Factorio Calculator — Belt Throughput, Machine Counts and Furnace Columns

The machine count formula, and why rounding up matters

Every production question in Factorio reduces to one equation:

machines = ceil( target × craftTime ÷ (60 × machineSpeed × itemsPerCraft) )

The craft time is the recipe's own time, straight out of the game. Machine speed is the multiplier of the building doing the work, so a 2.0-speed furnace finishes a 3.2-second recipe in 1.6 seconds. Items per craft handles recipes that output more than one item at a time.

Two mistakes are worth calling out because the calculator exists to prevent them. First, do not scale the craft time and the machine speed together — a steel furnace with speed 2.0 on a 3.2-second recipe produces one item every 1.6 seconds, but if you enter 1.6 as the craft time and 2.0 as the speed you have counted the speedup twice and will under-build by half. Second, rounding up always leaves you over target, and the calculator prints by how much so you know how much buffer the line carries.

Belt throughput: the real ceiling

A fully compressed belt carries a fixed number of items per second, and no amount of machines changes it:

• Transport belt (yellow) — 15 items/s
• Fast transport belt (red) — 30 items/s
• Express transport belt (blue) — 45 items/s
• Turbo transport belt — 60 items/s

Those numbers come from speed × density. A yellow belt moves at 1.875 tiles per second and a straight belt tile holds 8 items, so 1.875 × 8 = 15 items/s. Red doubles the speed and blue triples it, which is why the capacities are 30 and 45 rather than rounder numbers.

A belt that looks busy can still be underperforming: a belt with one full lane and one empty lane carries half its nominal throughput. When a line starves despite a correct machine count, check lane balance before adding machines.

Smelting columns, the one ratio worth memorising

Iron and copper ore take 3.2 seconds to smelt. A stone furnace runs at speed 1.0, so it produces 18.75 ore per minute and a full yellow belt of 900 per minute needs 48 stone furnaces. Steel and electric furnaces run at speed 2.0, halving that to 24.

Because a smelting column feeds furnaces from both sides, the practical build numbers are 24 furnaces per side for stone and 12 per side for steel. Scale the count with the belt tier: a red belt carries twice as much, so it needs twice as many furnaces — 96 stone or 48 steel — and a blue belt needs 144 or 72.

Three worked examples

Filling a yellow belt with green circuits

Electronic circuits take 0.5 seconds and output 1. At an Assembling Machine 2 speed of 0.75 you get 90 items per minute per machine, so 900 per minute needs 10 machines. That is exactly one yellow belt, with no headroom — the lesson being that the belt, not the machines, is the limiting factor at that scale.

Filling a blue belt with the same recipe

At 45 items/s a blue belt carries 2,700 per minute. At 1.25 assembly speed each machine makes 150 per minute, so the belt needs 18 machines. If you only have 12 machines you are leaving 900 items/min of belt capacity unused; if you have 24 you are producing more than the belt can move.

A smelting column for a red belt

A red belt carries 1,800 ore per minute. At 18.75 per stone furnace that is 96 furnaces, or 48 per side. Switching to steel furnaces halves it to 48 total, 24 per side — the same footprint doing twice the work.

Why belts beat machine counts in the endgame

The formula above assumes the ingredients can physically arrive. In a large base they usually cannot: several machine blocks compete for the same belt, splitters divide rather than multiply, and a saturated belt feeding a splitter leaves branches underfed no matter how many machines sit on them.

So the order of checks is always the same: work out the machine count, work out the belt demand, compare the demand against the belt tier, and only then decide whether to split the load across belts, move intermediate production closer, or upgrade the tier. The calculator reports all three figures at once so the decision is made with numbers rather than by building and watching.

Sources and standards

FAQ