Help with choosing balance between belt and bots transportation in a blueprint

[suxkar]

Hello!

I'm planning to build a 5k science "megabase". I've just started designing the various blueprints. I'm trying to make them as efficient as possible while still retaining some creative freedom (eg random spaghetti is accepted). For example, this is the basic block for green chips production:

Screenshot (19).png (5.14 MiB) Viewed 1832 times

Raw ore is unloaded by train in the south and feeds Nilhaus style smelters, which produce 3 full belts of iron and 3 full belts of copper. Thanks to productivity modules, I can put one extra smetler for each copper column. 3 iron belts and 3+copper belts are enough for 4 chips columns, resulting in 4 full belts of green chips.
In the picture you can see how i give priority copper and iron feed to the 2nd,3rd an 4th chips columns. The first one is fed by combining the leftover iron and copper, plus the copper from the extra copper smelters.

Considering that i have to repeat this build 12 times, my concern is about UPS:
1) How bad is this belt spaghetti gonna be along with everything else ups wise?
2) Would it be better to just build 2 smelters for each chip instead of optimizing with 6 to 4?
4) Would it be better to use bots? Bots for everything? Or may be only for the "leftover" chips column? How do I understand how many are "too many bots" for UPS?
5) If I keep I similar style with my other setups, will my base be playable?

Since I'm already asking,
6) how bad is it to use bots with miners instead of belts?

Thank you very much in advance!

[Guenni7]

1) How bad is this belt spaghetti gonna be along with everything else ups wise?
2) Would it be better to just build 2 smelters for each chip instead of optimizing with 6 to 4?
4) Would it be better to use bots? Bots for everything? Or may be only for the "leftover" chips column? How do I understand how many are "too many bots" for UPS?
5) If I keep I similar style with my other setups, will my base be playable?
6) how bad is it to use bots with miners instead of belts?

The last weeks I changed my complete iron/copper-smelting from bot-only to belts.
My observation is, that the ups gain from less bots is more than eaten up by the inserters-to-belt ups loss.
Inserters to boxes seems just highly more effective than inserters to belts.
Yesterday I started to build back to bot-only
The golden rule for bot networks is: Build small seperated networks, just enough bots in as needed (My biggest networks have only 500 bots).
As for bot-mining: Works perfect IF you use the ore nearby, as loading into a train or smelter array.
And always have a look at the ways bots have to go. Small changes can have huge impact, an example:

O = Ore Patch, T=Train loading Plates, S=Smelter
a)
TTTTTTTTT
SSSSSSSSSSS
SSSSSSSSSSS
SSSSSSSSSSS
OOOOOOOO
OOOOOOOO

b)
SSSSSSSSSSS
SSSSSSSSSSS
SSSSSSSSSSS
TTTTTTTTT
OOOOOOOO
OOOOOOOO

b) needs ~30% less bots than a)

[DaveMcW]

1) Spaghetti is great, underground belts and turns are highly optimized. What hurts UPS is side merges and splitters.
2) Every "thing" takes about the same amount of UPS. So using 14 splitters to save 2 furnaces and 4 inserters is a bad trade.
4) Bots are worse than full belts. The only time you should consider bots is if you are not producing enough to come close to filling a belt.
5) The game is so optimized that a 5k megabase should be playable even with some inefficient choices.
6) Who needs bots, belts, or inserters? Just mine directly into the train!

[suxkar]

Thank you!

6) Who needs bots, belts, or inserters? Just mine directly into the train!

How do you direct-load a train efficiently? Aren't long trains hard to fully load? Do you use multiple train stops to fill all wagons?

[NotRexButCaesar]

How do you direct-load a train efficiently?

Mining productivity research

Do you use multiple train stops to fill all wagons?

Yes, that is the most common way

[astroshak]

How do you direct-load a train efficiently? Aren't long trains hard to fully load? Do you use multiple train stops to fill all wagons?

You can put 4 miners next to a single cargo wagon, two on either side. 40 stacks, 50/stack, you can carry 2000 ore per wagon. 2000/4=500 ores per miner.

Miners output 0.5 ore/sec. Boost that with mining productivity tech, speed modules, and speed beacons. 3 SM3’s in a mine will decrease that cycle time of 1 second to .4 seconds, resulting in .5/.4=1.25 ore/sec. That would be 400 seconds to fill a cargo wagon with those 4 miners.

Mining Productivity tech would increase the output by 30% (MP3) or more. 1.25+30%=1.625ore/sec. This makes it about 307 second fill time for that cargo wagon. Speed Beacons, and space science level MP techs, will boost mine output further, reducing the fill-time.

Per https://factoriocheatsheet.com/#cargo-wagon-transfer it would take 12 seconds for a single side set of 6 stack inserters to load a cargo wagon with ore. 500/12=41.6667 ore/second miner output necessary to equal the six stack inserters loading from a chest. 41.667 / 1.25 = 33.3, so you’d need a tech level of 333? Obviously this will be lower with Speed Beacons (the 1.25 came from having modules already in the miners, not an unreasonable assumption if you are trying to directly load the train).