926/reactor

[MacThule]

So I just made a power grid with 317 modules that yields 293,580 power. That's 926 per reactor.

Is that pretty average for a good power grid?
[DOUBLEPOST=1423219524,1423218889][/DOUBLEPOST]scratch that - i'm getting 1,199 per reactor now

 

[Thalanor]

The less groups and the larger each group, the more efficient you get. The "best" reactor would get the full bonus with just one group. However, that rarely is possible - it all depends on the shape of your ship. There is no way to determine the efficiency of a reactor without knowing the space and shape constraints of the ship hull.

 

[MacThule]

Do other modules like hull affect the per-unit output of power reactors? I haven't seen that so far. I'll have to test that.

I hear what you're saying; I've watched the vids. Those guys either A) missed one little thing or B) deliberately held something back to keep an edge. Because power lattices are not this effective, and now everyone is using what's in the videos without testing to see if they can do better.

You can determine the efficiency of your reactor grid by dividing the total output by the number of reactor blocks you used to make it.

I made a grid that put out 1199 power PER REACTOR. 630 reactors yielding 755,136. And it's a single unit with no separations. I haven't even tested the idea of 8 such highly effective units forming corners of a cube or the like, but I did test separating this design into groups and it loses efficiency. Drastically. A cube lattice gains efficiency when you separate the corners to create separate groups, I know, I've confirmed this in testing.

I compared notes with some people using "textbook" power lattices (several individually effective groups extended over large 3D space), mostly cube-shaped. They were getting yields in the 200-300/reactor range from their reactor grids.

Just wanted to compare notes with someone who has built what they consider a "great" lattice power grid and see if it beats what I put together. I may be just doing something wrong with the lattices and not getting a fair result when comparing their efficiency to what I made.

 

[Thalanor]

The thing is that it is impossible to compare power setups if the context (the ships they are built into) is not the same. As the shape of the hull, interior etc of a ship determines how much efficiency can be achieved, it only makes sense to compare reactors under the same space constraints. In no case does it make sense to use a prebuilt reactor that does not make use of the ship's dimensions as well as possible.

If no ship is involved, the most efficient reactor in terms of power output per block would be one long line that extends about 1200 blocks, or alternatively a 3D cross with each dimension being a 400m long stick. After that, no further box dimension bonus is gained and all further blocks will merely contribute 25 additional regeneration per second per block, regardless of where they are placed, as long as they don't touch the reactor supplying the bonus.

 

[CyberTao]

To add onto what Thalanor said, each reactor gives a flat 25 e/sec, and there is a shipwide grouping bonus for up to 1 million e/sec (total). The grouping bonus is dependent on the dimensions of the groups, so stick-framing is best in that regard. Once you understand how grouping works, you can move on to "power nesting" which is wrapping the reactors around the ship's shape to get the maximum bonus (and allow multiple groups if you don't reach the cap in just 1).

There is the wiki at the top of the forums, it has a much better breakdown and explanation for power than we could probably explain, plus charts, graphs, and formulas.
http://starmadepedia.net/wiki/Power_Systems

 

[MacThule]

Okay... I want to communicate this without coming off wrong, so please bear with me

FIRST: I am familiar with the accepted canon of what makes an "optimal power grid" (i.e. separate grouping, framing, nested lattice structures). Watched the vids. Built the grids. No one's made a T-shirt yet. I just thought ppl on the forums would probably know of advanced ways beyond the tutorials.

Now the way I read it, you're telling me two things, guys:

1. There's an optimal way to build power grids
   
2. There's no way to compare power grid construction to determine which is optimal

Those are mutually exclusive; if you can't compare grids, how do you know nested lattices or spirals are more efficient?

I disagree with #2, and agree with #1 but firmly believe that everyone is watching tutorials on YouTube and taking those for fact without challenging the information being presented. Like people believing everything they read in the Bible or see on Fox, without question.

You CAN compare power grid efficiency between ships regardless of shape because addition of other modules does not affect reactor output and because it's simple math: total/reactor=output per reactor. This is not affected by shape or size of the ship, though the result of the equation is, and this is a measure of efficiency.

Test it please. Build your best lattice power grid of any size and I'm guessing you aren't likely to get efficiency much better than 300e/sec/reactor block even if you DID base the entire ship on what the YouTube tutorials say the is the best power grid (line segments, spirals, nested lattice, etc), which should yield the optimal power output.

I've test-confirmed 1,198e/sec with 5 straight lines all connecting. Nothing fancy. So far no one I've talked to has beaten that with any other supposedly optimized power gird. The scary thing is that I don't even know if that's maxed out, because so far every time I increase the size on this layout the EFFICIENCY (i.e. the e/sec, not merely total output) increases. I just stopped around 600 because I heard there's a soft-cap around there where you get diminishing returns. And it was 3am.

So I'm not trying to be dick, but unless someone can beat a per-reactor efficiency of 1,198, the conventional wisdom regarding optimal power grid construction is completely misleading because those things aren't even 1/4 as efficient as what I'm doing.

 

[Megacrafter127]

1. There's an optimal way to build power grids
   
2. There's no way to compare power grid construction to determine which is optimal

Those are mutually exclusive; if you can't compare grids, how do you know nested lattices or spirals are more efficient?

1 is definetly true due to the boxdim bonus.
2 is also true depending on what aspect shall be optimal(e.g. reactor count, total space used by the reactor, extendibility, …)
If a reactor is optimal in one, it is very unlikely to also be optimal in the others, so there is no optimal reactor, just an optimal reactor for a specific set of limits.

 

[MacThule]

1 is definetly true due to the boxdim bonus.
2 is also true depending on what aspect shall be optimal(e.g. reactor count, total space used by the reactor, extendibility, …)
If a reactor is optimal in one, it is very unlikely to also be optimal in the others, so there is no optimal reactor, just an optimal reactor for a specific set of limits.

Well I understand what you are saying - there isn't an optimum that works for every ship. But I think you believe this because you assume optimization requires things like as much spiraling or nesting as possible in a given hull, things that cannot be done the same from hull to hull. You can't optimize a streamlined ship to incorporate a nested cubic lattice. Agree 100% given the premise that optimization requires such clunky structures.

But the vast majority of hulls can incorporate a version of what I'm using, and since it only requires 1/4 of the blocks of the reactor grids currently being used, it's worth it to run a couple simple lines through the interior.

At the end of the day though, if no one wants to take me up on my challenge that's fine. I'll be rocking 1200e/sec/block and own all of you with your YouTube batteries pushing 250e/sec/block because I can fit that many more blocks into a ship of the same mass. :-D
[DOUBLEPOST=1423261707,1423258306][/DOUBLEPOST]Eff it.
It'll make my PvP even sweeter knowing I've shared my best power grid and everyone still builds decorative junk out of sheer obstinance ^_^

An optimal Power Reactor unit is identical to an optimal Docking Enhancer unit: Four lines forming an "X" in two dimensions and a fifth line projecting into the 3rd dimension from the exact center of the X.

It's so obvious that I kicked myself once I realized it after hours of testing different power grids and hating them all. The clues are right there on all of our turrets and stardocks.

It apparently doesn't matter proportion, only number of blocks... although my testing on such units wasn't as thorough as testing equal arm sizes. More blocks = more efficiency per block (at least up to 630 reactors) and way more total power. In the example of 630 blocks, it doesn't matter if your main spine (length) is 626 long with 4 single blocks projecting out into the other dimensions, or if the ship is a cube or shaped like a frisbee with a power grid 5 long and 625 blocks projecting out into height and width, or a triangle or a pyramid. 630 blocks yields 755,000e/sec regardless of proportion as long as you pretend its a docking enhancer array. Which is really weird, BTW, because I don't think it's a standard volume calculation even though it appears to be at first glance. Mind, that at 25 reactors you only get like 200-something per block, but as you scale up it just continues to improve in efficiency.

So, you see, it's very flexible as well in terms of shape while still providing optimal power FAR beyond any nested nonsense or spiral. And it's so easy any noob can do it. Unless you're trying to re-create a ship from a piece of fiction you love, this can really be incorporated into almost any ship.

Enjoy!
(or don't - your game)

 

[CyberTao]

A single reactor group with 594 blocks all extending the dimensions grants 1258.7 e per block according to the wiki.
Honestly, the fact that you are using a X sounds like you are wasting blocks, try just using a / and a line, or just a cross. It's all exactly the same mate.
Why do people use less effective setups? Either cause they don't know power or the power demand is greater than the soft cap allows. After a certain point, efficiency goes down because the bonus from the group goes down.
How to optimize? Build a reactor that fits your ship's X Y Z dimensions, and not overdoing it. A single groups yields the best bonus but not all ships can fit something big enough.

Number of blocks don't matter, dimensions of the group matter.

 

[MacThule]

By "X" I meant a cross shape. This is semantic. If you know how to build a docking enhancer array you know what I mean :-D

My hat's off to whoever got 1258.7 from 594 blocks. That's good stuff. First statistic I've seen above what I've been building.