Module Production: To beacon or not to beacon?

[zOldBulldog]

Much is possible, but is there any experienced player consensus?

I originally thought that the standard beacon/module arrangement (productivity in the assembler, speed in the beacon) would be a no-brainer, making things more productive/efficient. But an assembler making modules will not accept productivity modules.

An alternative could be to use speed modules on both and I've seen designs that do just that. But what's the point? Module consumption goes in waves, so I could just go un-beaconed and allow a few hundred to accumulate while I work on other stuff.

So I keep debating with myself without a clear conclusion.

If there is no consensus for a best practice, what rationale did you use when you had to decide? (please skip simplistic things like "didn't consider beaconing module production" )

[astroshak]

I rather Beacon them.

They take a while to produce (using an AM3, the 15 sec for a level 1 is really only 12 sec, but the 30 and 60 sec for levels 2 and 3 become 24 and 48 sec respectively). When I start to want them, I start to want them in a hurry.

[Koub]

You need 5x the number of assembling machines when not beaconed.

For 1 T3 module/s :
unbeaconed needs :
48, 120, 240 assembling machines for modules tier 1, 2, 3
beaconed needs :
10, 24, 48 assembling machines for modules tier 1, 2, 3

Consumption difference may seem big at first, but when integrated in the total energy needed including intermediates and resource gathering, it's a drop in the ocean (over 2.5 GW are needed for the whole production line)

And last argument in favor of beacons : beacons and modules indeed cost a lot, but so do the 326 assembling machines level 3 you can spare by beaconing like there's no tomorrow. So if a high volume of modules is needed, just go for full beaconed setup .

[zOldBulldog]

Thanks Koub.

- I had made a small unbeaconed setup to start, but I will definitely make a beaconed setup for volume.

- It is placed at the end of my malls mini-bus, so that it does not starve my other production lines of circuits. Ironically, it is precisely to fill the smelter and circuit outposts beacons that I initially need beacons for .

[Serenity]

Module consumption goes in waves, so I could just go un-beaconed and allow a few hundred to accumulate while I work on other stuff.

That's only for the ones you put into machines. But you also consume modules for production science (PM1) and rocket control units (SM1)

[bobucles]

Machine modules only boost one machine. Beacon modules boost 8-12 machines. It's a good way of getting more mileage out of the very expensive T3 modules.

There's no need to use beacons for T1 speed modules. A pile of machines will work fine.

[Amarula]

I have just started learning to use beacons, and as Koub says, the increased production for the same number of assemblers, or the same production with drastically reduced number of assemblers, is a compelling argument whether using productivity or speed modules.
Beacons do use a lot of power, so I am using power switches everywhere I have beacons so that when production backs up, the lights go out. Not because I don't have enough power (all thanks and glory to the great god kovarex) just it seems so wasteful to have all those beacons active when the assemblers have nothing to do. As a bonus, a quick glance at the map shows where production is badly out of balance, with multiple idle sectors that stay idle waiting for other sectors to catch up. Now if there was a visual indicator of lights out that wasn't quite so hard on the eyes
I do keep my mall production (everything used for base-building including beacons and module 3s) separate from the science pack production lines, but I am moving to use beacons for all the science pack production as well. I am not at 1KSPM yet but I am heading in that direction

[zOldBulldog]

Module consumption goes in waves, so I could just go un-beaconed and allow a few hundred to accumulate while I work on other stuff.

That's only for the ones you put into machines. But you also consume modules for production science (PM1) and rocket control units (SM1)

I wasn't worrying about those, those go in the large scale science/rocket production lines needed for space science. Those are of course always beaconed.

For just unlocking research I use a small unbeaconed 0.5 (or sometimes less) science per second that I can build early on that is more than sufficient for the purpose.

My question was mainly aimed towards the making the large number of modules needed to scale up smelters and circuit production. Koub answered that one quite well.

[Koub]

I did some more math, and on second thought, benefits of beaconing are not so obvious.

Optimal layout being a mille-feuille one row of beacons with speed modules and one row of assembling machine 3 (with prod modules if the recipe allows it), for high number of layers of this kind of sandwich, the ratio between AMs and beacons gets closer to 1 (this is theoretical for very high numbers of AMs and beacons). In reality, at such scale, we'll have around 1.5 beacon per AM.

So let's assume we can choose between :
408 AM3s with no modules/beacons, totaling 153 MW (375 kW each)
Resources needed to build these AMs (assuming everything else is optimally prod-moduled + speed beaconed) :
4.5k coal
36.2k iron
22.7k copper
89.7k petroleum gas

120x480 kW for the beacons, 82x2.475 MW worth of AM3s consumption, that's 260.6 MW
Resources needed to build 82 AM3, 120 beacons, and 240 speed module 3 :
25k coal
126k iron
165k copper
523k petroleum gas

That's around 5x the resource investment, it's a very significant cost just to spare ourselves some space. I'm not so sure I'd speed beacon recipes that can't be prod moduled any more

[csduff]

The only thing I consistently beacon is copper cable production for electronic circuits. And even then, that is just a +50% total to speed, so it becomes a 1:1 factory ratio.

It is understandable that the resource demands, both materials and energy, are relatively minor for a large base, so modules are an acceptable solution. But for most games, real estate is also a minor issue, so that is also acceptable.

It really comes down to a play style. If you want modules, use them. If you want beacons too, use them. If not, just spread that factory wider and taller.

[Koub]

Actually, for items that support prod modules, it's a win solution (4x prod + speed beacons) on the long term. However for recipes that don't support prod modules (the ones identified as final product), it's far less interesting.

[astroshak]

It really depends on how fast I need my supply to build/rebuild.

If I need more, in a hurry, Speed Module those things. If I don’t, then I don’t. Nuclear Reactors, for example, once I get a bunch up and running I tend to not need more for a while. So those don’t need beacons.

Modules? Once I get to the point I’m beaconing and moduling everything, I want modules built as fast as possible, and that means beacons.

[FredHp]

I try to beacon everything. Production modules + speed modules saves a lot of resources.
The power consumption is almost irrelevant - a 2x2 reactor can power almost everything.... if it isn't enough, i just build a new one.

[zOldBulldog]

I try to beacon everything. Production modules + speed modules saves a lot of resources.
The power consumption is almost irrelevant - a 2x2 reactor can power almost everything.... if it isn't enough, i just build a new one.

I tend to agree that power is irrelevant.
- In vanilla, a 2x2 reactor generates about 480MW, plenty and easy to place multiples.
- In a modded game with an Advanced Electric mod I find that ten 1000x panels and 10 1000x accumulators generate about as much electricity without any worries of UPS impact. I just leave my relatively small modded panels/accumulator lines (just one assembler for each tier of panel/accumulator) running constantly and have an alarm setup to warn me when my power runs low. When power is low... I grab panels+accumulators from the chests, plop them down (or let my bots do it) and go on to continue playing for days until the alarm rings again.

[FredHp]

I try to beacon everything. Production modules + speed modules saves a lot of resources.
The power consumption is almost irrelevant - a 2x2 reactor can power almost everything.... if it isn't enough, i just build a new one.

I tend to agree that power is irrelevant.
- In vanilla, a 2x2 reactor generates about 480MW, plenty and easy to place multiples.
- In a modded game with an Advanced Electric mod I find that ten 1000x panels and 10 1000x accumulators generate about as much electricity without any worries of UPS impact. I just leave my relatively small modded panels/accumulator lines (just one assembler for each tier of panel/accumulator) running constantly and have an alarm setup to warn me when my power runs low. When power is low... I grab panels+accumulators from the chests, plop them down (or let my bots do it) and go on to continue playing for days until the alarm rings again.

The Advanced Electric mod is excellent, but i never use the solar panels because i feel they are too overpowered. I use coal power plants - the boilers and engines in advanced electric are perfect.
Another mod i love is the nuclear fuel. The plutonium requirement for some things adds a nice level of complexity. With nuclear fuel mod, Reactors turns to be a integrated part of industry and not only a power generation mechanism.

[MeduSalem]

Speed beacon them.

Main reason is just so they fit in with the rest of my layout which is beaconized.

Benefit is that they produce way faster thanks to the beacons, hence need to waste less space on that. Only downside is that it consumes energy when idle but that can be helped with a powerswitch hooked to a logistic network check, if one even gives a crap about energy consumption.

The modules necessary for science production are in their own department anyway.

[Selvek]

I did some more math, and on second thought, benefits of beaconing are not so obvious.

Optimal layout being a mille-feuille one row of beacons with speed modules and one row of assembling machine 3 (with prod modules if the recipe allows it), for high number of layers of this kind of sandwich, the ratio between AMs and beacons gets closer to 1 (this is theoretical for very high numbers of AMs and beacons). In reality, at such scale, we'll have around 1.5 beacon per AM.

So let's assume we can choose between :
408 AM3s with no modules/beacons, totaling 153 MW (375 kW each)
Resources needed to build these AMs (assuming everything else is optimally prod-moduled + speed beaconed) :
4.5k coal
36.2k iron
22.7k copper
89.7k petroleum gas

120x480 kW for the beacons, 82x2.475 MW worth of AM3s consumption, that's 260.6 MW
Resources needed to build 82 AM3, 120 beacons, and 240 speed module 3 :
25k coal
126k iron
165k copper
523k petroleum gas

That's around 5x the resource investment, it's a very significant cost just to spare ourselves some space. I'm not so sure I'd speed beacon recipes that can't be prod moduled any more

It would be interesting to see the math for speed module 1 or 2. Since S3 are SOOO much more expensive than S1, there might be a sweet spot at S1 or S2 where the extra cost is more manageable but there are still significantly fewer assemblers required. My first thought would be to use S1 inside all the machines and ditch the beacons. If the no-modules setup has 5x as many assemblers, it is around 2.5 the size of a beaconed setup since the beacons roughly double the size. Using S1 everywhere reduces the number of assemblers to about 55%, so that the space is roughly 1.4x the beaconed setup - and that assumes a perfect beacon sandwich.

[BlakeMW]

It would be interesting to see the math for speed module 1 or 2.

When using speed1, two speed 1 module, multiplied by a beacon to effect 8 assemblers, and with effect halved, is effectively providing a cumulative 160% speed boost, that is it is equal to 1.6 of the machines it is boosting. So if a Beacon + 2 Speed 1 Modules is cheaper than 1.6 of the machines it is boosting (which it is, in the case of Assembler 3), it can be worth doing, if you don't care about the power.

[Xeorm]

If you're looking to be efficient filling the machines with speed 1's is about the same cost (before worrying about the power increase) while reducing the space required. If you're optimizing for space then you'd want to fill the machines with speed 3's before worrying about beacons.

[Selvek]

Ok, here we go. The semi-exhaustive calculation!

I considered assembler 3 setups with each type of modules, then S3 modules and beacons. I found the number of assemblers required to produce one T3 module per second using the factorio calculator at https://kirkmcdonald.github.io/calc.html

Only the assemblers actually producing T1, T2, and T3 modules were included. Here's the screenshot for the S3 modules + beacons case:

Calc.PNG (84.69 KiB) Viewed 8241 times

Using the "Total Raw" costs from the wiki, I calculated the raw ingredients required to build the assemblers, modules, and beacons required for each case. For the case with beacons, I assumed 1.5 beacons per assembler, but your mileage may vary depending on layout.

Sheets.PNG (61.22 KiB) Viewed 8241 times

And here are the plotted results. "Relative space" is the total number of assemblers + beacons.

Plots.PNG (25.57 KiB) Viewed 8241 times

I'll let you all make your own decisions about what makes the most sense, but I think I'm sticking with S1 modules!

P.S. I also considered the S1 + S1 beacon case, which comes out to slightly lower cost and slightly more space than just S1. I didn't consider power consumption.