Hot-end melt-downs are a problem. Last year, I was experimenting with a thermally fused hot-end. While it worked great and I really liked the simplicity, the proper thermal fuses are not available for this application. Recently, on the #reprap IRC, there was a discussion about using two thermistors and having the firmware compare them and shutdown the hot-end if they didn't match.
Afterwards, I contacted a couple firmware developers and one of the Marlin developers, "EvdZ", was kind enough to work with me to test out this concept. As I only had Sanguinololu boards available, I picked up a RAMPS board, from Ultimachine, and tested the concept with just a couple thermistors. The thermistors were at room temperature and, as a test, I put one of the thermistors between my thumb and forefinger to see if the firmware would shutdown the heater when the thermistors no longer agreed on the temperature. After a little debugging, I sent my findings back to EvdZ and he made the proper adjustments to the code. Next up was a static test of a proper hot-end.
Tonight, I setup a test for a dual thermistor J-Head hot-end. The thermistors were not properly secured as I wanted to be able to pull either one out for testing. In addition, this hot-end has a heater cartridge and a set-screw/jam screw in order to hold the heater cartridge in place. The thermistor holes are drilled at 90 degrees to each-other in an attempt to ensure that if one thermistor were to fall out, the other thermistor would not be pulled out with it. (Creating a situation where both thermistors would still agree on the temperature.)
The test worked flawlessly. I heated the hot-end up to 230 degrees. Upon removing either thermistor, the heater cartridge was shut off in under 2 seconds. I repeated the test a couple times to ensure that the result was always the same.
Once the dual thermistor changes are part of Marlin, I'll post the complete instructions for the dual thermistor configuration.
Of course, in the near future, the hot-end needs to be tested printing. As the only differences, from a J-Head Mk V-BV, are the thermistor and heater cartridge hole configurations, I don't foresee any issues.