Hi all,
Time for another update.
MOSFET issues: Like a good mystery story, it seems that there's been more than one MOSFET killer lurking in the Flux Deluxe's circuit board design.
I've spent the last 2.5 months focused on taming the electrical aspects of this heater, especially its susceptibility to voltage spikes. It took multiple tries, but as of last week I truly believe I turned the corner on the voltage spike issue. My current TVS diode solution continues to look good on the scope and work as expected. As far as I can tell, that issue is genuinely fixed. Huzzah!
So you can imagine the churn in my stomach yesterday when my test unit (with the lastest TVS diode fix) died mid-heat cycle, absolutely NOT due to any sort of voltage spike.
WTF??!? 
It turns out there's more than one way to kill a MOSFET. If you don't have any excess voltage at hand, excess heat will also do the trick. Thanks, physics.
All ZVS induction heaters are susceptible to this, and unfortunately I now see this is an area of design to which I should have paid more attention earlier in the project. Poor MOSFET heat management wasn't a show-stopper with the 16mm coil FDs, but the 14mm and 15mm coils pull more power through their MOSFETs than the original 16mm coil. I didn't realize how detrimental this could be until yesterday, when I finally saw a MOSFET fail on my bench. Which sucks, but yay! for seeing it happen, and being able to eliminate excess voltage as the cause. Knowledge doesn't come cheaply, sometimes, but that's how you learn stuff.
Unfortunately, this situation was hidden behind the voltage spike issue, so it took me by surprise.
The good news - and once again I may be squinting to see the upside - is that there are other, smarter people that manage hot electronic circuits for a living, and
they have proven strategies for mitigating this sort of thing. With a quick search I found the formulas for adding the right number of vent holes under my MOSFETs, as well as how to mate them to add'l surface area on both sides of the PCB, how to add more copper to the PCB surfaces for better heat exchange, etc.
I had done NONE of these things to this point, and as a consequence the Flux Deluxe's MOSFETs can only shed heat through their small footprint on the PCB.
I now know better.
Last night, I revised the FD's PCB to incorporate more surface area to vent heat, and more copper to facilitate better heat transfer. On both sides of the PCB. Previously, each MOSFET only had its 1cm^2 footprint for cooling.
Now each MOSFET has 20cm^2 (!) of surface area for cooling, 10cm^2 on each side of the board.
These changes should provide for better temperature regulation and allow the MOSFETs to stay below critical temperatures.
The new boards should be here in a week or so. I think they will be HUGE improvement over the current PCBs in this regard.
What does this mean for existing FD owners?
If your heater is working, that's great! Long may it run! 
If your heater has failed, please contact me to arrange the repair. Repairs are on hold until the new PCBs get here and test out OK, but a stable circuit will be worth the wait. I will know more once the new PCBs get here. It's always possible there's yet another big issue hiding behind these first two. I don't think there is, but this project has been sneaky that way, so we'll need to wait and see.
Finally, production remains ON HOLD. I am not making ANY new heaters until I get this sorted out.
I assure you, once I get these issues sorted, I will be happy to crank out as many of these as I can, as they will be good to go.
I will be putting together a post soon to show off the new features I'll be offering, which are incorporated into the new PCBs. They are more evolutionary than revolutionary, but they make the device more complete.
Sorry to post a cliff-hanger instead of a season finale, but that's how it goes sometimes.
Cheers,
)
I have some ideas, but unfortunately these things don't write log files that explain why they failed, etc. So my post-mortem analysis involves a mix of forensics and guesswork. 
