And here’s how that entire third Anet A8 catches on fire thing could have been prevented.
Side note, I started on this video when that last A8 caught on fire. Didn’t think the next one would follow so shortly after…
https://www.youtube.com/watch?v=VK_K6fp4BIk
Great work as always Tom!
One thing I will note is that I’ve had PPTC’s catch fire in certain cases.
The way a PPTC works is it uses a thermal compound that heats up with more current (PPTC = Polymetric Positive Temperature Coefficient). As it heats up the resistance of that compound increases so that the max current is reduced (see https://en.wikipedia.org/wiki/Resettable_fuse for a technical explanation of the process). It never truly breaks the current like a fuse though, as there is some leakage. As the compound cools down, it slowly restores the current till the heat drives the process back up. The time it takes for a PPTC to trip at certain currents varies based on the current tripping, the max rating, the ambient temp, whether the unit is cooled (passive vs active), and the voltage. The trip time is usually measured in seconds, which is way too long to be safe.
This internal process can create a hysteresis loop when there is a fault (not always useful), rather than a definitive “go/no go” situation like a fuse. In cases where the PPTC value is too low or the ambient temp is high, this can lead to regular false trips, leading people to “work around” things, which is just as bad.
If the current is high, but not a short, then it’s possible for the compound in the PPTC to heat so much that the PPTC explodes or catches fire. A friend of mine forgot to remove the PPTC on a RAMPS board and connected it to his large printer, which has 2x 12V PCB heaters for the heated bed in parallel, drawing about 23A (on a PPTC that trips at 11A). This was below the limit for his power supply, so it kept on producing current. The PPTC eventually tripped but the amount of energy (and therefore heat) pumped into the PPTC was so much that the PPTC’s internal temp went way higher than it can handle and it caught fire. FWIW: The 11A PPTC takes roughly 20 seconds of over-current to trip (even if it’s pretty much a dead short).
Fortunately for him, here was there. If he wasn’t, then he may have had a true fire on his hands. Also Re: active cooling: A fan would have simply added oxygen to the fire. Plus it’s quite possible that the PPTC could have set fire to the plastic in the fan, which would be more likely to spread.
All in all, I recommend people replace PPTC’s with fuses every time, and avoid solutions with PPTC’s like the plague unless they’re willing and able to replace them.
PS: Sorry for the rant/essay, but it’s something I’ve had issues with since RAMPS went to PTC’s from fuses so many aeons ago. 
One of the easiest methods is to switch to 24v. Many of these cheap printers still use 12v, which uses twice the current. Replace the heaters with 24v versions and get a 24v power supply.
Cable chains are great, but must have the proper bend radius. Sleeves typically fail at their mount points, where they repeatedly bend the most.
Many boards Don’t have the connectors soldered down well enough. Typically the via is too small and the solder didn’t flow to the top side of the board.
It’s only a matter of time until someone dies. How many times do you get lucky with “thank goodness I was there when it started” or “no one was around so no one was hurt” before there are kids asleep in the next room?
The number of people running fire-hazard printers overnight in their homes these days is going to cause fatalities eventually.
Anyone concerned with fire hazards should keep acetone a fair distance from heat sources.
@Stephanie_A Switching to 24V is a great idea, but the electronics and various parts (heated bed, hot end heaters, even motors) do need to be suitable. I’ve seen people simply put higher voltages into their printers without mods and that in itself can cause all manner of problems (and fires). Things like the capacitors on a driver board or (once again) PPTC’s have specific voltage ranges for operation, and the bottom end manufacturers put in the lowest value components that are suitable for 12V (ie: 16V caps) because they’re cheaper than the higher-voltage parts.
This actually reminded me of another thing: Connectors rated for 12A at 250V aren’t the same as connectors rated for 12A at 12V (or 24V). The lower voltage connectors usually have greater heat tolerance and/or lower contact resistance (just look at RC car connectors for example). Any connector resistance is usually MUCH less a percentage of the load resistance the higher the voltage (it’s why 24V is usually so much better than 12V), so the voltage split across that resistance is less, so they can sometimes heat up less (depends on the specifics, but usually the case). The lower the voltage, the lower you want the connector resistance to be.
@Nathan_Walkner that’s the sort of thing people can talk themselves into not being worried about, because the guy did some unusual stuff and only killed himself… filling a small room with hairspray fumes, storing low explosives under the printer… we saw tons of “that wouldn’t happen to me” responses here, and there was no good reason for lawyers or regulators to get involved at the end of the day. Likewise with the Anet A8 that burned up – it had been worked on recently and that credibly contributed to the fire.
Eventually we’re going to have a printer operated by the book that kills an innocent bystander. That’s going to be a much more impactful event than what we’ve seen so far.
We have an power strip that is wired with a relay through a Smoke detector. Alarm goes off, so do printers. Not a certain solution, but good. It is on thingiverse.