There's probably a way to cut the 6th arm by turning one of them other arms into a multiarm that also does its job, but thats really difficult to do. Doing that would bring the min to around 170C, which does seem like the "vaguely reasonable min cost". Theres always the chance that 160C is possible if you cut the track but thats in the "completely nonsense but at the same time still theory possible min cost" tier. As for this solve, it's extremely annoying to tinker with so i dont want to bother improving it anymore. On a surface level, it seems like you have barely any breathing room, as both the lead is being merged into tin at input rate, and quicksilver is being purified at input rate. Thankfully (or unfortunately, depending on how much fun you had this puzzle) you can adjust when the hexarm rotates to leave metals being promoted for a shorter or longer period of time, and also store a small amount of lead on top of the glyph of purification (or projection, i keep forgetting which is which). If you pay attention to the solve, you'll see that the hexarm produces metals in the order of copper, gold, tin, silver, tin, gold, iron, silver, copper, iron - 2 of each metal, so the ratios work out. Alternating between high (gold/silver) and low (tin/iron) metals is essential for this layout so you dont have too many or not enough leads stored. Outputting tin is the hardest, as you need to output quicksilver to the side at the same time to slip the metal through without it getting promoted AND also have lead stored and ready in order for the hexarm to instantly grab a 2nd tin so you dont have an empty gripper (which would result in a quicksilver collision later). Gold is also difficult to make, as youre very likely to run out of spaaace to store lead while the hexarm is stationary for so long. However, with a little bit of trial and error this layout can work, even if my version of the solve is a bit jank (theres not much method to the order of the metals being produced). I'm expecting a lot of ties or semi-ties for rate, as it doesn't seem like theres a lot of wiggle room for rate solves without a wheel catrying all the metals, but I'll be glad to be proven wrong.