Easy Bake Vacuum Tube Oven
November 16, 2025
Introduction
Years ago there was a toy called The Easy Bake Oven that was able to cook brownies, cookies, and such with a 100 watt lightbulb. Since incandescent light bulbs are being phased out and harder to get, they changed their design to use a heating element closer to what a toaster oven would have.
This project was intended to be a parody of the original Easy Bake Oven where instead of running on a 100W lightbulb, the food would be cooked by two vacuum tubes. A thermocouple is used to monitor the temperature and report values to a computer in XML format. A microcontroller can also turn off the tubes if the temperature is high enough and turn them back on if the temperature drops.
Related Projects @mikekohn.net
| Food: | RC Food, PANCAKE-ROM, Cyborg Chicken, Cyborg Lobster, Edible MIDI Keyboard |
Video
YouTube: https://www.youtube.com/watch?v=vpWpraO1HWA
Explanation
I actually already tried this a couple years ago using 2 vacuum tubes I got from Gateway Electronics (R.I.P.) when it was around. The first time was done in a throw-away aluminum baking tin with with plastic wrap on top. It's been awhile, but I believe there were 12v heaters in the tubes with a max still of 5W. Being in a container that didn't keep in heat very well, the brownie of course did not cook.
A couple years later it seemed like this needed to be tried again, this time in a container that can hold in the heat and a thermometer to know how hot the air is.
The new vacuum tubes are only putting out about 1.89W each but this time they are contained inside of a thermos. A thermocouple inside the thermos is able to monitor the temperature. An ATtiny2313 microcontroller is able to turn the heating elements in the vacuum tubes on and off based on an upper / lower temperature threshold.
After preliminary testing (raising the temp to around 70C before stopping), it seemed like 120C might be possible, which should be high enough to cook something. The experiment was run twice where first attempt was done with a microwave brownie and a cookie. The second attempt was a saltine cracker, ketchup, and provel cheese which is a really gross sounding parody version of what would be a St. Louis style pizza.
On the first run the temperature rose at a pretty steady rate. The firmware in the microcontroller was set with an upper threshold of 120C and lower 110C with the thought that 120C would be enough to cook the food. After about 25 min or so, after the temperature reached the upper 70's or lower 80's degrees celcius, things appeared to start to level off. Seemed like it wasn't going to get much past 85C, at least not very fast. It felt like a good time to stop so the MOSFETs were disconnected to allow the oven to cool down. After opening the thermos, the cookie (as seen in the pictures below) appeared to have melted and the brownie mix is still liquid.
With what was learned from the first run it seemed like something that could cook with a lower temperature would make more sense. The threshold was dropped as 85C for the upper threshold and 80C for the lower. This time the XML data was logged to a file and could be plotted:
The firmware in the microcontroller worked really well, but 85C or the length of time the "pizza" sat at 85C wasn't really good enough. It's a bummer because it seems like cooking food like this is really possible, just probably requires more vacuum tubes and/or higher wattage on the bulbs. As a side note, over the last summer there were some super hot days here, so I got some cookies to cook on the dash of my car (something I've seen others do online and something I've done before). A thermometer in my car has shows on hot days the air temperature can reach 54C while the dashboard temperature can get up to 74C (I believe the metal tray that the cookies were on got higher, but the picture taken of the IR thermometer was lost). The cookies did cook, but they were also in the car for longer amount of time.
ATtiny2313 / Thermocouple
The microcontroller of choice here is the Microchip (Atmel) ATtiny2313. This seemed like the best choice since it can run at 5v which makes it easier to to open the gate of the MOSFETs that control the vacuum tubes. For temperature sensing, an SPI based MAX31855 is used.
;; Set up PORTB.
;; PB0: Software SPI_SCLK (MAX31855)
;; PB1: Software SPI_MOSI (MAX31855)
;; PB2: Software SPI_MISO (MAX31855)
;; PB3: Software /SPI_CS (MAX31855)
;; PB4:
;; PB5: MOSI (in-circuit programming)
;; PB6: MISO (in-circuit programming)
;; PB7: UCSK (in-circuit programming)
;; Set up PORTD.
;; PD4: LED
;; PD5: Vacuum Tube
The MAX31855 was really super easy to use. Just assert the chip select pin and clock in 14 bits (or up to 31 bits for some other info). The data is sign extended to 16 bit and converted to base 10 to it can be reported through the UART of the ATtiny2313 to a computer as XML. The XML is sent at a rate of around 1 to 3 seconds.
The microcontroller was programmed so that if the temperature hits an upper threshold, the vacuum tubes will be shut off and if it drops to a lower threshold, the vacuum tubes are turned on again. When making the cookies, the threshold was set to 120 as the high and 110 as the low. For the pizza it was set to 85 as the high and 80 for the low.
The ATtiny2313 also has a mildly accurate clock so it can report how long it's been running. Every time the high threshold of temperature is hit and the vacuum tube is shut off, the time this happened is recorded. All this info is reported in the XML format which is sent to a computer through the chip's UART:
<?xml version="1.0" encoding="UTF-8"?>
<max31855>
<raw>0145</raw>
<temp>81.25</temp>
<time>60:33</time>
<element>off</element>
<time_off>59:43</time_off>
</max31855>
The data above shows the raw celcius temperature, the 14 bit value received from the MAX31855. The temp shows that value converted to human readable base 10. The time shows approximately how long the system has been running while element shows if the tubes are currently active. The time_off section shows the last timestamp when the upper threshold temperature was reached and the tubes were last shut off.
For programming, the new Pololu AVR programmer was used, which was super nice since it has a built in UART (1 USB cable to the computer to do programming and test the XML) and it could be in-circuit programmed.
Vacuum Tubes
So first, maybe a quick explanation of what a vacuum tube is. Basically a vacuum tube is similar to a light bulb. They are made out of glass with the air removed. There is an element which in this case only creates heat (and an orange glow) rather than heat plus a bright light. Tubes have a metal plate which are usually connected to a high voltage potential so that eletrons that float away from the from heater wire can be attracted to the plate. Some tubes have a metal screen that another, smaller voltage can allow / deny access of the electrons to the plate.
For this project, two 6Z4 vacuum tubes with elements rated at 6.3v elements / 0.3A each (1.89W) are used. The actual voltage used here is 6.9v, so a little higher. The power supply is actually 7.5v, but 1N4001 diodes in series with the heater are used to drop the voltage down by 0.6V. The heaters are connected to IPI50CN10N MOSFETs so the microcontroller can turn the heaters on and off.
Pictures
Above is the setup of the first test with microwave brownie mix and cookie mix next to the thermos.
The raw cookie and brownie before being put in the oven.
Above is the result of sitting in the oven for an hour. The cookie was wanting to melt / cook, but the brownie mix didn't do anything.
Above is a picture of the vacuum tubes out of the thermos and powered on to show how they glow. These tubes cost around $10 for the pair, probably some of the cheapest new tubes around.
Source Code
https://github.com/mikeakohn/small_projects/tree/main/vacuum_tube_oven
Copyright 1997-2025 - Michael Kohn
