2 seconds then off for 10, for example.

Why? It's cheaper to build a microwave to simply turn the magnetron on and off at regular intervals than it is to include an inverter and more advanced software.

[John Heath]

A microwave oven is switching on and off every 8 m seconds as well as every second or so. The reason for this is it is against the law for a microwave oven to run off a filter DC power supply. It must be a unfiltered supply raw off the full bridge rectifier thus on and off every 8 m second. The reason for this law is to sweep a gigantically stupid law passed before under the rug. The gigantically stupid law was to have a 1000 watt microwave  and a tiny 100 m watt WIFI in the same frequency range. The solution , equally as dumb , was to out law filtered power supplies in microwave ovens so that the tiny 100 m watt WIFI can sneak a few bytes out in a 2 m second window between power cycles of a monster 1000 watt microwave. Is it just me or did someone drop the ball here? Thank god for 5 GHz WIFI.

[tooki]

A microwave oven is switching on and off every 8 m seconds as well as every second or so. The reason for this is it is against the law for a microwave oven to run off a filter DC power supply. It must be a unfiltered supply raw off the full bridge rectifier thus on and off every 8 m second. The reason for this law is to sweep a gigantically stupid law passed before under the rug. The gigantically stupid law was to have a 1000 watt microwave  and a tiny 100 m watt WIFI in the same frequency range. The solution , equally as dumb , was to out law filtered power supplies in microwave ovens so that the tiny 100 m watt WIFI can sneak a few bytes out in a 2 m second window between power cycles of a monster 1000 watt microwave. Is it just me or did someone drop the ball here? Thank god for 5 GHz WIFI.

Do you have any sources for this supposed law and its supposed effects?

Microwave ovens predate WiFi by around half a century. I doubt they're bound by laws regarding WiFi.

I'm not even sure what you mean about a microwave operating from unfiltered DC. I mean, the magnetron isn't operating off DC, nor off mains frequency.



As for why microwave ovens and WiFi operate in the same frequency range: it's on purpose, because of the same root cause, though for different intents. It's the frequency range that is best absorbed by water. In a microwave oven, this is desirable to achieve optimum efficiency in heating the food. In WiFi, which operates in the unlicensed ISM band, the ISM band was placed in that range for the same reason: it's absorbed by water. This naturally limits the range of the signals by absorption by atmospheric water vapor, allowing the same frequencies to be reused in relatively close physical proximity to each other.

[Halcyon]

A microwave oven is switching on and off every 8 m seconds as well as every second or so. The reason for this is it is against the law for a microwave oven to run off a filter DC power supply. It must be a unfiltered supply raw off the full bridge rectifier thus on and off every 8 m second. The reason for this law is to sweep a gigantically stupid law passed before under the rug. The gigantically stupid law was to have a 1000 watt microwave  and a tiny 100 m watt WIFI in the same frequency range. The solution , equally as dumb , was to out law filtered power supplies in microwave ovens so that the tiny 100 m watt WIFI can sneak a few bytes out in a 2 m second window between power cycles of a monster 1000 watt microwave. Is it just me or did someone drop the ball here? Thank god for 5 GHz WIFI.

I think it might be you who has dropped the ball this time. This is the first time I've heard of such a thing as well. This is certainly not the case in Australia. Yes, microwave ovens operate within the same frequency range as 2.4 GHz Wi-Fi however in my experience it's rarely a problem if your microwave is operating / built correctly. Microwave ovens are designed to keep as much of the RF inside the cavity as possible. If it's leaking significant RF, then you have bigger issues than just your Wi-Fi dropping out. Microwave leakage testing can be a bit of a wank (like PAT testing) and there are people making money off it, but significant levels can cause problems. There are a number of standards and studies out there which dictate "safe" limits, but leakage of 5 milliwatts per square centimetre or less is considered acceptable.

I have absolutely no dramas with my 2.4 GHz Wi-Fi when my microwave is operating at full power (1200 watts). When it first starts up and ramps up to full power, it can draw as much as 1900 watts @ 240v. Even old devices manage full signal strength with no perceivable difference in throughput.

My microwave also doesn't switch off the magnetron at those intervals, in fact, it's ALWAYS on, regardless of power setting as it has an inverter.

[Red Squirrel]

I can kind of tell when the magnetron turns on/off.  It's almost like I can "feel" it, not like feel myself getting zapped because that would mean my microwave is leaking, but there's an odd feeling of magnetic force in the area. Hard to explain.   Same thing when there is a power outage.  When the power comes back on I can feel and hear the magnetic force of the initial "hit".  If I'm in bed it will actually wake me up at times, and I look at the clock and sure enough it's flashing 12:00.    Not sure if that's normal though, maybe I'm crazy? lol.

But yeah microwaves just turn the magnetron on/off in couple second intervals based on the setting.  For defrost you actually want that, because any water will heat up, and that water will help melt ice to create more water.  If it was on continuously the water would evaporate.  Ice does not really get affected as well by microwaves so it would take a long time to melt.  I'm not sure what things like the pizza setting or popcorn setting do though, is there actually a science behind those? 

[Halcyon]

I'm not sure what things like the pizza setting or popcorn setting do though, is there actually a science behind those?

They just vary the duty cycle (or in the case of inverter models, the output power) of the oven throughout various stages of cooking based on a program. Some programs might start off slow then ramp up towards the end of cooking, some might stop part way through and prompt you to stir/turn whatever you're cooking. It's really nothing fancy or complicated, they are just based off tried and tested recipes by the manufacturer. The microwave doesn't "know" when the food is actually done, it just bases it off pre-programmed weights, cooking sequences and power levels.

For example, if I tell mine to defrost 500 grams of chicken, it knows based on that weight, it should take 12 minutes and 1 second (at a certain power level) to defrost that mass of particular meat.

My microwave (Sharp Carousel) also comes with its own cookbook and the microwave itself has built-in programs based on those recipes. Most brand name microwaves will have something similar.

[tooki]

They just vary the duty cycle (or in the case of inverter models, the output power) of the oven throughout various stages of cooking based on a program. Some programs might start off slow then ramp up towards the end of cooking, some might stop part way through and prompt you to stir/turn whatever you're cooking. It's really nothing fancy or complicated, they are just based off tried and tested recipes by the manufacturer. The microwave doesn't "know" when the food is actually done, it just bases it off pre-programmed weights, cooking sequences and power levels.

That's not necessarily true. Various microwave oven models over the years have used various sensors to actually try and gauge the cooking progress. For example, a different variant of my Panasonic microwave/convection oven/quartz grill/steamer uses infrared sensors for automated defrosting. And I know there have been microwaves with microphones in them to literally listen to the popping frequency of popcorn to cook it automatically. (Since the way to know it's done is when popping slows to about 2-second intervals.) And of course some models have thermometer probes.

[Halcyon]

They just vary the duty cycle (or in the case of inverter models, the output power) of the oven throughout various stages of cooking based on a program. Some programs might start off slow then ramp up towards the end of cooking, some might stop part way through and prompt you to stir/turn whatever you're cooking. It's really nothing fancy or complicated, they are just based off tried and tested recipes by the manufacturer. The microwave doesn't "know" when the food is actually done, it just bases it off pre-programmed weights, cooking sequences and power levels.

That's not necessarily true. Various microwave oven models over the years have used various sensors to actually try and gauge the cooking progress. For example, a different variant of my Panasonic microwave/convection oven/quartz grill/steamer uses infrared sensors for automated defrosting. And I know there have been microwaves with microphones in them to literally listen to the popping frequency of popcorn to cook it automatically. (Since the way to know it's done is when popping slows to about 2-second intervals.) And of course some models have thermometer probes.

Indeed, I was simplifying it. Mine has "sensor" printed on the front (what the sensor is exactly, I don't know, temperature, humidity? It's not obvious by looking inside the cavity). But despite those sensors, it's largely based on pre-programmed recipes and some fuzzy logic. The oven doesn't truly know when food is cooked or not, it just guesses. I find the auto-defrost function on mine actually a bit too much, it tends to start cooking parts of the chicken as opposed to just defrosting it and the pizzas come out ridiculously hot and over-done.

Having said all that, I suspect they err on the side of "over-cooked" as opposed to "not cooked enough" to meet certain food safety guidelines.

[NiHaoMike]

If regular microwave ovens are "off" a bit more than half the time it's running, why hasn't Wifi in 2.4GHz detect for that and try to put packets in those times? Or does the IEEE expect that most microwave ovens will go inverter and thus make such a mechanism less useful?

[Halcyon]

If regular microwave ovens are "off" a bit more than half the time it's running, why hasn't Wifi in 2.4GHz detect for that and try to put packets in those times? Or does the IEEE expect that most microwave ovens will go inverter and thus make such a mechanism less useful?

They do, in a sense. Wi-Fi won't broadcast if it detects that the channel is "busy". This is why if you have many radios nearby on the same channel, your performance sucks, because it essentially waits for a break in transmission to avoid collisions with other Wi-Fi devices.

Why don't they don't do this with other sources of interference? Probably because it's a little pointless. In the case of microwave ovens, they SHOULDN'T leak enough to cause significant problems with Wi-Fi (see my comments above).

The way TCP/IP is designed, if the client doesn't receive the expected response within a certain time, it just times out, disconnects the session etc... Same for your Wi-Fi access point, it won't just buffer every single TCP/IP packet that comes its way, if it can't deliver it within a timely manner, it won't deliver it at all. If your Wi-Fi router or access point had to wait for all sources of interference to cease in built-up areas, no one would be transmitting anything. Wi-Fi is a half-duplex system, while a radio is listening (or receiving data) it's not transmitting.

Which brings me to another point that I feel I should mention. Most consumer-grade Wi-Fi routers will have an "auto" channel function. Usually, this only operates at the time the router boots to listen for the best free channel. It does not continually monitor the spectrum and channel hop. For it to do this, it would need a dedicated radio that does the monitoring or disconnect all clients from the network, monitor the spectrum, then allow all clients to re-connect (the latter is obviously not ideal).

Most enterprise grade wireless gear will either have dedicated access points connected to the controller through various parts of a building/campus which act as "monitors", that's their full-time job, to provide the controller feedback on channel interference and other radios in that area so it can change the channel on nearby access points. Alternatively some dual-channel access points will have a third and sometimes fourth built-in radio to do this, like the Ubiquiti UAP-AC-SHD.

The auto channel feature on consumer gear is next to useless (like MAC address "security"). Let's just say you had a local power failure and all your neighbours lost power too. Assume your router is the first to boot up and decides that Channel 11 (on 2.4GHz) is the most "clear" so it picks that and starts accepting clients to connect. Then all of your neighbours' routers boot up, some have auto channel (which would at this point avoid Channel 11) but some are configured to only use Channel 11 (which is very common). All of a sudden your "clearest" channel is congested and performs like crap. Your router won't detect this and change channels, it will just continue doing its best in an RF-noisy environment.

[Red Squirrel]

Technically your wifi should not be affected by the microwave.  I don't know how good of a test this is but I did a ping -f to my phone's IP and microwaved a glass of water on high with phone sitting next to the microwave and I had no packet loss.

[John Heath]

I can see there are more than a few who are not buying the microwave oven problem in the 802 IEEE standards for WIFI , XBEE and  blue tooth. I can understand why as you would think they would be smart enough to separate 1000 watt microwave oven frequency from .1 watt communication frequencies. Sadly they did not separate them. Even Sadder if it is your phone that rings when computer communications are down.

[tooki]

Technically your wifi should not be affected by the microwave.  I don't know how good of a test this is but I did a ping -f to my phone's IP and microwaved a glass of water on high with phone sitting next to the microwave and I had no packet loss.

I suspect it depends on several things:
1) the selected WiFi channel
2) the specific microwave oven model
3) the relative location of the WiFi router and the microwave

The last, in particular, I’ve found to make a huge difference: if the microwave oven is between the router and the client, performance will suffer (sometimes to the point of not functioning at all). I just experienced this at a friend’s house. If they’re equidistant but in different directions, it usually works fine.

I can see there are more than a few who are not buying the microwave oven problem in the 802 IEEE standards for WIFI , XBEE and  blue tooth.

Yeah, I dunno. I’ve experienced it enough times firsthand to know it’s a real problem.

That said, I’m still waiting for documentation about these laws you’re claiming.

I can understand why as you would think they would be smart enough to separate 1000 watt microwave oven frequency from .1 watt communication frequencies. Sadly they did not separate them.

I already explained to you why they’re on the same band. It’s not random coincidence!

Even Sadder if it is your phone that rings when computer communications are down.

Huh what?

[richard.cs]

There is a microwave oven avoidance mode in one of the WiFi specs, I think it just keeps the packet lengths down to 8 ms or something relatively simple like that. Note that this is taking advantage of microwave ovens running with half-wave rectified power, but it's not the reason they work that way. Microwave ovens do that because it's cheap. The magnetron produces a fairly pure single frequency tone, but with a weak voltage dependence to it sweeps in frequency during the (slightly less than) half a cycle that it's active and this makes it more interfering than might be expected.

The one dual-magnetron microwave I've had apart runs them on opposing half-cycles. I suspect that this is both to clean up the input current waveform a bit but also to prevent them from injection locking - if they end up locking in frequency you get a single standing wave pattern and the cooking is no more even than a single magnetron microwave (which is one of the alleged advantages). This means that at full power there is minimal dead time.

Interference with WifFi is certainly a real effect, I have a pretty little graph of WiFi data transfer rate against time where you can see the effect of my microwave when on 50% power, the low-frequency PWM is visible in the data rate. This should be entirely expected given that the shielding on a microwave oven often only gives about 30 dB attenuation so gives leakage on the order of a Watt.

[John Heath]

There is also the crowding of WIFI channels. Each year it gets worse. As much as 50 WIFIs all frequency hopping trying to sneak a byte or 2 through a limited shared number of channels. There are free apps for cell phones to monitors how crowded it is.

There is also the window problem if you need to go building to building. Not sure why but there are some types of glass used in windows that acts like a mirror at 2.4 GHz. You can be better off trying to penetrate a brick wall than a glass window. 5 GHz does not help.

[Psi]

Microwaves that control the power level by ~0.25 HZ pwm are pretty annoying.
I would recommend replacing it with an inverter microwave.

When you need to warm up something slowly, like on 'low power' the pulses of full power usually burn/cook it instead.
A good example is anything with egg in it, like warming up hollandaise sauce, it's almost impossible to not ruin it.
But with an inverter oven you don't get these problems.