Posts by Omicron

Quote from masterP

I tried to make a design for the Tier 1 Reactor of Advanced Nuclear Reactors without any Chambers (3x3 Slots in the reactor, only single cells):
This is my first design generally:
http://www.talonfiremage.pwp.b…xxr9ejzsm0izitqf5racfx0jk
Mark I EC, 7 Mio. EU, 35 EU/t, Efficiency of 2,33

You're using reactor vents where you could use basic vents, which raises your cost by a lot. Swapping them out pushes copper costs from 186 down to 154. Also, in that configuration it becomes a valid choice to use basic heat exchangers instead of component heat exchangers, which saves you 8 gold. End result: http://www.talonfiremage.pwp.b…jlkz9txptovsidy8mzvvgn0u8

However, there's also this: http://www.talonfiremage.pwp.b…mql684uludc2bm7l4hi8knrpc
Which is a little bit cheaper still: -6 copper, -2 tin, +2 gold, -12 rubber, -4 redstone

EDIT: Here's another design: http://www.talonfiremage.pwp.b…r76q2ho3ej7hxntknyy4ypvy8
-11 copper, +5 tin, +8 iron, -6 rubber, +2 redstone. It has 5 EU/t more and an efficiency of 2.67 (2.52 after factoring running cost).
That is, of course, unless the mod somehow changes the concept or recipes of multi-cells.

That looks to be an 1.4.7 / GregTech 2.8x design. Is it intended as such?

1.) For every reactor tick, each component is processed once, one after the other. It starts at the top left, the progresses horizontally until the top right. Then returns left one row below and progresses horizontally again. Every component processed gets to do its thing, whatever that may be - drawing heat from the reactor hull or passing it there, spreading heat among itself and/or adjacent components, cooling itself, cooling adjacent components, generating heat and/or EU. That means the order of components plays a role, and also the reason why no two components in a symmetrical design have the same heat state. One of them simply gets to act before the other.

2.) If you remove the middle bottom cell, you remove a 3-neighbor heat source, drop another 3-neighbor down to 2 and a pair of 2-neighbors down to 1. That's a massive reduction in heat generated, and thus several of the cooling components simply have nothing to do anymore. When their turn comes, there is no heat left to dissipate/spread/transfer etc because the components further up who get to act first already took care of everything.

3.) Correct

4.) The component heat exchanger balances heat level between itself and surrounding components. If you remove the top right vent, the other two vents adjacent to the exchanger start overloading themselves due to the reactor now generating more heat than there is cooling capacity for. And the exchanger will dutifully balance itself with them. It will make sure its own heat% is as similar as possible to that of the vents, which are making a beeline for 100%. Thus the exchanger is also making a beeline for 100%. But because the vents melt first and the simulation stops there, it doesn't quite go all the way.

5.) No, each OC vent draws the full 36, as long as there is heat available to draw. That's why you have the two heat exchangers. They balance the heat level between the five upper OC vents so that all of them are getting an equal share, even though some of them will draw more than they can handle and some of them will have nothing to draw.

6.) The two bottom OC vents incidentally receive heat directly from the adjacent uranium cell. A fuel cell will divide its heat output between all possible adjacent components that can receive heat, if there is at least one. Now, almost all reactor designs place their fuel cells so that they are not adjacent to anything that can accept heat; in that case (and only that case), they dump all their heat to the hull. The component vent in the third-from-bottom middle slot cannot accept heat for example, and neither can other uranium cells. The OC vents however can, and because they're the only adjacent component that can, they get all of it. All 24 that the corner uranium cell in question generates. As it happens, their native cooling rate is 20, and the component vent adds 4, so the 24 heat they receive is exactly compensated. Also, because they are so far down in the reactor, there is no heat left to draw from the hull when their turn comes, and thus they don't melt themselves.

7.) Also unsure as to why that is, but it's an effect you can observe in almost all reactors. Most components will keep a baseline of a few dozen heat points before the adjacent exchangers and such start cascading the heat away. It depends on how many steps the heat must move, because there needs to be a small differential at every step. It's not really important though, you'll find that the components stabilize themselves in a certain state eventually and, after switching off the reactor, cool back down within a couple seconds at most.

7a.) Heat not dissipated or transferred away adds up every tick, yes. But as said above, sometimes you first need a certain treshold before the transferring away really starts working properly.

Quote from MinecraftRedstoner

But it should output 180 Eu/t. And the AESU get only 13940 or so Eu instead of 14400.

Measure for more than 11 ticks with your EU meter. You're not getting accurate readings that way. Nuclear reactors only output once per 20 ticks.

Try starting the measurement, then wait 30 seconds, then take the reading.

Since you aren't showing or explaining what's going on with your AESU, I can only guess as to the cause. The reactor output looks normal, at least.

I'm pretty sure it worked when I tried it in the cube. That said, I've found two different bugs in the cube in the past, so the only 100% trustworthy test would be a live reactor.

Also, replacing a component vent with a basic exchanger may not be worth it. You pay 6 iron and 1 tin less, but in return need to invest 11 copper, 6 rubber and 2 redstone more. I'd consider that a const increase, all in all. Component vents really are surprisingly cheap in comparison to most other components

Same output over time in 3.07, but the thorium cells need to be replaced twice as often.

Also, heat for thorium has gone down 20% compared to what the planner shows. You now only get 12 heat for a single thorium cell with 4 neighbours, which is very convenient because 12 happens to be the cooling capacity of two regular vents (or one advanced vent), and also the transfer capacity of a regular heat exchanger. Taking advantage of that, here is a reasonably-priced max size single-cell breeder that doesn't hurt you when you approach it, and here is a 4-chamber design that still manages the full 64 heating cell stack.

Redstoner's design with the five thorium cells will only need 32 instead of 39 cooling in 1.5.x versions, so Requia's variant can drop the extra heat exchanger. That still leaves it 4 cooling over the needed amount, but you can't drop a component vent or the overclocked vent will melt itself.

You can even set IC2 uranium to be much more common if you wanted to, especially if you let another mod (like TE/CoFH Core) generate copper and tin. Then uranium remains the only ore generated by IC2, and you can go into the config and set the ore density multiplier to something like 2 or 3. That'll give you plenty of uranium.

UE was never big about game balance in general... :p

I'm not ignoring the 3 out of 8 cells returned as depleted isotopes from the centrifuge... I never factor in 8 cells per centrifuge run to begin with.

My math assumes that you insert 5 re-enriched isotopes into the centrifuge and get out 1 plutonium and 4 thorium. True, in practical application you have to insert 8 isotopes and get a side product of 3 depleted ones, but those can be re-bred indefinitely. Therefore you don't actually lose or gain any reactor fuel, and the whole thing can be safely ignored. The only thing that counts is that through some process, the details of which are unimportant, the centrifuge converts exactly 5 isotopes into exactly 5 fuel cells with no loss beyond the cost of 100k EU per centrifuge run.

I'm also not factoring in breeding fuel costs because really, silktouching a single coal vein and throwing the blocks into an indsutrial grinder will give you enough thorium to breed hundreds upon hundreds of isotopes without ever touching your regular reactor fuel at all. Breeding is essentially free of charge beyond the initial build cost of the breeder reactor.

And yeah, that centrifuge recipe is very useful. Together with the fact that the industrial grinder will double uranium output from ore plus guaranteed plutonium dust, and gives guaranteed thorium dust from coal ore (you version's mileage may vary), you can still get a decent amount. You do however invariably end up with uranium cells that you need to run through a regular reactor.

Not much of a fan of that metric because it rolls breeding gains into the number, and thus isn't comparable to standard IC2 reactors... you can basically only rate GregTech reactors with that, and then it'll confuse people into thinking they can compare the numbers to pure uranium.

Admittedly isotope efficiency isn't the most straightforward of models, especially with multi-reactor systems, but I wanted to avoid any confusion and keep everything fully comparable. That's why I opted for a metric that exactly returns the reactor planner numbers for pure uranium reactors while also appropriately describing GregTech fuels with the exact same formula. I think that a little complexity is a fair price to pay (but I suppose that as long as there is no automated tool like the reactor planner that reports it, it won't really see widespread adoption).

Just for reference of other readers who may not have seen my previous post, here's how you would calculate isotope efficiency for the multi-reactor system presented in post #364:

First, figure out how many isotopes you require to craft enough fuel for one full cycle. Pick one reactor to start with, usually the plutonium reactor. This example runs one dual uranium cell and 2 quad plutonium cells, and due to lifetime differences, the dual uranium cell must be replaced once per cycle. You effectively have two dual uranium cells, for a total of four uranium and eight plutonium. 1 uranium is 1 isotope, but 1 plutonium is 5 isotopes because the centrifuge turns 5 isotopes into 4 thorium and 1 plutonium. Eight such centrifuge runs thus cost 40 isotopes. Together with the 4 for uranium, that's 44 isotopes to stock the plutonium reactor.

Is this enough to also stock our thorium sink? The runtime for the plutonium reactor is 20,000 seconds, and for each such cycle we have 32 thorium left over. The thorium sink reactor runs for 25,000 seconds, which is one quarter more. During that extra quarter, the plutonium reactor generates an extra 8 excess, for a total of 40. Since we only need 36, we don't need to spend any extra isotopes here.

And now you can also see why explaining things to people is a good thing, because I just noticed that I made a mistake earlier. God dammit why am I so sloppy

Second, determine the total EU generated for one cycle. 360 EU/t for 20,000 seconds, plus 204 EU/t for 25,000 seconds: 144 million + 102 million = 246 million.

Eight runs of the centrifuge cost 800,000 EU, subtract that: 245.2 million. This is also the point where you would subtract running costs for your reactor if you paid them in UU-matter, but since this is a GregTech reactor and thus subject to GregTech's UU-matter costs, that is a pretty silly proposition, so I'll skip it here.

Finally, divide your EU total (in millions) by your invested isotopes: 245.2 / 44 = 5.5727272727272727...

This number is your isotope efficiency. Or, in other words:
"At what efficiency would you be if you generated this much EU out of this number of isotopes using just a pure uranium reactor."

Third, check if there's anything left: our reactor system only consumes 36 out of 40 thorium every 25,000 seconds. The system is thus "+4T/25k" thorium positive. That thorium is free energy we're not yet taking advantage of, so we might want to do that.

Since 4 thorium is 1/9th of the full 36 load of the thorium sink, we'll just say that these 4 will, at some unkown point in time when we have run many cycles and amassed enough thorium to let the sink run a bonus cycle, give us roughly 1/9th the energy output of a full sink cycle. So 102 million EU / 9 = 11.3333333... million EU. We get to add this number to the total EU yield of the reactor system, which rises from 245.2 to 256.533333...

And then just repeat the division: 256.53333333... / 44 = 5.830830830...

Therefore, by utilizing all the thorium, the final isotope efficiency of the system is 5.831.
Sorry guys, not 6.474, I blame food coma for inventing an extra 10 thorium that wasn't there.

Nope, plutonium runtime didn't change - it's still 20k seconds, and 25k for thorium, and 10k for uranium.

Basically one plutonium cell equals 5 thorium cells to be neutral (as seen in the last reactor I linked). In 1.4.7 this was 1 to 10 instead.

That total output would be down was a given though, considering the unintended +75% output 'feature' for single plutonium/quad thorium cell pairs was fixed.

Same as in 1.4.7, just without hybrid effects, and slightly less heat (and thorium got halved in runtime). Look it up in my spreadsheet

Sample 3.07i multi-reactor system - one uranium/plutonium hybrid, one 3x3 thorium sink:

cooling system - 720 heat / tick
http://www.talonfiremage.pwp.b…ovydkz4gvw4bbb961u7t8vqio
ingame setup - 720 heat, +40T/20k, 360 EU/t
http://www.talonfiremage.pwp.b…o2w0unrhtmzv9tetav6ooj11c

cooling system - 572 heat / tick
http://www.talonfiremage.pwp.b…g1879k7e27a0n5hl4ua52zw8w
ingame setup - 550.4 heat, -36T/25k, 204 EU/t
http://www.talonfiremage.pwp.b…bgqhitpiwqtaui1pj892xc000

Total system: 5.573 isotope efficiency, +14T/25k, 564 EU/t
Or: 6.474 isotope efficiency, thorium neutral when running all thorium cycles (either through pausing the plutonium reactor/waiting for fuel, or by building a second thorium sink which runs once every ~3 cycles)

Cooling system on the thorium sink might need some tweaking.

EDIT: For the sake of comparison, here's a singular thorium neutral plutonium/thorium hybrid:

cooling system - 558 heat / tick
http://www.talonfiremage.pwp.b…b8tl8i5r372z12mdth8kmax34
ingame setup - 550.4 heat, 5.06 Isotope efficiency, 254 EU/t
http://www.talonfiremage.pwp.b…lhxmzix6lvgomf7m9vmr904jk

That's more than 1.4 efficiency less than the pair above, and even building 2 of these won't give you as much EU/t.

Which setups? Don't have any built in my SSP world yet.

With the multi-reactor systems, you can get isotope efficiencies >6 without crippling your EU/t. Let me throw something together after dinner to demonstrate.

Well, it certainly is a possibility, considering I only ever use SSP for creative mode testing purposes.

Will test again and report back later.

EDIT: Yep, that's it. Works fine with handcrafted isotopes. Good call, Geggo!

Not only that, he made it more or less mandatory to have a stage 1 reactor to burn through uranium cells in order to get depleted isotopes for breeding towards a stage 2 plutonium/thorium reactor, since you can no longer produce depleted isotopes for breeding and centrifuging on demand...

In my current world I currently have two of these (from this very thread) running, the breeder and stage 2 reactors are still in planning (mostly because of thorium/plutonium changing so often recently).

As a side effectof the whole shebang though, I figured out that neat isotope efficiency metric which is really darn useful when comparing all those different fuel types (I need to make a better/shorter how-to guide sometime). And in combination with looking at what the CRCS guys were doing, I also discovered the secret of multi-reactor systems. Basically, it seems so far that having a (mostly) pure plutonium reactor combined with one or more "thorium sink reactors" that simply burn huge groups of quad thorium cells at once will yield far better results than any single plutonium/thorium hybrid reactor can offer.

Yep, this happened. For 1.5.x, Greg introduced the plate bending machine, which makes plates out of all metal ingots. All multicells now use copper plates (costing 1 copper bar) instead of dense copper plates (costing 8 copper bars). You also have the option of using lead plates (costing 1 lead) in place of the copper plates if you run at least 3.05c. This is convenient because I tend to have way too much lead lying around.

However no GregTech buff comes without a nerf, and 1.5.2 versions of GregTech now disable the 1 uranium + 8 cells = 8 depleted isotopes recipe by default (can be turned back on in the config).

The problem with finding new GregTech designs currently lies with the fact that Greg changed something about thorium/plutonium with almost every subversion update since his 1.5.x port. Significant changes, too - he trialed a multipulse system that would make plutonium pulse multiple times per tick and thorium less often than once per tick. But it ended up causing a ton of bugs, which he tried addressing in numerous updates but ultimately ended up reverting to default behavior in 3.07. He's now set thorium/plutonium to a functional standard behavior with the footnote of "I am going to completely replace the whole IC2 nuclear reactor mechanic with something of my own in a future which may or may not be distant" (obviously paraphrased). They're basically the same values as in 1.4.7, except with a 20% reduction in heat output for thorium and a 11% reduction in heat output for plutonium. Also no hybrid effect obviously.

You can see a selection of the various states of being the fuel cells went through in the spreadsheet I linked in various places, last in post #350.

Currently I assume there are four distinct groups of people playing GregTech:
- one group that stays on stable 1.4.7 versions up to 2.8x, for which the existing list of GregTech reactors is relevant
- one group that plays FTB Ultimate in the latest version, which foolishly incorporated the buggy testing build 2.90h, which has GregTech reactors in so bad a state that they're barely even functional
- one group that plays either the FTB 1.5.2 beta pack or one of the various semi-public homebrew 1.5.2 packs people built in the absence of FTB updates, for which 3.05 figures with multipulse rules are relevant
- one group that constantly updates to the latest version, for which the 3.07 values without multipulse rules are relevant

Something else to note on the "current state of reactors": there appears to be a bug in 1.5.x versions of IC2 that makes breeders in SSP instantly charge depleted isotopes in a single tick. It doesn't happen in SMP, and all signs point to it being a basic IC2 bug, not related GregTech (but I still need to confirm it properly as all my modpacks include GregTech in some form).

Yep, there's a bug with thorium in GregTech 3.05g, it gets fixed in 3.07 I believe.

Good to see that you made the same observation with the instant breeding though. It seems to be an IC2 issue and not related to GregTech, do you agree?

I am looking for someone who has built and operated a breeder reactor in a Minecraft 1.5.x environment, because I think there's a serious bug afoot and I need to compare people's experiences with mine.

Anyone done so yet?

Mostly Shneekey's post #336 comes to mind. Welcome back!