Satisfactory Aluminium: The Water Loop That Breaks Everything

Aluminium is where a lot of players hit a wall, and it is not because the recipes are complicated. It is because the chain hands water back to you at two separate stages, and water that has nowhere to go stops everything.

The chain, with real numbers

Step	Recipe	In	Out	Byproduct
1	Alumina Solution	120 Bauxite + 180 Water	120 Alumina Solution	50 Silica
2	Aluminum Scrap	240 Alumina Solution + 120 Coal	360 Scrap	120 Water
3	Aluminum Ingot	90 Scrap + 75 Silica	60 Ingot	none

All rates per minute. Refinery draws 30 MW, Foundry 16 MW.

Look at what steps 1 and 2 hand back. The refinery stage produces silica, and the scrap stage produces water. Neither is optional and neither can be left unconnected.

Why it stalls

A machine whose output cannot be placed stops producing entirely — including the product you actually want. So:

If the water output from Aluminum Scrap backs up, scrap production stops. No scrap means no ingots, and your whole aluminium line goes dark while the refineries upstream still look busy.

If the silica output from Alumina Solution backs up, alumina production stops. Same failure, one stage earlier.

The confusing part is that the thing that stopped is usually not the thing you are watching. You are looking at the Foundry wondering why it is idle; the actual problem is a full water pipe two buildings upstream.

Closing the water loop

The elegant solution, and the one the chain is designed around: feed the byproduct water back into step 1.

Step 1 consumes 180 water per 120 bauxite. Step 2 returns 120 water per 240 alumina solution. Run the numbers for your actual ratio and you will find a large share of your water requirement is covered by the loop itself, with fresh extraction only topping up the difference.

Two things make this work reliably rather than intermittently:

Put a Fluid Buffer in the loop. Returned water arrives in bursts as scrap batches complete. A buffer between the return line and the refinery input turns that into steady supply.

Feed the loop water in first, extraction second. Use a Priority Merger with the returned water on the priority input. That way fresh water only draws when the loop falls short, rather than the loop backing up while extractors keep pushing.

Without a priority arrangement, the loop and the extractor compete, and the loop is the one that stalls.

The silica surplus

Silica from step 1 is partly consumed by step 3, but the ratios do not cancel. You will accumulate silica.

Options, in order of usefulness:

Use it. Silica goes into Alclad Aluminum Sheet, Crystal Oscillators and several other mid-game parts. If you are building aluminium, you probably need those anyway.

Sink the surplus through an Overflow Valve so a full silica line never stalls the refinery.

Do not simply leave it unconnected. That stops step 1 entirely.

Alternate recipes worth having

Three change the chain materially:

Sloppy Alumina — 200 Bauxite + 200 Water gives 240 Alumina Solution, and crucially produces no silica. Double the alumina throughput per refinery and the silica problem disappears. The trade is more bauxite and more water in.

Electrode Aluminum Scrap — 180 Alumina + 60 Petroleum Coke gives 300 scrap and returns 105 water. If your oil setup already produces petroleum coke, this converts a byproduct you were managing into aluminium.

Pure Aluminum Ingot — 60 Scrap gives 30 Ingot with no silica input and only 4 MW against the standard recipe’s 16 MW. Lower throughput per machine, but it decouples ingots from the silica supply entirely.

The comparator on this site ranks each of these on raw ore per unit if you want to check a specific choice against your own constraints.

A layout that does not stall

1. Bauxite and water in at one end, with the returned water joining on a Priority Merger ahead of fresh extraction.
2. Refineries for Alumina Solution, silica out on a belt to storage with an overflow to a sink.
3. Refineries for Aluminum Scrap, water out into a buffer and back to step 1.
4. Foundries for ingots, drawing silica from the same storage as step 2 output.
5. Overflow valves on every byproduct line, so a full consumer never stalls the producer.

Build step 5 at the same time as everything else. Retrofitting overflow handling into a stalled aluminium plant means draining pipes to get at them.

Why aluminium is worth the trouble

It is reasonable to ask whether a chain this fiddly is worth building rather than working around, and the answer is that you do not get a choice for long.

Aluminium gates the second half of the game. Alclad Aluminum Sheet, Heat Sinks, Radio Control Units and the frames above them all trace back to it, which means every progression phase past a certain point wants aluminium in quantity.

The other reason is belts. Alclad Aluminum Sheet is what Mk.5 conveyor belts are made of. If your factory is bottlenecked on throughput and you want to move past 480/min per line, the route runs through aluminium.

So the practical framing is not whether to build it, but whether to build it properly the first time. A chain with overflow handling on every byproduct line takes maybe twenty minutes longer to lay out and saves the afternoon you would otherwise spend working out why a Foundry three buildings away has stopped.

Sizing from bauxite

Work forward from your node rather than backward from a target.

A Miner Mk.2 on a normal bauxite node gives 120/min. At the standard Alumina Solution recipe that supports one refinery exactly, producing 120 Alumina Solution and 50 Silica per minute. Two more refineries turn that alumina into 180 scrap, and 90 scrap per Foundry means two Foundries producing 120 ingots per minute.

That whole set draws roughly 120 MW before the water extraction, and it is a reasonable unit to build, test, and then repeat rather than trying to lay out a huge plant in one go.

The mistake worth avoiding is scaling the refineries without scaling the byproduct handling alongside them. Water and silica output grow at the same rate as your product, and a byproduct system sized for one module will stall three.

Common mistakes

Leaving the water output unconnected. Scrap production stops completely.

Leaving the silica output unconnected. Alumina production stops completely.

Merging returned water and fresh water without priority. The loop stalls while extractors keep pushing.

Diagnosing the Foundry when the Refinery is the problem. Trace forward from the stalled machine’s outputs, not backward from where you noticed.

Sizing water extraction for the full requirement. Much of it comes back through the loop; sizing for the total means constant surplus with nowhere to go.