This has been bothering me since before version 2.0 even came out.
In https://factorio.com/blog/post/fff-428 a developer stated that the new feature allowing the circuit network to read the temperature of a nuclear reactor, made it unnecessary to store excess heat from nuclear reactors in steam tanks.
According to https://wiki.factorio.com/Nuclear_reactor , "Nuclear reactors have a heat capacity of 10 MJ/°C. Thus, they can buffer 5 GJ of heat energy across their working range of 500°C to 1000°C"
But even if you insert just one fuel cell at a time when the reactor's temperature is below the 500 C minimum, one fuel cell produces 8 GJ of heat energy in a reactor - before multiplying for the neighbor bonus.
That's much more than enough to heat the reactor to the 1000 C maximum, and any energy above that is wasted since the fuel cell keeps burning. It gets much more severe in reactors with neighbor bonuses.
So if you're relying on just reading the reactor temperature to control when you put in a fuel cell, you have to be immediately consuming most or all of the energy as electricity to make a "lossless" setup as the FFF post describes. But if you're constantly using most of the reactor output anyway, then you don't need to regulate it by reading its temperature either.
As far as I can see, it is still advisable to store the excess heat from at least 1 fuel cell cycle as steam in tanks and regulate the reactor fueling by reading the amount of steam.
I didn't post this until now because I've been revisiting these numbers sometimes to figure out where I got it wrong. In the meantime, all the factorio nuclear setups I've seen on blueprint sharing sites and on youtube have stopped using steam storage (if they were using it to begin with) and adopted temperature reading. If I'm not wrong, countless engineers have been wasting countless bazillowatts of nuclear power this way.
So what am I overlooking here?
Correct me or correct FFF#428
- Mad Inventor
- Burner Inserter
- Posts: 16
- Joined: Sun Mar 13, 2016 1:10 am
- Contact:
Re: Correct me or correct FFF#428
You overlooked this was not stated by the developers. Read the fff again. It was stated: "This makes it quite easy to set up a lossless smart reactor, which doesn't insert any fuel unless it is needed." and "... may prove to be quite useful ..."Mad Inventor wrote: Fri Aug 22, 2025 5:54 pm This has been bothering me since before version 2.0 even came out.
In https://factorio.com/blog/post/fff-428 a developer stated that the new feature allowing the circuit network to read the temperature of a nuclear reactor, made it unnecessary to store excess heat from nuclear reactors in steam tanks.
[...]
So what am I overlooking here?
That's all. It was not stated you don't need to buffer steam any more.
You don't need to buffer steam any more to get an idea if your reactor is heating up or down any more, but it's not being stated you don't need any steam buffer at all for a wasteless setup. You're completely right that the reactor alone cannot buffer the heat of one fuel cell alone.
By reading the temperature, you're able to react way faster to rising or falling temperatures. That's the only difference to reading the fill state of a steam buffer before.
By the way, you write newer blueprints are missing steam buffers - I personally still add steam buffers for a wasteless design. I do read the temperature for reactor control of course and store steam to save the heat. Without steam storage tanks, it's simply not possible to store all the excess heat during low power usage phases. At least if you use a bigger plant such as a 2x6 reactor setup and more than that. For 2x2 plants it's not very important, but with those bigger setups you have tons of neighbor bonus that's creating additional heat to store.
This is my 2x6 tileable temperature reading wasteless reactor setup including steam buffers: viewtopic.php?p=678041#p678041
Re: Correct me or correct FFF#428
You are thinking small, and only looking at the static / unloaded case. During the 200 seconds that a Uranium Fuell Cell burns some of the energy will be used to produce Electricity - not all of it has to be stored as Heat or Steam.Mad Inventor wrote: Fri Aug 22, 2025 5:54 pmThus, they can buffer 5 GJ of heat energy across their working range of 500°C to 1000°C"
But even if you insert just one fuel cell at a time when the reactor's temperature is below the 500 C minimum, one fuel cell produces 8 GJ of heat energy in a reactor - before multiplying for the neighbor bonus.
So what am I overlooking here?
2 Reactors absorbs 8GJ of heat energy across a working range of 600C - 1000C. Or with 8 reactors you can buffer 8GJ between 900C - 1000C from a single active Reactor. You do not need to feed all the reactors at once or activate the Neighbor bonuses - just use them for Buffer capacity.

There are many ways to determine how many Reactors need to be fed a Fuel Cell, without over-heating the entire system. You can measure the amount of Water (and thus: Steam) flowing through the Turbines using a Pump circuit, convert to MW, and figure out Neighbor bonuses; but the math gets complicated quickly and it is prone to failures from brownouts. A simpler method is a set of Inserters with "Threshold temperatures" Combinator controls, wired directly to the Reactor that it Feeds. Each Inserter is given a slightly-different set point(EG: 900 , 875 , 850... 500). When the reactors cool down due to the Heat Exchangers consuming heat they will self-throttle as more Energy is required to keep the same Temperature - like a primitive industrial PID controller. I am only showing 6 Reactors here because that is the Minimum to obtain a +300% Neighbor Bonus (when everything is active). Most of the time this setup will run on only the middle 2 Reactors.
Good Luck!
- Mad Inventor
- Burner Inserter
- Posts: 16
- Joined: Sun Mar 13, 2016 1:10 am
- Contact:
Re: Correct me or correct FFF#428
Tertius, this was the part of the FFF that I was referring to - "The previous solution of people reading the amount of Steam in storage tanks felt like it needed a bit more brainpower and engineering, but it's hard to justify missing capabilities with this reasoning."
I could have misinterpreted it, but other people including content creators have also interpreted this as meaning that steam storage is obsolete. My point is that steam storage is still necessary.
eugenekay, I did consider the fact that some will be used as electricity. The point of controlling the fuel cell insertion to begin with is for when you are not using all of it, that's true both for reading the temperature or the amount of steam.
I didn't think about the points you made after that - that you can activate reactors selectively, or that multiple reactors can buffer each other's heat output.
These sound cool and I'm going to experiment with setups like what you showed. I still wonder if they'll be necessary or an improvement on full steam storage, though.
Power usage changes in between inserting fuel cells - you can activate # of your reactors at the same time and then drain your stored steam (if usage increases) or overflow your energy storage (if usage drops). If you insert fuel at different times then the neighbor bonuses will change constantly and the control system will have to keep track of that and of how much fuel-cell time is left in each reactor, etc.
When you have enough steam storage for an entire, 200-second fuel cell cycle at full output, you will just activate the reactors less often if you don't need the full output, and activate them together to replenish the stored steam when it runs low enough that you can buffer the next cycle. Steam turbines adjust their consumption to electricity use, so this system already accounts for changing power needs.
And this way you are always at the full neighbor bonus, so always getting the maximum energy per fuel cell. I know this is just an optimization concern, since most of the time you have plenty of fuel cells, but storage tanks and space to build them are also plentiful.
I could have misinterpreted it, but other people including content creators have also interpreted this as meaning that steam storage is obsolete. My point is that steam storage is still necessary.
eugenekay, I did consider the fact that some will be used as electricity. The point of controlling the fuel cell insertion to begin with is for when you are not using all of it, that's true both for reading the temperature or the amount of steam.
I didn't think about the points you made after that - that you can activate reactors selectively, or that multiple reactors can buffer each other's heat output.
These sound cool and I'm going to experiment with setups like what you showed. I still wonder if they'll be necessary or an improvement on full steam storage, though.
Power usage changes in between inserting fuel cells - you can activate # of your reactors at the same time and then drain your stored steam (if usage increases) or overflow your energy storage (if usage drops). If you insert fuel at different times then the neighbor bonuses will change constantly and the control system will have to keep track of that and of how much fuel-cell time is left in each reactor, etc.
When you have enough steam storage for an entire, 200-second fuel cell cycle at full output, you will just activate the reactors less often if you don't need the full output, and activate them together to replenish the stored steam when it runs low enough that you can buffer the next cycle. Steam turbines adjust their consumption to electricity use, so this system already accounts for changing power needs.
And this way you are always at the full neighbor bonus, so always getting the maximum energy per fuel cell. I know this is just an optimization concern, since most of the time you have plenty of fuel cells, but storage tanks and space to build them are also plentiful.
Re: Correct me or correct FFF#428
I guess it's mainly laziness if people don't include steam buffers any more. Did you actually try this with the intention to buffer steam in case consumption is less than production? If you tried, you will realize the steam tanks are already full when you need them empty and they are empty, if you need them full. That's due to the new fluid mechanics that fills every fluid segment instantly and evenly. There's no flow any more.Mad Inventor wrote: Fri Aug 22, 2025 8:32 pm I could have misinterpreted it, but other people including content creators have also interpreted this as meaning that steam storage is obsolete. My point is that steam storage is still necessary.
You need active buffer management now. In Factorio 1.1 the delay due to steam flow is big enough so the buffer is just half full if you need it empty and half empty if you need it full, so you were able to use at least half of the steam buffers. Directly and without pumps.
If you inspect my blueprint, you see circuit controlled pumps. They are not there for the lulz or because I didn't realize in 2.0 any nuclear power plant can work completely without pumps. Instead, they are there to keep the buffers empty as long as there is no need to buffer and allow filling it only to cool down the reactor at low power consumption to avoid fuel wasting. This was quite a challenge to design, because it requires a bit of understanding of the dynamics involved with reactor heating and cooling, heat flow through heat pipes and new fluid mechanics.
Re: Correct me or correct FFF#428
None of this is Necessary; it is just a fun mini-Optimization problem.Mad Inventor wrote: Fri Aug 22, 2025 8:32 pmThese sound cool and I'm going to experiment with setups like what you showed. I still wonder if they'll be necessary or an improvement on full steam storage, though.

In terms of Game-Performance-per-Gigawatt; all of these setups are terrible for performance because of the relatively low energy density of Heat Pipes. It is always a "better idea" to upgrade to using Fusion power. Even with the Hot / Cold Fluoroketone network it is still less of a Performance / UPS impact (in fluid-network) because of the much higher Power Density of the Reactors/Generators: you get more Gigawatts per Entity.
- Mad Inventor
- Burner Inserter
- Posts: 16
- Joined: Sun Mar 13, 2016 1:10 am
- Contact:
Re: Correct me or correct FFF#428
Tertius, you've lost me here. I've used exactly the same lossless nuclear power and steam buffer setups before and after 2.0, and fluid flow or pumps aren't relevant to it.
Heat exchangers adjust their water usage and steam production to the heat they receive from reactors. Steam turbines only consume as much steam as they need to satisfy the electric network demand. Both of these were unaffected by 2.0.
Heat exchangers output steam into tanks, steam turbines draw steam from the tanks as needed. I have enough tanks between them to store the output of all the reactors for at least 200 seconds, which is the amount of time one fuel cell will burn.
I read the amount of steam in the tanks to a combinator that enables the inserters putting fuel into the reactors. The only conditions in the combinator are no fuel in the reactors, and low enough steam in the tanks.
I really don't see how the fluid system would affect this.
To calculate how many tanks I need - according to https://wiki.factorio.com/Storage_tank - 2.4 GJ per tank. With (for example) a 2x2 square of reactors that each have 200% neighbor bonus, each fuel cell yields 24 GJ, so one in each reactor is 96 GJ.
With around 40 tanks I can store one fuel-cell cycle of energy, but I only need to have a wire reading the steam in one of those tanks. So long as they are connected directly or by pipes, the fluid will equalize in all of them.
Heat exchangers adjust their water usage and steam production to the heat they receive from reactors. Steam turbines only consume as much steam as they need to satisfy the electric network demand. Both of these were unaffected by 2.0.
Heat exchangers output steam into tanks, steam turbines draw steam from the tanks as needed. I have enough tanks between them to store the output of all the reactors for at least 200 seconds, which is the amount of time one fuel cell will burn.
I read the amount of steam in the tanks to a combinator that enables the inserters putting fuel into the reactors. The only conditions in the combinator are no fuel in the reactors, and low enough steam in the tanks.
I really don't see how the fluid system would affect this.
To calculate how many tanks I need - according to https://wiki.factorio.com/Storage_tank - 2.4 GJ per tank. With (for example) a 2x2 square of reactors that each have 200% neighbor bonus, each fuel cell yields 24 GJ, so one in each reactor is 96 GJ.
With around 40 tanks I can store one fuel-cell cycle of energy, but I only need to have a wire reading the steam in one of those tanks. So long as they are connected directly or by pipes, the fluid will equalize in all of them.
-
- Filter Inserter
- Posts: 375
- Joined: Sat Oct 07, 2023 6:44 am
- Contact:
Re: Correct me or correct FFF#428
Neglected here is the heat capacity of any attached heat pipes or heat exchangers, which each contribute 500MJ of heat capacity. So if you have a reactor with a 300% neighbor bonus (8GJ * 4 = 32GJ) you'd need the reactor itself (5GJ), 16 heat exchangers (8GJ), and 38 heat pipes (19GJ). Even if you don't strictly need that many heat pipes, they are more energy dense than steam tanks (4.5 GJ per 9 tiles vs 2.425 per 9). In actual application you will lose a little bit of working capacity because of the temperature gradient, but even building a straight line of 38 heat pipes will only result in an 8% reduction.Mad Inventor wrote: Fri Aug 22, 2025 5:54 pm But even if you insert just one fuel cell at a time when the reactor's temperature is below the 500 C minimum, one fuel cell produces 8 GJ of heat energy in a reactor - before multiplying for the neighbor bonus.
That's much more than enough to heat the reactor to the 1000 C maximum, and any energy above that is wasted since the fuel cell keeps burning.
You only have to use what you can't store. So if you can store 90% of the heat, you'd only have to use 10%, or 16MW, of a 300% neighbor bonus reactor. And it's not hard to store enough heat even by accident.Mad Inventor wrote: Fri Aug 22, 2025 5:54 pm But if you're constantly using most of the reactor output anyway, then you don't need to regulate it by reading its temperature either.