[DIYbio] Re: Making my own incubator

 

One good failsafe is to use the laws of physics to eliminate any potential of runaway (I'm not licensed so I can say "eliminate any" instead of "eliminate most except in rare cases", hah).  Use a PTC in the supply.  Or as a cheap (non-approved) substitute, couple the heatsink of the voltage regulator to the incubator heater element, for example by using a machine screw with spacers so that normal operation doesn't affect the voltage regulator, while excessive heat is conducted through.  If the heater goes into runaway, the voltage regulator will shut down, which cuts off voltage to the software-controlled circuit.  With some rudimentary tweaking of the mechanical design, this would (should) provide a failsafe cutoff (about 150C) prior to flash point of other materials.  Also this could tend to brownout the supply near this threshold so both the digital side and software side need to protect against brownouts, which they should be doing anyway.   (For example, PIC chips have a brownout detection bit.)   This is the difference between DIY and professional equip.  Mostly polishing on the design, though very important.  

As for the "software error (never had one in the last years)" .. well there have been a few studies of mission critical software, for example NASA satellite firmware, and a significant yet small part of code on all projects studied found the software to be doing crazy things inside like attempting to write to read-only memory, running code from non-existent locations, wiggling pins that didn't exist, and so on....  Much like biology, software finds a way - it wants to be free!  :-D