When I was young, I was into Heathkits in a big way. But as you all know, in the 80s and 90s the industry transitioned to SMD and, well, through-hole kit assembly stopped being even remotely cost effective (if it ever was) and I said "goodbye" to hardware hacking and became a software weenie.
I bought an electric car almost exactly a year ago, and stumbled across OpenEVSE. Long story short(er), that introduced me to Arduino, Eagle and OSHPark. Very soon I was back to designing my own stuff again - this time with proper tools and fabrication support. But I was still doing through-hole stuff.
Just to see what it would take, I decided to try a SMD version of one of my simpler things - an EV Simulator board (a very, very useful diagnostic tool if you're into designing and building EVSEs). I can only imagine what the DigiKey picker must have thought of someone buying single quantity 0805 resistors!
So I got the board and the parts and went to the hardware lab where I work with a microscope and it took me two hours to hand solder what I could have done in 20 minutes through-hole. I swore it off as a big waste of time, but happened to post about my experience on the OpenEVSE mailing list. In no uncertain terms, Chris (the main OpenEVSE developer guy) told me I was doing it wrong and that I should use reflow instead.
I quickly got a google education on the subject and saw that tons of folks had done their own reflow ovens. I looked at them all and... I dunno... I sort of found something different with each one that I didn't really like so much. Lots of folks used SSRs, but those are awfully pricey. An opto-isolated triac circuit can be done a lot cheaper. I also wanted to separate out the AC power control systems from the PID and thermocouple stuff, and actually build the former into the oven itself. It would wind up being a lot safer that way.
Skipping forward a bit, I call my result "Toast-R-Reflow." I went to Fry's (those not in the western U.S. may not be familiar with Fry's. It's a quite unique store. They sell components, computer parts accessories, appliances, office supplies... a quite eclectic mix) and semi-randomly picked a toaster oven. It turns out I made a not-too bad choice. It was easy to nondestructively disassemble it, while being "sporty" enough to keep up with at least a SnPb profile (which is good enough for me).
My power board is a pair of opto-isolated (MOC3020) BTA-20 triacs. By splitting the problem in two, I was able to simplify the thermal management. Each channel is thermally designed to switch up to 8 amps, which means the two combined (here in the U.S.) will do just under 2 kW. Plenty. Since there are two channels, this opens up the possibility of setting the oven to "broil" if you wanted to do double-sided reflow, for example. I haven't contemplated that because this particular oven struggles to follow the profile sometimes even with both elements in play.
The power board is through-hole - that solves the chicken-and-egg problem of building your first reflow oven. My first controller was an uno + breadboard with an AD595 thermocouple amplifier. At the time, AD8495 breakout boards weren't available. I now sell them and Adafruit sort of stole my thunder. :/ That controller's last job was to reflow its own replacement.
My latest controller is an ATTIny84 with a 2x16 LCD, an AD8495 thermocouple amp, a 1.8v LDO fed into AREF (to improve the resolution of the A/D), and a button. It has just enough firmware space for a dedicated-profile system. I have a design in mind for an ATMega + MAX31855, which conceivably could allow for a profile editor UI so that altering the set points wouldn't require an ISP.
Anyway, if there was a next step for me, it would be to try and either boost the capabilities of my current oven (perhaps by stuffing the inner walls with fiberglass) or replacing it with a beefier one so that I could use RoHS solder paste, which would open up Europe as a market for my stuff.
I've also wondered if there might not be a market for a
real hobbyist reflow oven as its own product. At its core, it's a double-walled sheet metal box with a pair (or four) heating elements and the control system I've already designed. I'd be surprised if I couldn't arrive at a retail price of $199.