Does "Quality" Make Fusion Reactors Worse?

[fryyyy]

Fusion reactors get their productivity bonus proportionally to their current load. Quality reactors have a higher max output. This means for the same load (able to be met by your max output,) a set of quality reactors will always have a lower productivity bonus than a set of normal reactors.

If you do encounter a scenario where your max load is greater than your max output, it's by far easier and more efficient (with higher max productivity bonuses) to add more normal quality reactors than it is to add quality reactors.

The only benefits I can see to quality reactors is less space required. Even then, the reactor itself is a minor contribution. Most of the footprint comes from the generators.

Am I missing something? I've looked around a bit and nobody is talking about this. It feels like having reactors be worse with higher quality violates the concept of quality and is particularly egregious for such an end-game item.

[mmmPI]

you are not missing much by having a lower efficiency in the core, the cells are very cheap, so saving on them is not super useful imo, whereas having a smaller footprint for power is always good

[fryyyy]

fusion.jpg (506.53 KiB) Viewed 1156 times

Here are two fusion set ups with almost the same power output (1.7 GW vs 1.71 GW.) One of them technically takes up less space, but requires 14% more coolant (offsetting its already meager space advantage,) eats through 10%-25% more fuel (despite the reactors' collective passive draw being 20 MW less) and costs a couple orders of magnitude more manufacturing to produce. Higher quality reactors are just worse, and the higher the quality the worse they are.

The issue isn't "Why isn't quality good enough," it's "Why does quality make it worse."

[Zavian]

I look at those two designs, and the first thing I notice is 2 less reactors, and the consequent the loss of neighbour bonus. And that loss of neighbour bonus is a massive loss. Personally I think that is why you feel the design with quality reactors is worse than the normal reactors.

[mmmPI]

The question seemed to me "Does quality make fusion reactors worse" and it's not the case imo. The setup shown doesn't even use quality generator, only the reactor and in a suboptimal configuration where it doesn't show the fueling footprint, it's not conclusive to me.

[fryyyy]

plasma 2 fixed 2.png (2.9 MiB) Viewed 1007 times

I look at those two designs, and the first thing I notice is 2 less reactors, and the consequent the loss of neighbour bonus.

I was being generous with my setups. Yes, there is lost potential neighbor bonus, but you never reach that potential because the neighbor bonus you get is proportional to the "current load / max output" of the reactor. Adding two more reactors drastically increased the "Max Output" and therefore damages the effective neighbor bonus. The 3 epic-reactor setup actually has higher efficiency, with avg plasma temp of 3 M°C.

The 5 epic-reactor setup burns more fuel for the same output than the 3 epic-reactor setup in the original post.

The setup shown doesn't even use quality generator

Quality generators are indeed better than normal generators. Their use does not change the efficiency of the reactor, and would benefit a setup with either normal reactors or quality reactors equally.

[mmmPI]

The 5 epic-reactor setup burns more fuel for the same output than the 3 epic-reactor setup in the original post.

Similarly a 5-normal-reactor, burns more fuel for the same output than a 3-normal-reactor, following your logic.

Therefore it is unrelated to "quality" since the same phenomemon occurs only considering "normal" quality.

[Orum]

Similarly a 5-normal-reactor, burns more fuel for the same output than a 3-normal-reactor, following your logic.

No, they don't¹. However, let's stick to the original problem of just varying quality and nothing else.

¹ This can depend on the build order of the fusion reactors, but it's beyond the scope of the original issue with quality and reactors, so I'm not going to go into depth on it here.

Therefore it is unrelated to "quality" since the same phenomemon occurs only considering "normal" quality.

I actually tested this in-game, with two nearly identical setups; the only difference was the quality of the fusion reactors and the generators. It looks like this:

setup.jpg (847.69 KiB) Viewed 506 times

The top side is normal quality fusion, the bottom is legendary. The steam "batteries" on the right are just there to get everything started, and they are both connected to their respective power grids on the left at the same time via the circuit in the middle (though each side is timed independently, so they don't actually have to start simultaneously; more on that below). After they are connected, they're both allowed to run dry so that both sides get exactly the same amount of power from them.

The "Electric energy interfaces" are both set to consume 800 MW of power, though the top electrical network actually draws slightly more power than the bottom, as fusion reactors require power to produce plasma, and more reactors are running simultaneously on the top. However, as you'll see, the adjacency bonus greatly outweighs the additional energy draw of running more reactors. As we're only interested in how efficient they are for external loads anyway, it's important that we don't consider the load from the reactors themselves.

To accomplish this, I did a count of how long each quality ran for, using a tick counter that you can see in the image. As soon as power starts flowing, it starts counting, and as soon as it runs out of power, it stops. Both sides started with 50 fusion power cells in a 'sushi-belt' style delivery system, which started as soon as the steam batteries were connected.

The results? Well, they weren't even close. The 'normal' reactors/generators provided 800 MW external power (i.e. power to the electric energy interface & the tick counter) for ~441K ticks, while the legendary only powered it for ~292K ticks. That's right, I got more than 51% additional energy out the fuel cells by using lower quality fusion reactors/generators. Why? The answer is obvious, and is what the original post is about: adjacency bonuses.

That's not to say that higher quality reactors are useless. They will give you more power per area than lower quality ones, but this comes at a severe cost to efficiency (as in energy produced per fusion power cell). As area is rarely an issue in this game, outside of space platforms to a small extent, and maybe to an even lesser extent, Aquilo, in general I prefer using lower quality fusion reactors/generators over higher quality ones. If this is not what the devs intended, then they should fix it.

[Junorus]

Nice demonstration. Could you run the legendary setup with 125 fuel cells with 2000 MW drain for a comparison?

[Orum]

To compare it to what, exactly? I'd need to increase the number of generators and reactors on the normal side to make it possible of producing that output, though it should still hold true that it would be more efficient.

[Rubicj]

To rephrase the problem:

Fusion reactors are most effecient when running at 100% capacity.

If you blindly upgrade reactor quality you will have more capacity, and therefore less efficiency.

If you replace 5 normal reactors with 2 legendary reactors, which have the same output, you'll have less neighbor bonuses, and therefore less efficiency.

Phrased this way, I see how it can be surprising to upgrade quality and see more fuel usage for no upside, but I think it makes sense.

[mmmPI]

Similarly a 5-normal-reactor, burns more fuel for the same output than a 3-normal-reactor, following your logic.

No, they don't¹. However, let's stick to the original problem of just varying quality and nothing else.

¹ This can depend on the build order of the fusion reactors, but it's beyond the scope of the original issue with quality and reactors, so I'm not going to go into depth on it here.

Yes they do. You demonstration is useless in prooving or disprooving this because you didn't try to get the same amount of energy at the end. You are joining a conversation that you may not have understood judging by the pointless argument.

To compare it to what, exactly? I'd need to increase the number of generators and reactors on the normal side to make it possible of producing that output, though it should still hold true that it would be more efficient.

Lol x)))), no you don't just do a normal 5 reactor setup and a 3 reactor setup, try to get 100MW out of them both, then obvisouly the most efficient will be the 3 reactor setup. Which proove my point that it has nothing to do with quality, the same phenomenon that was mentionned by OP occurs here also despite no quality being used at all.

Aquilo, in general I prefer using lower quality fusion reactors/generators over higher quality ones. If this is not what the devs intended, then they should fix it.

This is user mistake imo there is nothing to fix, you do what is not the mot efficient because you don't undertand how to properly use the better setup imo, the fix should come from user.

To rephrase the problem:

Fusion reactors are most effecient when running at 100% capacity.

If you blindly upgrade reactor quality you will have more capacity, and therefore less efficiency.

If you replace 5 normal reactors with 2 legendary reactors, which have the same output, you'll have less neighbor bonuses, and therefore less efficiency.

Phrased this way, I see how it can be surprising to upgrade quality and see more fuel usage for no upside, but I think it makes sense.

Indeed, therefore this not a problem related to "quality", it is problem of bad fusion setup. ( which is probably why this post is in gameplay help )

[Orum]

Yes they do. You demonstration is useless in prooving or disprooving this because you didn't try to get the same amount of energy at the end. You are joining a conversation that you may not have understood judging by the pointless argument.

Huh? The goal is not to get the same energy in the end, it's to see which is more efficient--lower or higher quality.

Lower quality will always be more efficient because it can get greater neighbor bonuses for a given output, unless your output is so low that it doesn't exceed the output of a single normal quality reactor.

Lol x)))), no you don't just do a normal 5 reactor setup and a 3 reactor setup, try to get 100MW out of them both, then obvisouly the most efficient will be the 3 reactor setup. Which proove my point that it has nothing to do with quality, the same phenomenon that was mentionned by OP occurs here also despite no quality being used at all.

You're comparing 3 normal reactors to... 5 normal reactors? What does that prove?

This is user mistake imo there is nothing to fix, you do what is not the mot efficient because you don't undertand how to properly use the better setup imo, the fix should come from user.

The most efficient setups require lower quality. My example demonstrates that in a painfully obvious way.

Indeed, therefore this not a problem related to "quality", it is problem of bad fusion setup. ( which is probably why this post is in gameplay help )

It is directly related to quality. More active reactors gives higher efficiency bonus. How do you get the most active reactors? Use those with the lowest output. It's that simple.

[mmmPI]

You're comparing 3 normal reactors to... 5 normal reactors? What does that prove?

That the phenomenon of loss of efficency is not happening only due to quality. You should follow the original reasonning, you are just misunderstanding at this point.

Edit from original post :

Fusion reactors get their productivity bonus proportionally to their current load. Quality reactors have a higher max output.

This means for the same load (able to be met by your max output,)

a set of quality reactors will always have a lower productivity bonus than a set of normal reactors.

In case you missed it, my comparaison is just aiming at showing that if you replace "quality reactor" by "more reactor" you arrive at the same conclusion. This means the problem is not about quality, it's about how the reactor are designed.

[spacedog]

It is directly related to quality. More active reactors gives higher efficiency bonus. How do you get the most active reactors? Use those with the lowest output. It's that simple.

This is the part where you went wrong. It may seem that way from your measurements, but you're interpreting the results incorrectly. Rubicj explained it pretty clearly, but I'll try another way. The problem isn't with quality, it's with how the neighbor bonuses are designed.

As you're aware, the neighbor bonuses scale the plasma temperature based on how much a reactor is producing versus its maximum. The inefficiency directly stems from when the reactors can produce more plasma than the generators consume (in quantity, temperature, or both). When that happens, the reactors scale back their production until equilibrium is reached. This causes the reactors lose some of their neighbor bonuses, and thus their fuel efficiency. So this isn't directly related to quality. Quality is just one way you can end up with too much reactor capacity.

The people I've seen getting stuck on this concept are in a nuclear power mindset. There, you get the full neighbor bonuses whether you use the electricity or not, and it's pretty easy to visualize the whole thing since the fuel gets used up at a constant rate.

With fusion power the fuel consumption corresponds directly to the amount of fluoroketone being converted to plasma. Specifically, it takes 25 MJ of energy to convert one unit of fluoroketone into plasma, regardless of the temperature the plasma is heated to. So to get the most "laws of thermodynamics be damned" Factorio Magic™ out of your fuel, you need the plasma to be as hot as possible -- which means running as close to the maximum capacity of the reactor as possible.

The reason you get worse efficiency by adding reactors (of any quality) isn't extra fuel consumption. If you could somehow keep the plasma temperature the same, then the total amount of plasma generated would just be split evenly between the reactors, they'd collectively consume the same amount of MJ to heat that plasma, and your total fuel usage would be exactly the same. The only reason the fuel usage increases is because the plasma temperature goes down due to the surplus reactor capacity, and you lose more neighbor bonuses.

This is also why all of your tests are flawed. They would only be valid if the plasma temperature were the same in all of them. Your intuition is correct, however, that the only real benefit of quality reactors is space savings.

Once you wrap your head around all this fully, you'll realize the fusion model is actually a lot nicer than nuclear. As others have said, like nuclear fuel cells, it's trivial to be drowning in fusion fuel cells and efficiency is a premature optimization. IMO the right way to look at the neighbor bonus is like a buffer of sorts -- the more your power demands increase, the harder the reactors work to meet that demand. The extra efficiency from the increasing neighbor bonuses just gives you that much more breathing room to bring more reactors online before you start having power issues. Fusion reactors are also amazing at handling sudden large power spikes (i.e. railguns and lasers) for a small fraction of the fuel cost of other power sources.

[fryyyy]

In case you missed it, my comparaison is just aiming at showing that if you replace "quality reactor" by "more reactor" you arrive at the same conclusion. This means the problem is not about quality, it's about how the reactor are designed.

The problem isn't with quality, it's with how the neighbor bonuses are designed.

If you have Effect X, which can be caused by Cause A but also Cause B, then showing that Cause B can cause Effect X does not disprove that Cause A can cause Effect X.

If someone gets to the top floor of a building by elevator, you cannot say "There are stairs, using the stairs can get you to the top floor, therefore the elevator does not get you to the top floor."

If someone comes to you and says "help, the currency in my pocket is changing into lower denominations; I'm losing money" and then they show it to you. You demonstrably see it happening in front of you. You can't just say "Money can be lost by spending it, your problem is spending money." Spending money is a normal expected thing that happens, currency changing denomination is unexpected and alarming. Your response is not a solution to their problem, it doesn't even accurately reflect their problem, and you're not helping.

Nobody is saying using more or fewer reactors is an issue. Managing variable neighbor bonuses is a logistical problem in a game about solving logistical problems. That's fine, good even. The issue is that QUALITY, the thing that across every other end-product item provides only benefit, will detrimentally effect the function of one particular relatively expensive and late game item, and not in an insignificant way either. This seemingly violates the concept of quality, the core goal of quality. Nowhere in the FF#375 discussing quality are we led to believe quality is about managing pros and cons of the effects of quality.

Saying that some other factor produces the same negative effect that "upgrading" quality does, does not disprove or discount the fact that higher quality makes that thing worse.

[mmmPI]

Saying that some other factor produces the same negative effect that "upgrading" quality does, does not disprove or discount the fact that higher quality makes that thing worse.

You asked if quality made fusion reactor worse, and it's not the case, what make reactor fusion worse is when they are not running at full efficency, if you happen to make a design with reactor not running at full efficency it's not because of "quality".

If you happen to have the same utilization of your reactor in % of their max output, then the quality one will be just as efficient per fuel cell and produce a lot more power for the same footprint.

Of course if you make a bad design with quality reactor where they are not working at 100% and you compare it with a okayish design of normal quality it's easy to blame quality, but really the problem is you made a innefficent quality reactor, hence why this is in gameplay help section imo

[spacedog]

Saying that some other factor produces the same negative effect that "upgrading" quality does, does not disprove or discount the fact that higher quality makes that thing worse.

It makes that thing worse for your use case.

It sounds like you want quality to be a multiplier on the neighbor bonus instead of the max plasma/fuel rate. That would give you the behavior you're expecting. That would also be the opposite of what people trying to save space on their space platforms want, if they just want more punch out of a single reactor.

I can understand why you would see legendary quality being bad. It will reduce your plasma temperature for a given electrical demand. That's true, but it's not specifically the fault of quality. A few different people on this thread have tried to suggest a different way of looking at this, but you seem kind of stuck. I'm not sure how to help with that.

[Orum]

This is the part where you went wrong. It may seem that way from your measurements, but you're interpreting the results incorrectly. Rubicj explained it pretty clearly, but I'll try another way. The problem isn't with quality, it's with how the neighbor bonuses are designed.

Neighbor bonuses are a direct function of how many reactors are providing power (and of course, how many reactors are adjacent to them). As reactors turn on "one-by-one" until an equilibrium is reached, you'll always be better off with as many reactors as possible, each providing as close as possible to 100% output.

Lower quality reactor use directly correlates with larger amounts of active reactors (for any non-negligible load), and thus, higher adjacency bonuses. I don't understand why this point is so hard to get across, because it means that lower quality reactors are more efficient. Could it be changed such that it doesn't work this way by changing how neighbor bonuses work? Obviously yes, but that doesn't change the fact that that's not how neighbor bonuses work now, and quality has a big effect on the efficiency of a fusion reactor setup.

The people I've seen getting stuck on this concept are in a nuclear power mindset. There, you get the full neighbor bonuses whether you use the electricity or not, and it's pretty easy to visualize the whole thing since the fuel gets used up at a constant rate.

Nuclear fission power is much easier to get max (or near max) efficiency out of, as you can trivially store the power produced as either steam or heat. As you cannot store the power from fusion, outside of the crappy accumulator (even at legendary, it's awful compared to steam/heat energy storage), this is a much bigger problem there. That said, I still generally prefer normal quality fission reactors, as there are no benefits other than space savings for quality, and my reactors don't take up much of my base's footprint anyway.

This is also why all of your tests are flawed. They would only be valid if the plasma temperature were the same in all of them.

The test is for efficiency, and plasma temp is just a measure of efficiency. Obviously the temps will be lower when they're less efficient. You're asking for a test of efficiency, saying that the efficiency has to be fixed for the test to be valid. That doesn't make any sense, as then you're not testing anything.

Once you wrap your head around all this fully, you'll realize the fusion model is actually a lot nicer than nuclear.

I disagree. The only place I use fusion is on space platforms, and the only reason I prefer it there is due to the small footprint compared to fission. Even on the easiest planet to use it on (Aquilo, as you don't have to ship anything other than holmium there to sustain fusion), I don't bother with it, because I can't get heat out of it that I desperately need on that planet.

All in all, fusion is a major disappointment for me in SA. I had hoped I would see uses for it outside of just the niche of space platforms, but I just can't justify the hassle of doing so.

[mmmPI]

[Moderated by Koub : Off topic]