A Linux-powered microwave oven
lwn.net
lwn.net
After searching LWN you find a comment at the bottom that you need to run fuser on /mnt/microwave/0/door/mechlock and kill the dbus_microwave_door_status daemon.
Search google again to discover a blog from 2004 detailing how you use dmesg and modprobe to find out which number your microwave door is bound to.
[1] http://whizoo.com/reflowoven [2] http://www.nxp.com/products/rf/rf-power-transistors/rf-cooki...
[0] https://www.google.com/search?q=solid+state+microwave+oven
But, in the wrong hands, that's a directed-energy weapon, isn't it? (Hell, stick one of these in the right kind of tube and you've now got a coherent RF beam—a maser pointer. You could probably burn your name into a wooden building from across a football field with one.)
Thinking of that book, I don't think it's likely that libraries would carry that kind of book any more, sadly.
http://www.amazon.com/Build-Laser-Phaser-Working-Projects/dp...
Looking at NXP's "RF Cooking" page, it appears they're pushing this stuff with some optimism... (http://www.nxp.com/video/sage-freescales-vision-of-an-intell...)
My personal, electronic microwave, has values 125/250/500/750W so not completely exponential (I guess, that's what you meant?).
One thing is, you can always (?) set arbitrary time, so you can compensate to gaps in power settings.
Other thing, maybe electronics are more simple with power settings like these?
When I see exponential settings I would always think of analogy to photography and 'f-stops' [1]. Some devices have exponential settings due to the physical nature of something, but is it the case with microwaves, I don't know.
[0] http://www.firstiraq.com/sites/default/files/products/5002-1...
I have almost no idea how a magnetron actually works.
You can see this effect if you run a microwave on low power for a short period - sometimes the magnetron will be running for the full duration.
Some GEs claim to use inverters but only for defrosting. I've been wondering why they wouldn't use it for other functions.
http://electronicdesign.com/blog/say-good-bye-last-vacuum-tu...
Using the MHT1003N, manufacturers can use from one to eight of these 250 W units to build a microwave oven with the desired power level. And the magnetron’s 4 kV power supply goes away in place of a supply of 28 to 50 volts. Furthermore, the crude on-off control of the magnetron can be replaced with full variable power control.
However, as far as I know there's no consumer devices containing things like this yet - the only ones I've seen were concept devices from Freescale. This article suggests we won't see them until 2017, and then only in high-end cookers:
http://www.electronicsweekly.com/news/design/the-future-of-m...
The thing is, kitchen appliances are HARD. Appliance designers spend careers doing the stuff and have manufacturing know-how and are backed by serious resources. A dude installing linux in a microwave is going to have a good time and perhaps get other folks excited and I am all for that (and have done ghetto-retrofits myself).
But there are good microwaves out there, they just cost more than others. If you just want a good microwave and want to encourage better designs, fork over the money and buy them! On the other hand, it is just a microwave. There are better ways to cook food... don't overthink it!
What the heck are you talking about? Linux has rather decent realtime facilities. It's not a hard real-time, it's soft real-time, but that's good enough for many things. It's used successfully by the LinuxCNC project for controlling CNC machines all the time, and that requires millisecond precision for running stepper motors through a parallel port. There is no way that running a toaster oven is more demanding than that.
Hell, a reflow toaster oven can be run with a controller with latencies of a second or more! It's not going to make a difference there; there's far too much latency between activating power to the heating elements and seeing a temperature change for millisecond precision to be necessary. The only problem with Linux on a toaster oven is that it's just plain overkill; an Arduino is simpler and cheaper. SMD reflow does NOT require "delicate control of temperature and timing". Reflow ovens are not that accurate. They just have to follow the recommended profile well enough, and it's not that hard once a controller has learned the characteristics of the oven.
I wouldn't run avionics systems needing hard real-time guarantees on Linux, but for soft real-time systems, Linux works well enough. To say that its RT capabilities are "so poor" is just ridiculous, unless you're looking at Linux circa 2001 or something.
Somehow, the software on microwaves has gotten worse over time (even without becoming "smartmicrowaves"), which I didn't think was possible. Limited sample, but I've seen three new microwaves, in '99, '10, and '14.
The first one was really easy to use: buttons 1-6 were a quick-start for 1 to 6 minutes. e.g. hit 3 and it immediately starts with a 3-minute timer. If you hit start, it would immediately start with 30 seconds (or add 30 if it was already started).
Second one ('10) was basically the same model, but with worse software: if you used the number quickstart, it somehow couldn't accept I/O for the next half second. So if you wanted to do 90 seconds (common operation), you would have to pause after hitting 1, then hit +30. A really annoying inconvenience, and completely unnecessary.
The third microwave ('14) was admittedly a different model, but doesn't seem to have learned from previous microwave innovation. The start/+30 button doesn't work for a quick start, only for adding to an active, ticking timer. And cancel no longer deletes previous input, but emits an error sound.
How hard is it not to make microwave UI software worse?
If he gets the voice control worked out to where I can say "Tea, Earl Grey, Hot" I might be sold.
Is it really that hard to push "Time cook -> 2 -> 3 -> 0 -> Start"? I mean come on I can understand wanting to optimize your life or whatever but complaining about microwave buttons seems a bit ridiculous.
TL;DR: pressing buttons is easy, but sometimes they press your buttons back.
I only upvoted because I felt smart for recognizing "au". (Not a high-school graduate.)
Haven't pushed more than two buttons in almost a year.
The microwave oven I've got doesn't let you use the number keys to specify the power level. Each press of the power level button decrements the power level by 10%, and the first press doesn't count. So to cook something at 30%, you need to press the power level button 8 times before typing in the cooking time.
One knob with rotary encoder for time (turn faster for minutes).
One knob with rotary encoder for power.
Both settingscan be changed even while microwave is running.
Not a crappy brand, a mid-range product I think. Costs around 180 EUR here.
I have trouble finding US shops that have it, maybe it's Europe-only brand, but hope you have better luck!
http://www.amazon.co.uk/Russell-Hobbs-Microwave-Litre-White/...
One knob for time, the other for power. Done.
http://www.radar58.com/radarange/fullsize/2Panel.jpg
Mine is like that too, and it works really well and simple to use compared to the alternative of a dozen not-very-tactile buttons. It's based on a mechanical timer, so no buggy firmware either.
It integrated a raspberry pi with a barcode scanner and microwave. All you needed to do to cook prepackaged food was scan the barcode. It had all the times set in and all of the mixing and instructions.
I think that feature is useful.
This is important
So he might actually be able to eat corn under GPL license.
I'll show myself out.