Whats the deal with beacons?

[Templar]

Does using beaconed assemblers/furnaces save you on UPS vs the same through put of just having more assemblers/furnaces? I am using bots, so I do understand how it would help with belts but with bots it would create more entities then it would save so wouldn't that be bad? My plan would be to use all speed moduals in the beacons and assemblers/furnace as resources are no issue and neither is space.

[Synical]

Beacons essentially let you transform power into "free" items if done correctly. A single T3 assembler making a processing unit will produce 7.5 per minute while consuming 210 kW. In that time, you would have consumed 14.4 mJ, so that comes out to about 1.92 mJ per item.

A T3 assembler with 4 T3 productivity modules and surrounded by eight beacons with 2 T3 speed modules will produce 46.2 per minute while consuming 2 mW. Of course, we need to factor in the beacons; in my typical tiled setup, there's around a 1.6 ratio of beacons to assemblers (too lazy to calculate the actual convergence), so we'll add an additional 480 * 1.6 kW per assembler, for a total of ~160 mJ spent per minute. So, that comes down to about 3.46 mJ per item. We've eaten less than twice the energy for more than 6 times the items per minute.

Of course, there are initial setup costs and concerns about return on investment, but usually once I'm considering building like this, I'm at the point where the 130 R, 100 G, 30 B chips needed to make a module is not a large hit from my production. And sure, it spews pollution like no one's business, but biters stop being a threat once you have a nuclear arsenal that would rival the USSR.

And while you could just ignore beacons completely and just build more assemblers (around 6.16x more if you wanted to match), but that means more belts or travel time for your bots, and though space is infinite, keeping things compact is important if you want to reduce latency and keep your bots happy.

[Templar]

Power is no issue with me. What I am concerned with currently is raising up my UPS as I go for a RPM+1000 science PM base.

[Synical]

Well, less things operating will give you more headroom on your UPS, and keeping things so compact that you can just ignore belts entirely and use bots will also help. A single "module" of my green chip factory is only about 120x25 tiles yet will produce ~33k chips a minute using only a single row of roboports along the long side to exchange with the adjacent train station.

[Synical]

Also bots are great for UPS. Belts have to do way more processing than a bot, because a bot is just "fly in a straight line to one location and then fly in another straight line." Belt logic is way more convoluted.

[Roxor128]

Originally posted by Synical:

In that time, you would have consumed 14.4 mJ, so that comes out to about 1.92 mJ per item.

Millijoules? That's an amazingly efficient assembler.

Seriously, though, this seems to be the single most common mistake I see made when discussing this game.

SI prefixes are case-sensitive.

m = milli- = 10^-3
M = mega- = 10^6

When you make this mistake, you are out by a factor of a billion. You are being more inaccurate than a creationist who thinks the Earth is only 6000 years old (they're only out by a factor of 750,000).

[Harry]

Originally posted by Roxor128:

Originally posted by Synical:

In that time, you would have consumed 14.4 mJ, so that comes out to about 1.92 mJ per item.

Millijoules? That's an amazingly efficient assembler.

Seriously, though, this seems to be the single most common mistake I see made when discussing this game.

SI prefixes are case-sensitive.

m = milli- = 10^-3
M = mega- = 10^6

When you make this mistake, you are out by a factor of a billion. You are being more inaccurate than a creationist who thinks the Earth is only 6000 years old (they're only out by a factor of 750,000).

Take this "Like" since there is no button for it^^.

[Templar]

So at this point I am very unimpressed with beacons...I created beaconed setups for smelting and circuit production and tore down my old infrastructure and started replacing after doubling my power capabilities. I started out at about 60 UPS and before I could even get all the way to my previous capacity I was down to about 26 UPS and was using a starteling large portion of my new power as well.

[Killcreek2]

@Templar
Yeah, beaconed setups are a BIG power drain.

What I do is have them operating in pulse mode: A belt counter on each output linked to the power switch connecting that assembly/furnace line to main power grid. If the belt is not full it powers them up, & if it is full they go to power-save mode.

[Templar]

I would be cool with it if it would have helped with my UPS issues but it made it worse instead of better at even less capacity..

[Killcreek2]

Well, if the arrays are shutting down regularly, then there is almost nothing for the cpu to process from them during those times. No power = no activity = (almost) no impact on UPS.
Might be worth a try to see if it helps.

Just to check: you are running 4x prod3 in assemblers, 2x in furnaces, with speed3 in the beacons nearby [as Synical advised]? Not just speed3 in everything, as your first post indicated.

[Lothos]

fwiw, i was under the impression if your factory, regardless of beaon/no beacon design, would hinder UPS if it started/stopped on a frequent basis. That a steady running factory was smoother on the UPS than a burst one.

[Killcreek2]

Originally posted by lothos:

... That a steady running factory was smoother on the UPS than a burst one.

That is the ideal goal, for sure. Harder to actually put into practise on a large scale, imho.

Example: I have a train-bus factory, so my main iron smelter often has 5+ trains arrive at once all wanting plates, instead of at regularly spaced intervals, as they deliver to different destinations that take different times to use up those plates.
The smelter must be beefy enough to supply such a burst in demand, but then what does it do during a lull in the outbound trains?
I simply have each smelting line go to sleep once it fills its belt, as calculating the average required plates /sec to build exactly the right number of furnaces, and calculating the exact size of the train buffer to allow for burst-loading but not require burst-crafting is a lot of rather dull math I would prefer not to do. [& gets even more complex when you factor in different research flasks that may or may not be needed, eg; military].

I have the same power controls at the green & red chip assemblers too. As each smelting line / chip line is power-controlled separately, it rarely has everything running at the same time.
A few temporary hits to the UPS due to spikes in demand, is an acceptable compromise in my book. ;)

[Lothos]

i had a similar issue with iron and honestly i found it better to use longer ore trains than the pickups....no more burst runs on plates.

[Killcreek2]

I have at least 1 ore train queued up in the stacker waiting for a free station at all times, so longer ore trains would not make any difference there unfortunately.
The burst runs are due to the plate output trains arriving at random times, often in groups, whereas the ore dropoff is silky smooth clockwork.
Idk how to "fix" it without some complex combinator train scheduling &/or drowning in the math.

[Lothos]

do you not have enough buffer storage for plates? I've seen some add a second row of chests/inserters to increase the output buffer from smelters. Might help a bit, but you'd also add more load.