Embedded power

I'm looking to create a solder reflow oven similar to the one created on the Ben Heck show. I want to take the AT MEGA chip and put it on it's own circuit board. I know I'll need a crystal, but I'm not sure what else is needed exactly. What do I need to take 120v off the power supply cable to the toaster oven and get it to 5V DC for the arduino? I know I could pull apart a wall wart, but that seems a bit messy. Perhaps that is the best option though. Thought I would get everyone's take on that before starting.

"similar to the one created on the Ben Heck show"
Got a link?

Reflow Oven controller Reflow Toaster Controller - WIG-00081 - SparkFun Electronics

5Vwallwart Switching DC Wall Power Adapter 5V / 1A - dipmicro electronics

If you don't want to take a wall wart apart, you could use something like this http://www.newark.com/xp-power/ece05us05/power-supply-ac-dc-5v-1a-5w/dp/89T3513.

Would this work?
https://www.jameco.com/webapp/wcs/stores/servlet/ProductDisplay?storeId=10001&langId=-1&catalogId=10001&productId=1954906

Do I also need a linear regulator between this and the ATMEGA?

You can use that to power the atmega directly. The datasheet doesn't say it has a minimum load, however I suggest you load it with at least 40mA, because switching regulators don't always behave nicely when unloaded. If your project includes an LCD display, the backlight will provide this load.

That PSU has a specified output ripple of 80mV p-p, which is likely to cause some jitter on analog readings if you use the 5v supply as the analog reference. You might want to use a 3.3v linear regulator to provide a more stable analog reference.

aurelius:
I'm looking to create a solder reflow oven similar to the one created on the Ben Heck show. I want to take the AT MEGA chip and put it on it's own circuit board. I know I'll need a crystal, but I'm not sure what else is needed exactly. What do I need to take 120v off the power supply cable to the toaster oven and get it to 5V DC for the arduino? I know I could pull apart a wall wart, but that seems a bit messy. Perhaps that is the best option though. Thought I would get everyone's take on that before starting.

May I suggest a different tack than going with the Sparkfun controller? For one, no control code is provided. Secondly, it runs on a PIC microprocessor. Third, the folk at Sparkfun have so much confidence re: the American legal system that the unit is only sold as kit.

My suggestion would be to wait for the folk at Rocketscream to make their Atmel-based controller available again. It may take a while to arrive, but the ease of use is great because they use the Arduino IDE to program the thing. I had my first experience with forking a open-source project when I modestly upgraded the code to include a switch between leaded and lead-free solder profiles, giving a abort message when the initial temperature is too high, and making use of the fan relay option.

So what I think makes this unit great is the hardware is based on the Arduino form factor (it's a shield that you can use with the Uno and similarly-sized units), it's fully assembled, it has two built-in switches, it uses a K-thermocouple (easy to source) and it leverages the great work by the Adafuit team (thermocouple support) as well as Brett Beauregards PID library. Simpletons like me can then easily make small improvements as described above.

Rather than follow in my footsteps as far as ripping out the existing control hardware and installing a Crydom D2425D SSR inside the toaster oven, a better option may be to buy a fully-assembled PowerSwitch Tail and ignore the fan setting. Should take a lot less time and less dangerous to those new to the business of line-voltage electronics. One thing you will likely have to do though is to insulate the inside of the oven cavity to allow heating rates that are compatible with the solder profiles published by manufacturers.

For those that also want to install the SSR on the inside, go over the circuit before ordering one. My circuit included a fat diode to reduce the duty cycle to 50%. If your circuit also has a similar diode, use a SCR-based SSR, not a Triac. On the other hand, if no diode is present, a Triac-based SSR will give you more output. As it were, the SCR-based toaster is running at 1300W when input is 100% and the heat sink on the back of the SSR is more than a good idea. I simply sandwiched the exterior casing of the toaster oven between the SSR and the heat sink, works great.

One other tip is the use of the great pre-crimped wires that Pololu sells, along with a plastic housing that the crimped connections click into. Makes for a very clean install, especially if you heat-shrink the wire harness after making the connector.

I am looking to take the chip from the Arduino and pop it into a custom setup that I create. I figure that if I'm going to go, I might as well go all out. Buying a shield seems to be cheating. Finding out what it takes to put together an embedded solution that I can swap out with the toaster oven's circuit board I am sure will teach me a lot, and be a hell of a lot more interesting than plugging together two things that "just work". There is a .ide on github.

In that case you may find Prototyping small embedded projects with Arduino | David Crocker's Solutions blog useful.

aurelius:
I am looking to take the chip from the Arduino and pop it into a custom setup that I create. I figure that if I'm going to go, I might as well go all out. Buying a shield seems to be cheating. Finding out what it takes to put together an embedded solution that I can swap out with the toaster oven's circuit board I am sure will teach me a lot, and be a hell of a lot more interesting than plugging together two things that "just work". There is a .ide on github.

The beauty of schematics, etc. being available at rocketscream (among other places) is the ease with which you can then roll your own solution, being able to stand on the shoulders of the giants that brought you the hardware design, the K-thermocouple library, the PID control software, etc. Or do you want to write that too?

Hence, I disagree with buying a shield == "cheating" . I see it as supporting the folk who put together a working solution that gets me faster to what I want to do (i.e. design and manufacture my own boards) than working on developing that solution first. Comes down to how you want to spend your time, I guess.

I finished the install today and tried to make it look pretty using some silver spray and bondo to cover control panel holes. The Arduino and the shield sit inside a plastic case I salvaged from inside my Neuton lawn mower. It's fairly thick plastic, protecting the two boards from bumps and bruises while keeping the switches and terminals accessible on the outside of the oven. The only cable running into the high voltage part are the control wires for the SSR and they're insulated with additional heat shrink and high-temp socks.

As you can see, the oven cavity is partially filled with 2000F rated insulation to reduce the mass of the oven being heated / speeding up the PCB heat transfer. Also insulated the inside of the oven between the cavity and the outer surface with same. On the backside, a large heat sink keeps the SSR cool. I have a small PCB attached to the tray with the thermocouple woven through it. I have some rigid insulation on order for the roof of the cavity, the door, and the side of the cavity with the convection fan. Once that insulation is in place, the oven is in the best position possible to reach difficult performance goals such as raising the interior temperature at 3C/second for some parts of the reflow cycle.

Last but not least, I only found three analog switches inside the oven, nothing more. No diagram, PCB, or anything else that would help decipher the circuit. Interestingly, the lamps, the fan, etc. are designed to be "hot" all the time, the controls were all on the Neutral side of the circuit. Not sure why, my guess is that it's cheaper.

Cool build Constantin. Where did you get the insulation from?

In drawing up the schematic, if I'm getting 5v out from the encapsulated power supply, would adding a 7805 regulator create a stable enough analog reference? I'm thinking that if 3.3 is better for the analog reference, I'll go with a LM1117T. Anytime someone says "jittery", I think of my first car and that was a nightmare.

What I've got so far is PSU>10 uf electrolytic>7805>10 uf electrolytic>10 uf tantalum>LM1117>10 uf tantalum
Does that look correct? For the extra $1.50 or so, is the 7805 setup going to do me anything?

If anybody knows of a better 5v psu that might provide me with a better bang for the buck and less mess, I'm all ears.

Depending on the actual load current there are several "Better" methods for an Accurate 5V reference. I think this is where your Education will truly begin. This is NOT a low power design so I might recommend a TIL 431 which is an "Adjustable zener diode, it takes a pot and a resistor to make a Stable any voltage reference from 1.25 to 30+ volts and is better as a reference than a 7805. On the subject of 7805's I personally remove them any time I can use a switcher instead, If noise is an issue several caps and a resistor or small coil w\ill deal with that Easily. (I do own an analog scope a Tektronics 2213) and finding and dealing with noise isn't a big deal for me. The TIL 431 is the simplest answer to your reference issues, simple because it is a variable shunt regulator and although there is always the pink noise issues with Zeners this device has it's Zener buried well enough that any generated noise can be dealt with with a .1uF and a 10uF cap.

Doc

Not sure a stable 5V supply is actually needed for this project.

Looking over the screaming rocket shield, I see a dedicated chip for the K-thermocouple, whose use makes sense because the thermocouple head has very little mass and is very temperature resistant. You can get pre-configured K-style thermocouples with fiberglass insulation @ $9 for 2 at a time on e-bay. Makes sense to me.

As for controlling the SSR, I suppose just about any external transistor will do, some may even be able to be driven off the Arduino itself. Depends on the current draw, see the spec sheet. My personal preference for this task would have been an opto-isolated darlington transistor array with a isolated power supply. Just in case the SSR-control wires ever rub / melt / whatever and could get shorted on the high side. Adds cost, but also adds a little bit of fool-proofness. A nice touch would be a fuse and removable darlington chip (think ULN2003A in a DIP form factor)

The rocket scream display is functional though I may modify the messages a bit. A larger display (like the Nokia 5110 series, for example) could be nifty from the point of view of being able to show more relevant data. For example, besides announcing the state of the process (i.e. soak, reflow, etc.) show the temperature, rate of change, how well the oven is doing vs. the manufacturer spec sheet, etc. This can be achieved on the 8x2 display to a limited extent but it requires frequent display wipes.

As for the insulation, McMaster Carr is your friend for all types. I only use insulation that is rated well above the intended temperatures. Today, I hope to glue the semi-rigid stuff to the roof of the oven and the door (leaving a slit for observation) with furnace cement / adhesive to minimize the heat transfer and maximize the homogeneity of the temperatures inside the oven. Yesterdays trial board came out great, though I had to move it closer to the center of the oven during reflow as the edge by the oven door was not reflowing in tune with the rest of the board.

This also marks my transition to lead-free solder. The Kester 96/3.5/0.5 lead-free paste I have works much better than the leaded stuff I had been using. Even tricky chips like the ADE7753 with its 0.5 or 0.65 mm pitch came out clean with no bridges. Also ordered lead-free solder wire...