Dyson Sphere Program: The Ultimate Guide to Ratios, Math, and Megabase Scaling

Introduction

Mastering Dyson Sphere Program isn't just about building factories; it's about understanding the brutal math that governs every belt, smelter, and assembler. From the deceptive speed of your machines to the explosive scaling of late-game science, your entire production empire lives or dies by its ratios. This guide breaks down the core principles and exact numbers you need to build efficiently, from your first smelter array to a universe-spanning megabase.

Core Production Principles & Belt Math

Belt Throughput Fundamentals

Here's the harsh reality about belts in DSP - they're not just pipes; they're your absolute ceiling. Every production line you build lives or dies by three numbers, so memorize them now:

Belt Tier	Speed (items/s)	Speed (items/min)
Yellow (Mk.I)	6	360
Blue (Mk.II)	12	720
Green (Mk.III)	30	1800

That green belt number looks delicious, but it comes with a catch: if you've got a single yellow belt bottleneck anywhere upstream, your entire factory slows to yellow speeds. This means belt capacity determines your whole scaling strategy - you're not just planning factories, you're planning belt highways where the slowest lane wins.

Machine Speed Multipliers

Your assemblers are lying to you. Well, not exactly, but the tooltip numbers don't tell the full story, and that's where ratios start falling apart.

• Mk.I Assemblers run at 0.75× base speed, so they're actually 25% slower than advertised.
• Mk.II Assemblers run at 1.00×, making them the true baseline.
• Mk.III Assemblers crank it to 1.50×, which is 50% faster than the tooltip.

But then you add Proliferator spray, and everything changes. Mk.III Proliferator in production-speed mode doubles your entire recipe speed at 2×, which means a Mk.III assembler with Mk.III spray hits 3× real speed (1.5 × 2). This is where the math gets wild - your old ratios completely shatter, and you'll need way fewer machines than you think.

The 12-Smelter Rule

This is the golden ratio that early-game players swear by: 12 Arc Smelters per Yellow belt. Here's why it works and how it scales.

Step 1: The Base Math Each Arc Smelter churns out 30 items per minute. Twelve of them gives you 360 items/min, which perfectly saturates a Yellow belt. Simple, right?

Step 2: The Hidden Catch Mk.I Sorters move at only 1.5 items per second, and they create micro-gaps that can starve your belt. Twelve smelters is the magic number that hedges against this sorter delay - you're essentially building in buffer capacity.

Step 3: Power Cost Those twelve smelters will cost you 360 kW continuously (30 kW each), so don't forget to pad your power grid accordingly.

Step 4: Proliferator Speed If you spray with Proliferator Mk.I at 150% speed, the raw math says you only need 2.4 smelters in theory. In practice, you'll want 4–5 smelters to keep the belt full and compensate for sorter quirks.

Step 5: Scaling to Blue Belts When you upgrade to Blue belts (1800 items/min), you're looking at 48 Plane Smelters with proliferator spray - that's 12 smelters per blue line across four parallel lines. The rule scales, but your smelter count explodes.

Smelting Ratios (Ore to Ingots)

Alright, let's talk smelting math before your factory turns into a beautiful, inefficient mess. The ratios in DSP look weird at first glance, but they're actually consistent once you see the pattern - most recipes fall into three camps: 1:1 simple conversions, 2:1 compressions, and that weird outlier we call steel.

1:1 Conversion Recipes

These are your bread-and-butter smelting recipes where one unit of input becomes one unit of output, but don't let that fool you - timing matters way more than you'd think.

Iron and Copper Ingots both take 2 seconds per craft, which means a single smelter only spits out 30 per minute, so you'll need two of them running side-by-side to fully saturate a basic blue belt at 60/min.

Stone Bricks are the speed demons here, cranking out one brick per second for 60/min on a single smelter - if you need walls fast, this is your guy.

Titanium Ingots look trickier on paper since it's 2 Titanium Ore → 1 Ingot, but the timing works out to the same 30/min throughput as iron, so we still call it 1:1 in practice.

Then you've got your exotic materials: Kimberlite → Diamond and Fractal Silicon → Crystal Silicon both crawl along at 4 seconds per craft, giving you just 15/min per smelter, which means you'll want to double up on these bad boys if you're building processors.

2:1 Conversion Recipes

Here is where the math gets spicy. These recipes chew through two inputs to make one output, which means you need way more smelters to keep pace.

Magnets are everyone's first headache: 2 Iron Ore → 1 Magnet in 3 seconds per smelter, which comes out to 20/min, so to hit 60/min, you need three smelters running full tilt.

Assembler Production Chains

Early Game Components (Red/Blue Science)

When you're first bootstrapping red and blue science, your entire factory hinges on three components that look simple but will eat your lunch if you don't plan ahead. Magnetic coils are the first real test - one Mk-I assembler can pump out 60 coils per minute, but it needs 60 magnets and 30 copper ingots to do it, which means you're running two smelters on overtime just to keep up. Circuit boards need their own assembler at the same 60/min rate, and they're even thirstier for iron: 60 ingots plus 30 copper ingots will drain your patches faster than you think. Gears are the chill cousin here - one assembler makes 60 per minute while only asking for 60 iron ingots from a single smelter, so at least something's easy. For a perfect 60/min science setup, you're looking at three assemblers and four smelters total, which sounds manageable until your starter ore field disappears in the first two hours.

Mid-Game Electronics (Yellow/Purple Science)

Yellow and purple science are where the training wheels come off, and your production lines get real. Processors become your new bottleneck - two Mk-III assemblers can hit 60 per minute, but only because the craft time is so slow that you'll swear they're moving through molasses. Those two assemblers need 60 circuit boards and 60 microcrystalline components fed in constantly, which means your early-game setups have to be bulletproof or everything stalls. Microcrystalline components are their own special hell: two Mk-I assemblers running at 2 seconds per craft just barely hit 60/min, and they're screaming for 60 high-purity silicon ingots from two dedicated smelters. Plasma exciters are just as sluggish, with two Mk-I assemblers at 2-second crafts demanding 60 magnetic coils and 60 plasma per minute - that plasma alone needs four refineries running flat-out. So mid-game isn't just 'build more assemblers,' it's 'how do I keep these things fed without tearing my hair out.'

Late Game Advanced Components

Late-game components are where DSP stops being polite and starts getting real. Titanium alloy is your first wake-up call: three Mk-I assemblers running at a brutal 12-second craft time can only squeeze out 60/min, and they're guzzling 180 titanium ingots, 180 steel, and - this is not a typo - 360 sulfuric acid per minute. Plane filters are the real monster, though, demanding eight Mk-III assemblers running at 6 seconds per craft just to hit 60/min, and they need 120 processors, 120 particle broadbands, and 180 graphene fed in constantly. Frame materials are almost a relief by comparison: four Mk-I assemblers at 3-second crafts need 120 titanium alloy and 60 high-purity silicon, which feels reasonable only because plane filters exist. So yeah, if you're serious about white science, you better have a whole planet cleared and ready because these numbers aren't joking around.

Chemical Plant & Oil Refinery Ratios

Oil Processing Fundamentals

Here's how you turn crude oil into actual progress. The community has been obsessed with finding the 'perfect' oil setup, and they've landed on a 6:8 pairing - six Plasma Refineries feeding eight X-Ray Crackers. In theory, this should be clean, but the math gets a little weird depending on your proliferators.

Each Plasma Refinery runs 120 crude oil per minute through plasma refining, which spits out 90 hydrogen and 30 refined oil. Six of them together generate 180 refined oil/min and a baseline of 90 hydrogen/min, though that hydrogen number jumps to 210-300/min if you're proliferating your inputs.

Now those eight X-Ray Crackers are hungry. They devour 240 refined oil/min and 240 hydrogen/min to produce 360 hydrogen/min and 120 energetic graphite/min. You can already see the problem - your six refineries don't make enough refined oil. You'll need to pull in 60 extra refined oil/min from an external source just to keep everything fed.

The weird part? Despite burning through hydrogen, this setup is still hydrogen-positive in practice, meaning you'll have a surplus to siphon off for other uses.

Red & Blue Matrix (Early Game)

Getting your first research cubes running is simpler than it looks. For 60 red matrices per minute, you only need 2 Matrix Labs chugging away, but the real work happens upstream. You'll need 3 Oil Refineries (one normal, two running X-ray cracking) to squeeze out 150 hydrogen per minute, plus 2 Graphite Smelters burning coal. That whole setup pulls around 3.3 MW, which isn't terrible for early game, but you'll want to belt-feed it 120 graphite and 120 hydrogen continuously or the labs will stall.

Blue science is where you'll first juggle multiple assemblers. 60 blue matrices per minute also needs 2 Matrix Labs, but now you're building a mini-factory: 1 Assembler for Magnetic Coils, 1 Assembler for Circuit Boards, 4 Smelters splitting duty between iron and copper, 2 more Smelters dedicated to magnets, and don't forget the Water Pump for those circuit boards. The coil line alone eats 120 iron ingots per minute and 60 magnets per minute, while circuit boards need 60 iron ingots and 30 copper ingots. It's a lot of belts, but at least everything's still on your starting planet.

Yellow & Purple Matrix (Mid Game)

Here is where your factory's footprint explodes. Yellow matrices at 60/min demand 3 Matrix Labs fed by a sprawling network: 2 Diamond Smelters, 4 Titanium Crystal Assemblers, 6 Organic Crystal Assemblers, 3 Plastic Chemical Plants, and a staggering 12 Oil Refineries to manage all the cracking loops. You'll also need 6 Titanium Smelters, 1 Water Pump, and 3 more Graphite Smelters just to keep the inputs flowing. The final recipe wants 120 Diamonds and 60 Titanium Crystals per minute, which breaks down to 120 graphite for the diamonds and 180 titanium ingots plus 60 organic crystals for the crystal line. If your oil setup wasn't organized before, it will be now.

Purple science is the painful part. 60 purple matrices per minute requires 15 Matrix Labs and a forest of 11 Assemblers split across processors, circuits, photon combiners, prisms, and glass. You'll need 3 Smelters for iron and copper, 2 Chemical Labs converting Fire Ice into graphene, and 2 Miners pounding stone for silica. The raw material bill is brutal: 120 iron ore, 60 copper ore, 180 fire ice, and 120 stone every minute. The particle broadband branch alone - a single assembler - needs 2 graphene per second from 2 fire ice plus a photon combiner, which itself needs prisms from glass, which needs silica from those stone miners. It's a nested nightmare of dependencies.

Green & White Matrix (End Game)

Green matrices are where rare resources gate your progress. 60 green matrices per minute needs 4 Green Matrix Assemblers fed by 3 Graviton Lens Assemblers, 2 Diamond Smelters, 1 Titanium-Glass Assembler, 4 Strange-Matter Colliders, and 2 Processor Assemblers. The ingredient list gets exotic: 60 optical grating crystal per minute joins the usual iron, copper, silicon, fire ice, and coal. Each graviton lens sub-factory running at full speed burns 4 diamonds per second (240/min), 2 titanium glass per second (120/min), and 1 strange matter per second (60/min), so those colliders can't keep up unless you feed them pure fire ice and titanium consistently.

White science is the final boss. 60 white matrices per minute needs 12 Universe Assemblers running non-stop, but it also forces you to keep producing 60/min of every other matrix color simultaneously. That means you're not shutting down your old factories - you're adding 8 blue, 6 red, 6 yellow, 8 informational, and 6 green assemblers to the mix, plus 4 Ray Receivers pulling roughly 1 GW from your Dyson sphere for antimatter. If you're using Proliferator Mk-III for the 60% speed bonus (and you should), the power draw hits 9.5 MW just for assemblers and sprayers. You'll be importing iron, copper, silicon, coal, fire ice, titanium, and optical grating crystals from multiple star systems, and you'll pray your logistics network doesn't buckle.

Starter Smelting Arrays

If you're still hand-feeding smelters in the early game, this 12-smelter tile will save your sanity. Twelve Arc Smelters (T1) hit the magic number that perfectly saturates a green belt at 12 items per second, which means you can run this on blue belts without any weird compression issues. The whole thing stays lean on power, too.

Here's how tight it is: four Pile Sorters in a 3-wide column handle both the ore input and the ingot output, so the footprint clocks in at just 11×13 tiles - 143 tiles total with 95 buildings. You can slam six MK-I Miners on a normal vein, crank them to 100%, and they'll feed exactly 12 ore per second, no underground belts required for parallel setups.

Power is built right into the frame. Two Tesla Poles dropped down the middle give you 3-wide coverage, so you can tile this north-south or east-west without ever hitting a power gap. And when you're ready to scale up, swapping to T2 Smelters keeps the same 1:1 ratio. The footprint even supports Plane Smelters later, though you'll need to tweak the sorters for that catalyst input.

Oil Processing Block (6:8 Ratio)

Oil processing loves to jam, but this 6:8 layout doesn't. Six Plasma Refineries churn out 6 hydrogen and 12 refined oil every 4 seconds, which is exactly what eight X-Ray Crackers need - except those crackers demand 8 hydrogen every 4 seconds, so you're short by 2. That's where the priority splitter trick comes in.

Set one Priority Splitter to 'Crackers → output first,' and it will force-feed the crackers their 8 hydrogen before anything else gets through, which prevents the classic hydrogen deadlock loop. You'll also want to buffer the refined oil in a tank, then loop 2/3 of it back to the top row input. This keeps the crackers fed even if your crude supply hiccups.

The whole block pulls 12 MW - 4.32 MW from the Plasma refineries and 7.68 MW from cracking - so it fits on a single Mk-II Power Pole with a tidy 6×14 footprint. And because the math is self-balancing, you get a clean surplus of 6 extra hydrogen every 4 seconds that you can siphon off after the priority splitter for export or thermal power.

Fractionator Loop Designs

Deuterium loops are all about belt speed and priority management. On a Mk1 (yellow) belt, you can run roughly three Fractionators and pull about 0.06 deuterium per second - that's 6 hydrogen per second times the 1% conversion rate. Step up to a Mk2 (green) belt and you can double to six fractionators for ~0.12 deuterium per second, using the same single-splitter refill point.

Mk3 (blue) belts are where things get spicy: you can push 15 to 17 fractionators on one blue belt, cranking out around 0.3 deuterium per second. The key is a Priority Splitter set to 'loop input priority,' which recycles whatever hydrogen is already in the loop first and only tops off the gaps. This keeps the build compact and light on UPS.

If you're building mega-scale and every CPU cycle counts, swap the splitter for a T-junction side-load. The supply belt merges hydrogen only when a gap exists in the loop, so you skip the splitter logic entirely while keeping the same priority behavior.

Advanced Optimization & Tools

When your factory spans multiple star systems and you've got more smelters than you can count, that's when you stop eyeballing ratios and start reaching for calculators. Lucky for you, the DSP community has built an entire suite of tools, and they're all free.

Online Calculators & Tools

FactorioLab is probably the most flexible option out there, and it's not just for DSP. It supports Satisfactory and Factorio too, which means if you jump between factory games, you won't need to relearn a new interface every time. You get one-click game data sets, a color-blind friendly UI, and handy toggles for proliferator and belt tiers. It's great, but it's generalist by design.

If you want something laser-focused on DSP, DSPC (dyson-calculator.com) is the community gold standard. It's been updated through EA build 0.8.23.9832, includes every single recipe, and handles the tricky math for arc-smelters and Mk.III proliferators without breaking a sweat. The interface is clean, and you won't have to second-check whether the game version matches.

Now, maybe you don't need full production chains - you just want to know how many machines feed one belt. DSP Ratios (dspiratios.com) gives you exactly that: clean, simple ratio tables with no power calculations or map clutter. It's perfect for quick checks when you're mid-build and just need a fast answer.

For the megabase builders who think in city blocks, Sage Calculator takes a different approach. Instead of targeting a specific items-per-minute rate, it calculates the exact machine counts you need to fill a given building area. This is clutch when you're optimizing compact layouts and every tile matters.

And if you're really lazy (no judgment), the Production Rate Calculator mod on Thunderstore lets you drag-select any area in-game with Alt+X to see real-time production rates. No alt-tabbing, no spreadsheets - just point, drag, and fix your bottlenecks on the fly.

When Ratios Don't Matter

Here's the thing about chasing perfect ratios: DSP is engineered to make that obsession mostly unnecessary. The community mantra 'just add one more machine' exists for a reason. You've got infinite resource nodes and belts that move absurd volumes per second, so the pressure for 100% efficiency just isn't there. In the early game, slapping down an extra assembler is almost always faster than recalculating your entire line.

This mindset is so baked into the culture that Youthcat Studio even hid an achievement called Just One More Machine. You get it for hand-placing 1,000 buildings in a single session without copy-paste, which is both a joke and a warning. Chasing perfection without enemies or time pressure can burn you out fast, and the veterans know it. So yeah, use the calculators when you're bored or planning a megabase, but don't let them slow you down when 'eh, one more should do it' works just fine.

Scaling to Megabase Levels

When you start targeting 1,000+ white science per minute, 'one more machine' stops being cute and starts being expensive. At this scale, you need every efficiency multiplier stacked and working together.

Veins Utilization (VU) is your first priority. At VU-60, ore veins last roughly 30 times longer and mine about 7 times faster. That means a single patch can feed a factory for dozens of hours instead of running dry after one. It's exponential, so every level matters more than the last.

Proliferators are the next multiplier. Mk-III spray doubles your output - that's 100% extra product - which effectively cuts your raw material demand in half for every colored cube. Stack this with beaconized labs: two Mk.III beacons plus Speed-3 proliferators give labs +100% hash speed, which drops your white science lab count from 60 down to 30 for 3,600 cubes per minute. That's half the footprint, half the power, half the headache.

Speaking of targets, most megabase guides aim for 3,600 cubes per minute to guarantee 60 hashes per second regardless of research track. To support that, you'll need around 75 ray-receivers (each producing 8 antimatter per second) for 600 antimatter per second, plus about 25 artificial suns to cover the 22 GW power draw. These numbers aren't suggestions - they're the math that separates a big base from a stable one.

Conclusion

Whether you're planning a compact starter block or scaling to thousands of white science per minute, the key is balancing precision with practicality. Use the community's powerful calculators to plan, but remember the game's forgiving nature: sometimes, the best optimization is simply adding one more machine. Now, go forth and build your sphere.