Safety of Arduino self-bondage timer?

Post your thoughts and ideas on safety here.
Username
*
Posts: 7
Joined: 11 Sep 2021, 10:36

Re: Safety of Arduino self-bondage timer?

Post by Username »

I`ll use an industrial timer (battery powered) and just an simple connector close to the magnet. One pull on the cable and the magnet lets loose.
User avatar
ZipMeUp
*
Posts: 36
Joined: 11 May 2020, 02:15

Re: Safety of Arduino self-bondage timer?

Post by ZipMeUp »

As someone who has used MOSFETs quite a bit in various personal projects there are a few sneaky things that can creep up and cause MOSFETs to fail... and by that I mean fail shorted.

First and foremost, I'm quite worried about the watchdog circuit that has a RC time delay connected to a MOSFET gate. Modern power MOSFETs are composed of a very large number of smaller FETs in parallel, and there is bound (no pun intended) to be some variation in parameters across the die. If the gate voltage switches too slowly, some of the sub-FETs can turn off before others and the last ones to turn off will get stressed. This will slowly lower their threshold voltage over many cycles, making them take this stress for longer. Eventually the FET will fail short, leaving you :gag: until the battery dies. I have actually had this happen to me in my early days of playing with FETs. I thought it would be cool to use a FET to dim a 12v bulb as it turned on and off... until the bulb just stayed on. To prevent this, use a comparator with hysteresis before the FET to make sure that the FET turns on and off quickly. Add a flyback diode to prevent any inductive spikes.

Static shock while assembling the board can cause the FETs to fail outright, or even worse, cause the FET to be weakened to the point where it fails later. This isn't usually a problem with big power FETs, but with small FETs rated for an amp or so it can happen pretty easy.

Short circuits will cause the FETs to go shorted unless it's bad enough to blow the FET to bits. Having some type of e-fuse as discussed earlier would be a great way to stop this. I would do this with a shunt and comparator to trip a latch but a purpose-built E-fuse chip would do the same thing. Or just get massivly overrated FETs, ones rated for 50A at 30v cost like 50c each.

Finally, a crowbar circuit would be a cheap and reliable final fail-safe. If the board senses that an output is still on after telling it to go off, turning on a SCR across the main power rails to blow a fuse would be better than getting stuck.

I really like this idea of a standardized board to control electronic bondage rigs. It would even make a great power control shield for vanilla uses as well.
It it ain't inescapable... It ain't my kind of bondage.
Post Reply