Inspired by topic about traps against trapavoids, who also happen to avoid pressure plates, I did some experiment on supports. Basic idea is to use magma to trigger the plate, which makes support tumble. Other idea is not to use plates, when not necessary. And in my layouts they are usually not necessary, because supports tend to be located next to the plates anyway. Supports are buildings, and as buildings they are as resistant to magma as much as they components are. So I checked time it takes to bring supports down and make a cavein. My setup was as this:
These are basically "landmines", which are built in my obsidian caster. Whitish tiles next to the ramps are recreated marble floors, remnants of walls used for building traps. Floors to be caved are made of dolomite. Supports are made from
nether-cap, diorite, wood (acacia), zinc,
dolomite, copper,
nickel and tin. Bolded ones are supposed to survive magma, all others - not so much.
I dropped magma and measured time needed to collapse the support, and to collapse the cave-in which usually happens shortly afterward - one to two frames usually. But tha game doesn't pause for a moment, so we usually see a state of matters a few ticks after the cavein was triggered. I was curious which materials can be used as reasonable timers.
After first couple of runs I changed setup to involve more supports of the same material, and to involve other materials, potentially better materials, namely pewters and lead. Due to some light randomness the exact time of survival differed amongst runs, but it was always very close, and the sequence of materials was generally . So for zinc it could be 182 frames or 187 frames, but not 195, for example, and lead was always before pewters.
Measured times, from magma contact to collapse, were as follows:
lead 45
tin 47
pewter 48
zinc 185
wood 335 (wood starts smoking around 100 ticks from contact, it really hinders observations…)
copper approx. 1300-1400
diorite approx. 1700-1800
Due to expected errors it is possible that all first metals there would melt somewhere between 40 and 50 ticks, though most likely around 45, the same error is for zinc, and for wood it is around 10. Other values are approximate, due to way they were calculated (from video), though the ranges shouldn't be off for more than 10%. Yes, melting diorite or even copper takes more than one in-game day.
As seen the sequence of melting follows two properties: melting temperature and specific heat capacity (called "specific heat" in RAWs). The lower the melting point, the faster the object melts. The lower the specific heat capacity, the faster it gets hotter and eventually melts. The best short fuses are apparently lead, tin and fine pewter. I like zinc too. The sequence follows a sequence of destroying cages in magma - goblins in wooden cages tend to outlive even copper ones (though it isn't a comfortable life) while zinc cages almost explode shortly after submerging. These supports are "dumb" because there's no logic nor sensors, only basic material properties involved.
How good they are in practice? For my traps they are lethally good, not least because I use shaped cave-ins, which kills everything that happens to be in area. Building destroyers usually stand one tile from a door or floodgate, often directly under the support (supports are passable). When using tin or lead, they have around 45 ticks to run, which, with Agility=2000 would mean about 6-8 tiles, and for the average troll it would be around 4-5 tiles. The most agile denizen of Underworld on my map has agility=2000, though even maximum agility doesn't give much more headroom, plus they like to crowd much, which makes them slower. For comparison, time taken from triggering upright spike trap to make it move is 40, and time from triggering a bridge/floodgate to open it is 100. Supports aren't, but are fast enough for many uses. In fact it may be wiser to use zinc, or maybe some alloy or wood, and give enemies more time to fill the corridors-to-be-they-tomb.
Magma tumbled supports have one disadvantage, which is need for magma to stay at they location. With short time "fuses" that's not an issue, but longer time (like stones) would probably need additional measures. Like setting them in a 3x3 trough with ramps, to collect magma, and making them two levels high (or three levels in shaped caveins). To install one support atop another it is required to make a floor on the first one. Or simply putting a constant source of magma, maybe through a diagonal depressurizer, then it tends to stay for a long time. I use them mainly because my mechanics are really busy in other parts of the fortress, but also because they are means of disposal of unwanted metal...