Thanks for your extensive work here! I think there are still some conflicts between the various angel's mods and marathon regarding ore yield. I compared with and without Marathon (and with and without this mod) to document them. Sorry for the very long post.
tl;dr some ore -> plate interactions are missing and some are strange, leading to unbalanced gameplay. I have some suggestions to model this mod closely after Angel's, plus Marathon's plate multiplier.
If I understand correctly, Marathon's biggest change is that iron ore yields 2.5 times more plates and copper ore yields 5 times more plates (and costs are adjusted even more than that, so many more plates are floating around).
So, I think the goal of this mod (correct me if I'm wrong) is primarily just to modify Angel's stuff so that iron production increases by 2.5, and copper production by 5, in the same way; and, also to increase structure costs similar to marathon. There's a lot going on in the mod I don't understand (why change times for ore crushing? etc.) but that's what I imagine is the ultimate goal.
In Angel's Refining without Marathon, ores follow this pattern:
(Mechanical) crushing takes 0.5 time units per ore and produces 0.5 stone per ore.
Raw ore turns into plates at 0.67 efficiency (3:2 ore to plates)
Sorting raw ore gives primary (iron for saph, copper for stir) at 4:2 efficiency and secondary at 4:1, so overall efficiency is 0.75. A modest increase, and you have to deal with extra hassle -- not an enormous leap in efficiency by any means. You also get slag, which is a detriment early on but a modest efficiency boost later.
Now for smelting:
Smelting for iron takes one coke and turns 4 iron ore into 5 plates (eventually). It only works on sorted ore. You could call this a 25% efficiency boost, but it costs coke which means you only break even, though you can do cool stuff with the intermediate products later on.
Smelting for copper is a break-even prospect, five ore to five plates.
Finally, processing copper gives you a 25% boost, 4:5. Processing iron gives you a similar boost, 4:6 but costing one coke (I'd call it a ~20% boost, then)
All of this means that the benefit you get from going into a full refining supply chain is modest and the effort is substantial: it makes sense for huge operations but it's not essential.
This Bob's Angelthon mod as of now has the following changes:
Ore crushing is much faster and produces much less stone. I'm unclear why; if there's not a good reason, maybe this should be removed? It causes a surprising degree of stone scarcity since stone can't be mined directly!
Direct smelting of saph and stir to their primary metals is unchanged from Angel's vanilla at 0.67. (Costs are still modified way up because of Marathon, so this is an enormous penalty for direct smelting).
Sorting ore still gives you the same ratio from saph and stir to iron and copper + slag, 4:2:1:1. The efficiency from ore to ore is the same as in Angel's, except that it happens faster (why?)
However, the sorted ore is affected by Marathon and produces way more plates. So instead of an increase of 0.67 ore->plate efficiency to a modest 0.75, it becomes an increase of 0.67 to 1.88 (for iron) and 0.67 to 3.75 (for copper). The post-sorting productivity is consistent with marathon; the problem is that the pre-sorting productivity is not.
Advanced iron smelting is also a bit strange. If you consider coke and iron ore equivalent in cost, then this is break-even in Angel's w/o Marathon. But with this mod, now it's a negative: it costs 6 iron and 3 coke (9 total) for 18 plates, which means the 2.5 ratio for iron ore falls to 2.
Advanced copper smelting: 3 ore to 18 plates, so a boost in productivity (normally would be 3:15) where Angel's was break even.
Lastly, steel smelting via molten iron is also out of whack -- it costs half of what it should (because Marathon doubled steel cost in terms of raw materials) and probably could stand to have its energy/time increased a lot too.
My suggestion is that the mod be brought in line with the Angel mod ratios, but multiplied by 2.5 or 5 for iron and copper plates respectively. AFAIK you have to end up with plates to craft into anything useful, so I think the best thing would be to scrap most of the intermediate iron/copper/steel pipeline changes (to how much ore is crushed at a time, to ingot crafting, etc.) and ONLY modify the steps that produce iron or copper plates: the mod will be much more manageable in terms of complexity and will have ratios that match Angel's. So, for example, iron ingots would produce 2:5 ingots to plates instead of 2:2, and unsorted ore would be 0.67 * 2.5 or 0.67 * 5 for iron and copper respectively.
