It's nice when things come together [utitility-science-pack]

Clever and beautiful constructions, bigger than two chunks
- Defense: killing biters as an art
- Castles, Throne Rooms, Decorations (comfortable living in the Factorio World)
- Main Bus Concepts
- Modular Systems, Factory Streets, show how all works together
- Megabases
Please provide us with blueprints or saves, if that makes sense of course.
Forum rules
Clever and beautiful constructions, bigger than two chunks
Post Reply
mrvn
Smart Inserter
Smart Inserter
Posts: 5709
Joined: Mon Sep 05, 2016 9:10 am
Contact:
Contact mrvn

It's nice when things come together [utitility-science-pack]

Post by mrvn »

This came together quite beautifully to produce 1.5 utility science packs per second:

utility-science-pack.png
utility-science-pack.png (1.15 MiB) Viewed 1386 times
This uses a lot of copper plates. I might have to add a second copper train station to bring the train throughput more in line with other stations.

I like how the ratio of copper wire, green circuits, blue circuits, utility science packs and flying robot frames at the bottom line up near enough so they are all just piped through.
Top
mmmPI
Smart Inserter
Smart Inserter
Posts: 2747
Joined: Mon Jun 20, 2016 6:10 pm
Contact:
Contact mmmPI

Re: It's nice when things come together [utitility-science-pack]

Post by mmmPI »

It is very beautiful indeed.
Top
mmmPI
Smart Inserter
Smart Inserter
Posts: 2747
Joined: Mon Jun 20, 2016 6:10 pm
Contact:
Contact mmmPI

Re: It's nice when things come together [utitility-science-pack]

Post by mmmPI »

I have been playing with this blueprint this WE and i found out, although it is very beautiful there are quite some ways to improve it :

1) This blueprint doesn't use module instead different tier of assembly are necessary to achieve ratios, this means it doesn''t scale well in the future.

2) This blueprint waste lots of room by not using underground belts and instead placing power poles between assembly. Though not enough that it will be easy to add beacons.

3) There is no need for 2 rows of fast long-handed inserter per warehouse, only 1 is enough, or some stack inserters, it's not like 1.5 science per minutes require tremendous amount of material.

4) The warehouses themselves are not necessary, they can be replaced by chests easily for space gain. ( they are used for only a single ressource and throughput of train was suboptimal see (6).

5)The belt based balancer is 4=>4 balancer and there is only 2 red and 1 yellow belt of output but 4 red belts of inputs. It is not necessary to use a balancer in such case it is not doing anything more than the wire already connected to the belts to prevent the loader to empty the warehouse if case of imbalance. Just 3 priority splitters is enough.

6)There is no room for a second train behind the one that is unloading and no good signaling on the right side to allow train to smoothly take turns, i improved improved by a lot the thoughput by adding signals, i can't tell exactly how much, that would require precise testings, but it didn't feel like adding another copper station was necessary, one can provide for 60/ 70 item per second with 2 red belts and a yellow belts ,that's not even 2 blue belts for the full train of 4 wagons. It is possible to unload 1 blue belt per wagon, using only 1 side, without too much trouble in the space gained by removing the warehouses.

Now it's not per say an improvement, but i found it a way to avoid wasting ressources, for the station that unload plastic and steel, there is no need for filter inserters, those are expensive, and all 4 warehouses on each side are containing the same item, it's possible to send only steel or plastic trains, and as such reading the content of the train from the train station with a red wire and only activating one side or the other. Because I couldn't see an absolute need for sending trains that contain the proper plastic/steel ratio instead of alternating full steel and full plastic thanks to LTN making the frequency of each delivery varying but that may be because i haven't seen the plastic or steel production area.

I hope you don't mind me giving a few advices back as you often do when i post some creations, in case feel free to ignore my attempt :)
Top
mrvn
Smart Inserter
Smart Inserter
Posts: 5709
Joined: Mon Sep 05, 2016 9:10 am
Contact:
Contact mrvn

Re: It's nice when things come together [utitility-science-pack]

Post by mrvn »

mmmPI wrote: ↑
Mon Mar 25, 2024 9:36 am
 I have been playing with this blueprint this WE and i found out, although it is very beautiful there are quite some ways to improve it :

1) This blueprint doesn't use module instead different tier of assembly are necessary to achieve ratios, this means it doesn''t scale well in the future.
Didn't have beacons and very few modules. It's not mend to scale, just work for now.
mmmPI wrote: ↑
Mon Mar 25, 2024 9:36 am
 2) This blueprint waste lots of room by not using underground belts and instead placing power poles between assembly. Though not enough that it will be easy to add beacons.
The width is determined by the train length and the assemblers fit with the gaps. So there was no need to pack them more densely. It's not enough to add beacons later and that would not work anyway with the belt speed and completely changes the ratios.
mmmPI wrote: ↑
Mon Mar 25, 2024 9:36 am
 3) There is no need for 2 rows of fast long-handed inserter per warehouse, only 1 is enough, or some stack inserters, it's not like 1.5 science per minutes require tremendous amount of material.
That depends a lot on the inserter bonus. The game was a long time ago and since i used fast/filter inserters instead of stack/stack filter inserters that indicates it was early in the game. So maybe the double row of inserters was actually needed to keep the trains moving fast enough for the science production. Or if not for the science production it frees up the trains twice as fast so they can be retasks to service some other LTN request. Fewer trains needed overall.
mmmPI wrote: ↑
Mon Mar 25, 2024 9:36 am
 4) The warehouses themselves are not necessary, they can be replaced by chests easily for space gain. ( they are used for only a single ressource and throughput of train was suboptimal see (6).
Then you have to balance 24 chests instead of 4 warehouses. It's also just simpler using a single station blueprint even if you could save a few concrete by using chests here and there.
mmmPI wrote: ↑
Mon Mar 25, 2024 9:36 am
 5)The belt based balancer is 4=>4 balancer and there is only 2 red and 1 yellow belt of output but 4 red belts of inputs. It is not necessary to use a balancer in such case it is not doing anything more than the wire already connected to the belts to prevent the loader to empty the warehouse if case of imbalance. Just 3 priority splitters is enough.

6)There is no room for a second train behind the one that is unloading and no good signaling on the right side to allow train to smoothly take turns, i improved improved by a lot the thoughput by adding signals, i can't tell exactly how much, that would require precise testings, but it didn't feel like adding another copper station was necessary, one can provide for 60/ 70 item per second with 2 red belts and a yellow belts ,that's not even 2 blue belts for the full train of 4 wagons. It is possible to unload 1 blue belt per wagon, using only 1 side, without too much trouble in the space gained by removing the warehouses.
3 priority splitters would work. Wouldn't change anything beyond saving 3 splitters though.

In that game all stations are layed out in columns the width of a train + curves and scaled to require a train maybe every 5-10 minutes. The copper stations needs far more trains than the other stations and is an outlier in the design. Hence why I said "to bring the train throughput more in line with other stations".
mmmPI wrote: ↑
Mon Mar 25, 2024 9:36 am
 Now it's not per say an improvement, but i found it a way to avoid wasting ressources, for the station that unload plastic and steel, there is no need for filter inserters, those are expensive, and all 4 warehouses on each side are containing the same item, it's possible to send only steel or plastic trains, and as such reading the content of the train from the train station with a red wire and only activating one side or the other. Because I couldn't see an absolute need for sending trains that contain the proper plastic/steel ratio instead of alternating full steel and full plastic thanks to LTN making the frequency of each delivery varying but that may be because i haven't seen the plastic or steel production area.

I hope you don't mind me giving a few advices back as you often do when i post some creations, in case feel free to ignore my attempt :)
Getting LTN trains to ship multiple items on purpose reliably is very hard. But getting them by accident is very possible. I have a garbage dump where I send trains with leftover items or when I deconstruct stuff. And if there is steel and plastic in the garbage dump then LTN might just order a train containing both.

If I were to change the steel/plastic station I would reduce it to just 4 warehouses storing both items. Then you really don't need any filter inserters to unload. But then you need filter inserters to fill the belts instead of using loaders.

Anyway, I didn't post the picture because the train stations where so beautiful....
Top
mmmPI
Smart Inserter
Smart Inserter
Posts: 2747
Joined: Mon Jun 20, 2016 6:10 pm
Contact:
Contact mmmPI

Re: It's nice when things come together [utitility-science-pack]

Post by mmmPI »

mrvn wrote: ↑
Mon Mar 25, 2024 12:04 pm
 Didn't have beacons and very few modules. It's not mend to scale, just work for now.
Ah that make sense, i was in editor mode so it was easy for me to have the modules and beacons to make some improvements. It found it helps a lot as yellow science is expensive and late game enough that i don't build a version without module and another one different later like i would do for green or red science.
mrvn wrote: ↑
Mon Mar 25, 2024 12:04 pm
 The width is determined by the train length and the assemblers fit with the gaps. So there was no need to pack them more densely. It's not enough to add beacons later and that would not work anyway with the belt speed and completely changes the ratios.
Yeah i prefer having room for 2 trains everytime, so that when the first one leave, there is another one just behind, and then the 3rd can be late it's ok. Because if the trains are scheduled to arrive every 5-10 minutes, but you need 2 row of inserters to unload them fast enough for the rest of the factory not to starve, i found it's risky. A bit of congestion can add 10 to 15% to the travel time, say, and then it would break, i found it's better to have more margin.
mrvn wrote: ↑
Mon Mar 25, 2024 12:04 pm
mmmPI wrote: ↑
Mon Mar 25, 2024 9:36 am
 3) There is no need for 2 rows of fast long-handed inserter per warehouse, only 1 is enough, or some stack inserters, it's not like 1.5 science per minutes require tremendous amount of material.
That depends a lot on the inserter bonus. The game was a long time ago and since i used fast/filter inserters instead of stack/stack filter inserters that indicates it was early in the game. So maybe the double row of inserters was actually needed to keep the trains moving fast enough for the science production. Or if not for the science production it frees up the trains twice as fast so they can be retasks to service some other LTN request. Fewer trains needed overall.
Sorry i meant second not minutes, 1.5 science per second is bigger than 1.5 per minutes x).
I realize it's a problem outside the blueprint now, copper station require a train every 4 minutes, and if there is delay on trains transit it would require a 3row of inserter or a 4rth to free them fast enough by reducing unload time. There is a limit obvisouly to the time you can save like this , at this point it's just better to add trains i think, as such the blueprint can stay as is but adding copper station seems risky if there is not enough trains for a single row of inserter though.
mrvn wrote: ↑
Mon Mar 25, 2024 12:04 pm
 In that game all stations are layed out in columns the width of a train + curves and scaled to require a train maybe every 5-10 minutes. The copper stations needs far more trains than the other stations and is an outlier in the design. Hence why I said "to bring the train throughput more in line with other stations".
Well to me the copper station require a train every 4 minutes, it can't be more than that as there's not enough belts. But with 2 rows of inserters it goes super fast to unload, so i'm not sure why considering adding another station and not another trains , that's obvisouly trains the shortage if you need 2 rows of inserters to free them fast enough for the rest of the factory not to starve, adding stations would not solve the problem.

mrvn wrote: ↑
Mon Mar 25, 2024 12:04 pm
 Then you have to balance 24 chests instead of 4 warehouses. It's also just simpler using a single station blueprint even if you could save a few concrete by using chests here and there.
I have nothing against mods that make the game easier, i just thought of it as an improvement toward the not so simpler version because it solve an additionnal problem, if you have 24 chests, that makes 12 belts where you have 1 inserter on each side, 12 belts can easily combine into 6 using splitters, and since your outpout is 3 belts, i mean it's fairly doable without warehouses to have something that balances itself there, since it's only for science , and you can't add modules or beacons because it's inside rails, it's unlikely the ratios will change, it would work too.

Using chest makes it easier to use circuits to math the quantity too, because if you have millions of ore in outpost, which you can realistically do only with warehouse, then it makes it difficult to do stack division. To know how much LTN trains can/will carry. Also it is more efficient for ressources to not harvest too much ore, like millions, and put them in warehouse, as harvesting them later with more research, would yield more. As such , it was an improvement for me.
mrvn wrote: ↑
Mon Mar 25, 2024 12:04 pm
 3 priority splitters would work. Wouldn't change anything beyond saving 3 splitters though.
That's the simplest approach that would keep the wire reading the belts and the warehouse content every tick, just to remove the uncecessary splitter is a direct improvement for cost of ressources, the self balancing version is more of a conceptual challenge, like it's not even necessary to use balancer in this situation. It's like an habit i found, if i can share some advices without sounding too much like a prig, i thought you wouldn't mind, to try and spot where it is possible to avoid doing unecessary balancer. Having the wire to prevent warehouse to empty unevenly AND a balancer doing the same thing. You could do one or the other only, in both case it would still be an improvement for the CPU, and it wouldn't change anything.

I know you are yourself a perfectionnist sometimes i'm surprised you just wave that detail like it's nothing. If you save 3 splitters per station, you could build additionnal mining drills or modules, and generate free ressources, so sure you lose 3 splitters worth of ressources now, but in an hour you actually miss what the ressources could have produced if better invested. I haven't done the math, but there is the notion of counpound interest that comes to mind, i wouldn't be surprised if wasting 3 splitters early game, when you don't have stack inserter yet, and no modules or beacon, delay the moment when you can actually start generating infinite ressource with modules and can cause a huge descrepancy over time.

mrvn wrote: ↑
Mon Mar 25, 2024 12:04 pm
 Getting LTN trains to ship multiple items on purpose reliably is very hard. But getting them by accident is very possible.
I have no problem imagining the situation, it happened to me to request 1000 iron, but my trains only had 3 wagons. It was always leaving some leftovers and messing up with the counts until i realize my mistake.
mrvn wrote: ↑
Mon Mar 25, 2024 12:04 pm
 I have a garbage dump where I send trains with leftover items or when I deconstruct stuff. And if there is steel and plastic in the garbage dump then LTN might just order a train containing both.
That's how accident happens !
mrvn wrote: ↑
Mon Mar 25, 2024 12:04 pm
 If I were to change the steel/plastic station I would reduce it to just 4 warehouses storing both items. Then you really don't need any filter inserters to unload. But then you
need filter inserters to fill the belts instead of using loaders.
You can use filter on loaders sometimes, i'm not sure which one you used. It would be more efficient to use shared warehouse that's for sure but that kind of defeat the pupose of using warehouse to simplify logistic if you are going to use them in a complicated way x). I would still note that as an improvement, but for such cases it feel like trivializing things a bit too much for me.
mrvn wrote: ↑
Mon Mar 25, 2024 12:04 pm
 Anyway, I didn't post the picture because the train stations where so beautiful....
I thought you were sharing your plans of adding copper stations to improve the designs, you are severe, beauty is subjective, it's often in the eye of the person who look don't we say ? To me it looks beautiful, when i look at it , i get positive sensation.
Top
mrvn
Smart Inserter
Smart Inserter
Posts: 5709
Joined: Mon Sep 05, 2016 9:10 am
Contact:
Contact mrvn

Re: It's nice when things come together [utitility-science-pack]

Post by mrvn »

mmmPI wrote: ↑
Mon Mar 25, 2024 2:52 pm
mrvn wrote: ↑
Mon Mar 25, 2024 12:04 pm
 In that game all stations are layed out in columns the width of a train + curves and scaled to require a train maybe every 5-10 minutes. The copper stations needs far more trains than the other stations and is an outlier in the design. Hence why I said "to bring the train throughput more in line with other stations".
Well to me the copper station require a train every 4 minutes, it can't be more than that as there's not enough belts. But with 2 rows of inserters it goes super fast to unload, so i'm not sure why considering adding another station and not another trains , that's obvisouly trains the shortage if you need 2 rows of inserters to free them fast enough for the rest of the factory not to starve, adding stations would not solve the problem.
A second station will have a second train loading copper and on the way in parallel. It's similar to having a train limit of 2. The layout of the train tracks just don't allow for a station with 2 trains as limit without blocking the main line.
mmmPI wrote: ↑
Mon Mar 25, 2024 2:52 pm
mrvn wrote: ↑
Mon Mar 25, 2024 12:04 pm
 If I were to change the steel/plastic station I would reduce it to just 4 warehouses storing both items. Then you really don't need any filter inserters to unload. But then you
need filter inserters to fill the belts instead of using loaders.
You can use filter on loaders sometimes, i'm not sure which one you used. It would be more efficient to use shared warehouse that's for sure but that kind of defeat the pupose of using warehouse to simplify logistic if you are going to use them in a complicated way x). I would still note that as an improvement, but for such cases it feel like trivializing things a bit too much for me.
Loaders, as in the game entity, don't support any filtering, only inserters do. Which sounds wrong until you realize that the miniloader mod (and others of the same make) do use hidden inserters. That's also why they work with trains.

I've used both and one has to be careful with what mod one uses because they behave differently. E.g. loaders will fill an assembler to the brink, miniloaders aka. inserter only fill in 2 recipes worth of ingredients (supposedly more for faster machines). For high throughput assemblers you really want the game loaders, not anything using inserters as they take too long to insert items when needed instead of inserting items ahead of time.
Top
mmmPI
Smart Inserter
Smart Inserter
Posts: 2747
Joined: Mon Jun 20, 2016 6:10 pm
Contact:
Contact mmmPI

Re: It's nice when things come together [utitility-science-pack]

Post by mmmPI »

mrvn wrote: ↑
Mon Mar 25, 2024 5:57 pm
 A second station will have a second train loading copper and on the way in parallel. It's similar to having a train limit of 2. The layout of the train tracks just don't allow for a station with 2 trains as limit without blocking the main line.
It's similar but having a better rail layout allowing a limit of 2 allow to spare 4 warehouses and 48 insterters, almost a stack ! I prefer personnaly not having to limit my stations to just 1 incoming train, because then it's only when the currently unloading trains leaves that the next one start its travel, and it can be very far away. Whereas increasing the limits from 1 to 2 allow to have one train unloading when the second one is moving to replace it, or waiting just behind. That's what i would improve first.

It's easier when you continue building i found to add more trains on existing stations, to make up for the increasing distance at which you gather ore, when the starting one deplete, rather than adding a whole station that would be limited to 1 incoming train for fear of blocking the main lane. That works with LTN too, it's not always the case if you turn off stations yourself, but here it would have been possible to have 2 or 3 trains scheduled to deliver at the same place at the same time. It helps to have large buffers with warehouses, such stations with so many inserters and so much buffer look like high-throughput station to me. I think it's overkill for only a train every 4 minutes for copper , and 1 trains every 10 minutes for the others but that's more my personnal opinion i'm giving there, not really an advice.
mrvn wrote: ↑
Mon Mar 25, 2024 5:57 pm
 Loaders, as in the game entity, don't support any filtering, only inserters do. Which sounds wrong until you realize that the miniloader mod (and others of the same make) do use hidden inserters. That's also why they work with trains.
That answer another interrogation, because i thought loaders may have been used to unload from trains to warehouse, but i didn't know which mod you used, ( same for the long handed fast filter inserters btw). People keep making them look nicer and nicer and i can't recognize those i have used anymore. I suppose i would use loader to unload the train into warehouse if i could. Because that's where there seem to be the highest need for throughput. I suppose you have the loaders that are good for high throughput assemblers then but wouldn't work on trains that's unlucky.

I'm still a bit confused because despite your explanations it seem to me not more efficient than using chests, it would require 12 inserters per wagon too, 6 to put in chest, and 6 to put in belt, but that's it, whereas your design require an additionnal loader on top of the 12 inserters.That's one of the thing i changed after adding some room for a second train at the unloading. I'm saying this because the bottom part can not really be improved unless everything is changed to accomodate modules and beacons. But the upper part was to me responsible for the tentation to add another station and it could be improved i thought in a way that would not require more stations. ( additionnal trains do not make the blueprint "larger").
Top
Post Reply

Return to β€œMedium/Big/Gigantic Sized Structures”