Smallest possible Kovarex?

Circuit-free solutions of basic factory-design to achieve optimal item-throughput.
Involving: Belts (balancers, crossings), Inserters, Chests, Furnaces, Assembling Devices ...
Optimized production chains. Compact design.
Please provide blueprints!
Forum rules
Circuit-free solutions of basic factory-design to achieve optimal item-throughput
coffee-factorio
Filter Inserter
Filter Inserter
Posts: 336
Joined: Thu Oct 17, 2024 10:56 pm
Contact:

Re: Smallest possible Kovarex?

Post by coffee-factorio »

It took me some doing to finish that one as a square. In the process I realized that I could manipulate the process so u238 feed could work in stack order. Edit: I thought about what I was doing. The belt inputs items first in, first out, but the input areas closest to the belt entrance deadlock the ones closest to the belt entrance. It looks like only the last item on the belt leaves, which was fooling me.

So that ends up making 2 versions of, lets call it Early Variant. One is, the all beacon design.
variant-square
And the speed on that is 7. But for 12 beacons... you know, it is a study in early game diminishing returns: that's so much steel and advanced circuits that it hurts the nuclear reactor budget. We drop this down to 2 beacons on 3 units and for the amount of objects we throw at the station, the speed 4 we get makes sense.

This is why splitter mechanism was moved out, we can do this many to one and take that to its logical conclusion. It also drops the number of splitters down to 2 by handling that in a workable space. There are probably a couple variants of this that involve good loop backs. Edit: And this one is hard, because it's extenable, even though it doesn't fit into a beacon box it ends up using less parts because the belt is used in a one-many instead of one-one relationship.
variant-diamond
This would work till recyclers, so there's a variant of that for another day. Late game logic will tend towards what royce3 is working on I think.

Last edited by coffee-factorio on Thu Mar 19, 2026 10:52 am, edited 1 time in total.
royce3
Inserter
Inserter
Posts: 22
Joined: Mon May 08, 2017 8:31 pm
Contact:

Re: Smallest possible Kovarex?

Post by royce3 »

quyxkh wrote: Wed Mar 18, 2026 8:36 pm Stop feeding in external '235 once you've got started, it's why your output's clogged. The centrifuge will eventually buffer several full loads ahead of the one in the oven.

And, now you're not going for smallest, you're going for fastest. Here:
pic
drop 40 '235 in manually it reaches 100% uptime soon enough. Sounds gratifyingly like a steam locomotive getting up to speed.

edit: turns out you can still squish this some, you only need 5 inserters. Replaced the image.
yes I didn't state my title very accurately. I think YT clickbait has warped my brain, but once I start improving on something i don't like to stop until it's as perfect as possible. Making it small strips out anything unnecessary, then making it fast and look good are just satisfying enhancements. Collaborating with you on this has been a lot of fun so far.
royce3
Inserter
Inserter
Posts: 22
Joined: Mon May 08, 2017 8:31 pm
Contact:

Re: Smallest possible Kovarex?

Post by royce3 »

quyxkh wrote: Wed Mar 18, 2026 8:36 pm Stop feeding in external '235 once you've got started, it's why your output's clogged. The centrifuge will eventually buffer several full loads ahead of the one in the oven.

And, now you're not going for smallest, you're going for fastest. Here:
pic
drop 40 '235 in manually it reaches 100% uptime soon enough. Sounds gratifyingly like a steam locomotive getting up to speed.

edit: turns out you can still squish this some, you only need 5 inserters. Replaced the image.
not feeding in any more 235 would still clog up because new 235 is created by the machine. In order to trigger the splitter overflow, then priority output has to be blocked, and I think the only way to trigger that is for the belt to be full, or I guess if the priority is slower than the input, but that second scenario will spill excess when you don't want it. Maybe if the slow belt is far enough down the line that it can buffer 40 U235 it might work. I never imagined spending this much time thinking about Kovarex again, hahaha
royce3
Inserter
Inserter
Posts: 22
Joined: Mon May 08, 2017 8:31 pm
Contact:

Re: Smallest possible Kovarex?

Post by royce3 »

quyxkh wrote: Wed Mar 18, 2026 8:36 pm Stop feeding in external '235 once you've got started, it's why your output's clogged. The centrifuge will eventually buffer several full loads ahead of the one in the oven.

And, now you're not going for smallest, you're going for fastest. Here:
pic
drop 40 '235 in manually it reaches 100% uptime soon enough. Sounds gratifyingly like a steam locomotive getting up to speed.

edit: turns out you can still squish this some, you only need 5 inserters. Replaced the image.
I tried your new design and shutoff the 235 feed and it still jammed. I put one slower belt far enough behind the priority splitter so that at least 40 U235 are still on the belt and now it's working perfectly. and the machine is staying busy more often but it's still not non-stop. but I am exceedingly happy with this design, especially since the output feed is opposite the input feed which makes chaining them together look nice. Actually chaining them like that doesn't make sense because the input type doesn't match the output type.

FYI one little circuit trick right after the priority output disabling that belt if there's more than 40 U235 on that section of belt significantly sped up the process and keeps the machine busy all the time, but that's not allowed here, haha.
03-19-2026, 11-08-00.png
03-19-2026, 11-08-00.png (640.49 KiB) Viewed 662 times
And here's 2 of them chained together. I think it would be better to rotate them since the input is on the left and the output is on the right anyway.
03-19-2026, 11-22-15.png
03-19-2026, 11-22-15.png (1.21 MiB) Viewed 662 times
Last edited by royce3 on Thu Mar 19, 2026 3:23 am, edited 1 time in total.
quyxkh
Smart Inserter
Smart Inserter
Posts: 1064
Joined: Sun May 08, 2016 9:01 am
Contact:

Re: Smallest possible Kovarex?

Post by quyxkh »

Maybe try the actual designs people post to help you rather than telling the people giving you gifts how you imagine their gifts are dumb and broken and then building something different?
hunh. I'd swear there used to be a description available on profiles. "quyxkh" is a made-up word for fun, pronounced a lot like "cook".
royce3
Inserter
Inserter
Posts: 22
Joined: Mon May 08, 2017 8:31 pm
Contact:

Re: Smallest possible Kovarex?

Post by royce3 »

I apologize if I offended you, I certainly wasn't trying to do that. I'm really enjoying collaborating with you.
zig1000
Inserter
Inserter
Posts: 42
Joined: Fri Oct 25, 2024 9:57 pm
Contact:

Re: Smallest possible Kovarex?

Post by zig1000 »

Image

My take (note the yellow belt between splitters). Forget if the extra passthrough splitters were necessary or if I just wanted it to be able to buffer if the output backs up.
Tertius
Smart Inserter
Smart Inserter
Posts: 1596
Joined: Fri Mar 19, 2021 5:58 pm
Contact:

Re: Smallest possible Kovarex?

Post by Tertius »

Kovarex alone isn't the whole story.
A common issue is belt clogging because of the U-235/U-238 mix. Often happens in splitters that sort U-235 from U-238. Sometimes U-235 has no demand and stalls, sometimes U-238 has no demand and stalls. While one material stalls, the other might be blocked in a splitter that separates the 2 materials. Or a belt stalls and a kovarex centrifuge is unable to output the 2 U-238 items. Clogging usually happens 8 hours after your checked everything is fine and your Nauvis base suddenly has a blackout.

So I'd like to show my complete never clogging scalable uranium factory from ore to fuel cells/nuclear fuel.
Ignore/remove the circuit wires. They are just for stopping production on surplus. Replace them with reserved/blocked slots in the output chests, if you don't want them.
03-19-2026, 09-33-00.png
03-19-2026, 09-33-00.png (3.73 MiB) Viewed 596 times
Notable parts:
  • every single beacon is optional and the tier of the speed module inside isn't important. Use them to adjust processing speed of the various machines.
  • the tier of the productivity modules is also not important. Use the best you have available to get highest output for lowest input, especially nuclear fuel.
  • to increase ore processing, add speed beacons on the left part or add more ore processing centrifuges on the left.
  • to increase U-235 output, add speed beacons around the kovarex centrifuges. If you still need more, add 2 more kovarex centrifuges by splitting the factory between the 2 existing kovarex centrifuges and create a 6 tile gap. Then insert 2 new kovarex centrifuges into the gap and extend the belts accordingly. Don't forget to extend the underground belt from the loop. There are exactly 2 tiles available for underground exit+entry.
  • at higher throughput it might be necessary to change belts to red or blue. The factory is able to handle this fine. It's even possible to upgrade the quality of any of the machines, beacons and modules with no other change except perhaps upgrading the belt tiers. The U-235 inserters might need a corresponding quality upgrade as well. The factory just runs faster and is still working fine.
  • the bottom left belt is made so that with highest priority U-238 from reprocessed fuel cells is taken, so the system never overflows with used up fuel cells.
  • next priority is the bottom center requester chest that feeds surplus U-238 back into the loop. Surplus from the initial ore processing to get the first 40 U-235. For me, this stock usually lasts until I "win" the game.
  • last priority is U-238 from ore processing. The default.
  • the belt on top of the kovarex centrifuges is made so the belt lane used for output of U-238 is always free, so these centrifuges never clog and stop creating U-235 because of overflow of U-238 from the input side. U-238 from the kovarex output is fed back to the input and has priority over the U-238 we get as input from the ore processing centrifuges


Example for quality-upgraded factory but still the same
royce3
Inserter
Inserter
Posts: 22
Joined: Mon May 08, 2017 8:31 pm
Contact:

Re: Smallest possible Kovarex?

Post by royce3 »

quyxkh wrote: Thu Mar 19, 2026 3:03 am Maybe try the actual designs people post to help you rather than telling the people giving you gifts how you imagine their gifts are dumb and broken and then building something different?
Actually your designs are pretty awesome. Indeed, rotating them by 90 degrees makes them look better and I was even able to share beacons between two of them:
03-20-2026, 02-45-18.png
03-20-2026, 02-45-18.png (1.79 MiB) Viewed 519 times
quyxkh
Smart Inserter
Smart Inserter
Posts: 1064
Joined: Sun May 08, 2016 9:01 am
Contact:

Re: Smallest possible Kovarex?

Post by quyxkh »

royce3 wrote: Thu Mar 19, 2026 2:33 am yes I didn't state my title very accurately.
That does seem to be a theme here. Here's a blueprint of a reoriented-for-pretties-and-convenience version to make it easier, since I've found my own reasoning about things I can't represent accurately to be unreliable.

screenshot-tick-9708956.jpg
screenshot-tick-9708956.jpg (147.29 KiB) Viewed 487 times
edit: I somehow had unboosted s3's in the bp, fixed.
Last edited by quyxkh on Fri Mar 20, 2026 3:44 am, edited 1 time in total.
hunh. I'd swear there used to be a description available on profiles. "quyxkh" is a made-up word for fun, pronounced a lot like "cook".
coffee-factorio
Filter Inserter
Filter Inserter
Posts: 336
Joined: Thu Oct 17, 2024 10:56 pm
Contact:

Re: Smallest possible Kovarex?

Post by coffee-factorio »

Well, for the mass production one.
It took some tinkering and some figuring and...
Loewchen, you were right about it needing a dedicated u238 feed.
Screenshot 2026-03-19 211731.png
Screenshot 2026-03-19 211731.png (449.6 KiB) Viewed 462 times
42 IPS I/O Stable from 3 teams of 2 splitters.
Stacks output. Tiles Y axis only. Cold starts from 40 U235 tested.



I am a bit sad. As originally imagined, I wanted it to have 2 belt lines and there's no possibility of rotating it. The performance demand was higher than anticipated.
quyxkh
Smart Inserter
Smart Inserter
Posts: 1064
Joined: Sun May 08, 2016 9:01 am
Contact:

Re: Smallest possible Kovarex?

Post by quyxkh »

Coffee, you'll get better effective cs from P3's in the centrifuge, and you can squeeze it down by a 1m×12 row.Try the bp I fixed up to include boosted S3s in the beacons just now.

p.s. loved the through-line squishies so much I had to build it too. I don't see how to make it tilable without a wire, though. I missed that detail at first, got it now tho:
p.p.s. okay, last edit, I *swear*. Figured out how to not need the wire, only way a 238 gets to the exit is a cycle just started with unneeded excess so this cannot jam itself. ss1 on the exit speeds up the start but I don't even think that's strictly necessary.

screenshot-tick-27752296.png
screenshot-tick-27752296.png (2.44 MiB) Viewed 416 times
hunh. I'd swear there used to be a description available on profiles. "quyxkh" is a made-up word for fun, pronounced a lot like "cook".
coffee-factorio
Filter Inserter
Filter Inserter
Posts: 336
Joined: Thu Oct 17, 2024 10:56 pm
Contact:

Re: Smallest possible Kovarex?

Post by coffee-factorio »

No wires, it's some black magic involved with the underground belt. I use u238 entities coming off the Kovarex to jam u238 inflow line. If I'm allowed use wires to set conditions and want to save on belts, I'd just link Heligungir's passback build. Lowers speed, I cannot beacon a passback as much.

Feed mechanism is as in picture 1. I don't consider the turn at the end perfect, but it is rather necessary because a turbo belt should support about 12 [edit: 30x4 = 120, 120/5 = 24 ] builds. I barely noticed I could wrap the 238 through the line. So that can be done as a 3x4 tile with some belts hanging outside it. And it's an ugly scenario because the feed belts reverse. Makes a fine blueprint though when it is finished.
03-20-2026, 04-57-27.png
03-20-2026, 04-57-27.png (904.72 KiB) Viewed 388 times
S3L's in this case are being used to establish it will work & fully load itself under worst case load. P3 :any-quality: is the standard you want.
Screenshot 2026-03-20 045440.png
Screenshot 2026-03-20 045440.png (1.26 MiB) Viewed 388 times
\

I can get it to self load and start but not fully sustain i/o on less than 3 [edit: pairs of ] arms, it will give item ingredient shortage and not overrun itself with u235. Otherwise I'd tweak it 1 tile in. Might make 1 more to drop the station size down, it seems reasonable a person could fully load it rather than cold start off 40 exact.
quyxkh
Smart Inserter
Smart Inserter
Posts: 1064
Joined: Sun May 08, 2016 9:01 am
Contact:

Re: Smallest possible Kovarex?

Post by quyxkh »

One convenience wire, self-start, jam-proof, minimum size, can handle an all but full belt. The wire automates a one-time chore that guarantees external jam recovery, reverse the final-output loader at will to test.
To do without the wire disable or remove the wired '238'235 inserters yourself once their centrifuge starts running.
big pic
hunh. I'd swear there used to be a description available on profiles. "quyxkh" is a made-up word for fun, pronounced a lot like "cook".
quyxkh
Smart Inserter
Smart Inserter
Posts: 1064
Joined: Sun May 08, 2016 9:01 am
Contact:

Re: Smallest possible Kovarex?

Post by quyxkh »

Here it is again with the switchback to the second lane inside the beacon box, '238 recycling so it takes only what it needs from the main supply belt, and supply control done with an actually usable splitter and express belt rather than a not-so-usable-in-a-real-map inserter-to-magic-land. It does mean the top left splitter and the four beltsegs after it are part of the design proper.

Like the earlier one this stabilizes pretty quick at 670-671/m, with eight centrifuges on a 60s recipe that peak effective crafting speed is 83.875, /1.5 for the P3s is 55.91⅔, over the rated crafting speed of 55.875×1.5 so if it's throughput limited at all the loss looks to me like a literal rounding error on 0.

I've been bouncing in and out of game.speed 128 and banging on the loaders, can't make it jam after rearranging the internal splitters and going back to bulk inserters to deliver '235 to the exhaust belt.

hunh. I'd swear there used to be a description available on profiles. "quyxkh" is a made-up word for fun, pronounced a lot like "cook".
mmmPI
Smart Inserter
Smart Inserter
Posts: 4965
Joined: Mon Jun 20, 2016 6:10 pm
Contact:

Re: Smallest possible Kovarex?

Post by mmmPI »

You can use (filtered) car to make it faster for a centrifuge to loop the ouput back into the input compared to belt.
with car.jpg
with car.jpg (146.7 KiB) Viewed 279 times
The bp comes with 3 tiled , and i think similar to quyxkh for the uranium input.



There's room for another car in case for fancier logic
Check out my latest mod ! It's noisy !
quyxkh
Smart Inserter
Smart Inserter
Posts: 1064
Joined: Sun May 08, 2016 9:01 am
Contact:

Re: Smallest possible Kovarex?

Post by quyxkh »

Kinda thinking there may be an intended solution…
Screenshot (4).png
Screenshot (4).png (44.48 KiB) Viewed 244 times
hunh. I'd swear there used to be a description available on profiles. "quyxkh" is a made-up word for fun, pronounced a lot like "cook".
mmmPI
Smart Inserter
Smart Inserter
Posts: 4965
Joined: Mon Jun 20, 2016 6:10 pm
Contact:

Re: Smallest possible Kovarex?

Post by mmmPI »

i'm disappointed, it's the same one but more of them ?
Check out my latest mod ! It's noisy !
coffee-factorio
Filter Inserter
Filter Inserter
Posts: 336
Joined: Thu Oct 17, 2024 10:56 pm
Contact:

Re: Smallest possible Kovarex?

Post by coffee-factorio »

quyxkh wrote: Fri Mar 20, 2026 11:12 pm

I've been bouncing in and out of game.speed 128 and banging on the loaders, can't make it jam after rearranging the internal splitters and going back to bulk inserters to deliver '235 to the exhaust belt.
You know, you're making me work :lol:
So, what I ran into was, I can add a third belt to let this thing work like a demon.
Screenshot 2026-03-20 212918.png
Screenshot 2026-03-20 212918.png (388.01 KiB) Viewed 223 times
But at that point, I'm probably letting the layout drive me and not the other way round.



It can be simplified to 1 belt and we can still see rotation. It's essentially Loewchen's logic just with five arms. If they had put a 1 point gap on the red belt on the right hand side, I would swear they were forward allocating space. And the logic here is "these things really don't spit out enough to justify keeping stacking and splitting logic in them". If that's a loss, well. I'll take the loss. Much respect.


Screenshot 2026-03-20 222520.png
Screenshot 2026-03-20 222520.png (595.17 KiB) Viewed 223 times
quyxkh
Smart Inserter
Smart Inserter
Posts: 1064
Joined: Sun May 08, 2016 9:01 am
Contact:

Re: Smallest possible Kovarex?

Post by quyxkh »

mmmPI wrote: Sat Mar 21, 2026 4:06 am i'm disappointed, it's the same one but more of them ?
Yeah, that's fair, it does at least settle on that nifty rate and consume its input near-as-dammit exactly, I wasn't trying for either of those results and the pair of coincidences still tickles my fancy, that's all there really was to that. I did have to do some minor tweaks to get it all stabilized, it's not an exact twinning.

edit: jesus, coffee, are you seeing what I'm seeing? can you squeeze *more* than just the gap column out of that layout you posted?
Last edited by quyxkh on Sat Mar 21, 2026 5:00 am, edited 1 time in total.
hunh. I'd swear there used to be a description available on profiles. "quyxkh" is a made-up word for fun, pronounced a lot like "cook".
Post Reply

Return to “Mechanical Throughput Magic (circuit-free)”