Posts by Vyrebird

    Reflector lasts 10k ticks. If it does not, it is a bug. It's a full cycle.

    Be advised: Neutron Reflectors surrounded by multiple cells will diminish faster (2 cells adjacent to the same reflector will deplete it in half of a cell cycle).

    So with Neutron Reflectors adjacent to Quad cells, they'd last 10k/4, or 2500 ticks. Thus why you need Thick Neutron Reflectors.


    I noticed a crazy amount of mark 1 reactors but no mark 2 reactors so far. Anyone has the guts to design a nice one? (meaning it has to be competitive with the mark 1 reactors)

    Mark II reactors rely on storing heat somewhere in order to vent it during the cooldown phase, and there aren't many places to store heat in the current system. 60k coolant cells, over a 10k cycle time, average to storing 6 heat per reactor tick, the same as a normal vent - you do gain slightly with the coolant cells, since you can timeshift the cooling, but that's relatively small. Heat storage in 'normal' components amounts to a small sum as well - going with the generous assumptions of a 50/50 mix of vents and Component Exchangers across 50 reactor slots, that amounts to a 150k heat sink, or about 15 heat per tick.

    So it seems like the 1.106 reactor model heavily favors Mk I, III, and V, with only a very narrow window for Mk II to possibly operate in.



    My potential contribution to the list.
    Mark I, Efficiency 4, 160 EU/t.
    Iron 234, Copper 413, Tin 65, Gold 16 (5-chamber)

    It's higher efficiency than any Mk I currently in the list (aside from the reflector-based eff 7), and it is a different resource balance: lower gold, more copper/iron. I don't know if that's enough to make it useful to others.

    Edit- 4-chamber variant with lower resource costs. Remove the bottom-rightmost heat exchanger to save a couple copper and make it a Mark II-1 ;p

    Iron 215, Copper 364, Tin 64, Gold 16

    Could a "core transfer" field be added in the info panel, along the lines of the Vent Cooling stat? Listing the total heat extracted from the hull (by heat exchangers and heat vents) and the total possible (since exchangers can opt to use less than their maximum).

    I know this would be a more awkward statistic (since heat exchangers can also transfer heat *into* the hull, and I dunno how you'd want to integrate that), but it would be a very very helpful one to me. It's partly for making breeders (since they seem to have greater stability when balanced around core transfer than around cooling), but also just for general reactor design.

    1. Energy output per one piece of ore should be at least 50% of theoretical maximum.

    A quad cell with 4 neighbors yields 140 eu/t, making that requirement translate to 17.5 eu/t per uranium cell (or a bit below 140 eu/t per ore). Assuming the quad cell lifetime bug gets fixed first, you can achieve that rate with a quad cell + single cell: 90 eu/t for 5 cells, or 18 eu/cell, for 172 heat/t.

    Such a reactor can be made relatively cheaply, a Mark I in a 3-chamber reactor using no diamonds and 8 gold (4 overclocked vents). But that doesn't come close to your average output requirement.

    3. Average output of reactor(s) should be at least 300 eu/t. If you breed in same reactor as you get energy from, breeding time is factored in when counting average energy. You can use more than one or two reactors, if needed.

    300 eu/t puts us somewhere between a 2x2 square of 3 quads, 1 single (315 eu/t, 24 eu/cell, 744 heat) and a 2x2 of half quads, half dual (280 eu/t, 23 eu/cell, 640 heat). Or the simple single-cell version (the only one viable in 1.106 from an eu-per-ore perspective due to the quad/dual lifetime bug), a 4x4 square for 320 eu/t (20 eu/cell) with 656 heat. You might be able to manage something a little better with reflectors, but this seems like a reasonable rough number.

    A 6-chamber reactor has 6*9 = 54 tiles. If we consume 16 of those with uranium, we have 38 left for heat management. At 656 heat, that means we need an average of 17.26 heat removal per tile. The only heat removal pieces close to that are component vent (16 heat) and overclocked vent (20 heat). Ignoring edge effects, they'd come out to 18 heat on average, so with edge effects they'll be well below our needed 17 heat. So, this reactor will need to be a Mark III or higher.

    I would love to look at the designs but the links keep redirecting me to some site to download Java when I already have it installed.
    Could someone please post them elsewhere besides the redirect hell that talonfire is.


    If I remember correctly, the applet does require an optional Java package (it also requires at least version 1.6 of the base JRE).

    Just make sure you're downloading the Java-whatnot from ( also looks legit, but I've always used oracle, since that's where the old official redirects to)

    While I haven't tried it, as I understand it external lava does nothing to a reactor's heating. External lava is used by some people, particularly for breeders, presumably because it's an easy way to remove air-cooling without entombing the reactor in solid blocks.

    I've noticed that my breeder seems to slowly, slowly cool over time ( http://www.talonfiremage.pwp.b…wcfhtk8wngho7v5v3kcaooxkw ) as well. Since I don't use a thermometer addon, this is based on the heat bars of the components. I typically run the breeder with all components at 1 'bar' of health. If components are just barely heated to this point (ie they tick to 2 bars instantly with added cooling, but otherwise stay at 1 even after 'settling' into the breeding config), later checkins on the reactor sometimes have some of the components with 2 bars. I know any form of maintenance (fuel swaps, isotope swaps) will cool the reactor slightly, but this has occurred even within a single untouched ~80 minute breeding session.

    I would've expected a 'neutral' reactor would slowly heat over time, due to the water evaporation (if the reactor ticks before the block re-fills, that'd be 1 less cooling than expected). Yet it seems like there is a tiny bias toward cooling instead, which I'm at a loss to explain.