Friday Facts #382 - Logistic groups

[Qon]

Their solution is similar to what I had in mind. They did not talked about conflicts, but I hope we can order the same item type in different groups.

They can be set up completely independently, on several constant combinators. Obviously you can request things and set a constant combinator to have the same signal. And it would be weird if you couldn't put two groups in the same interface if they have the same signals. Constant combinators can have the same signal in multiple slots as well.

I'm assuming groups will be added together, like in a constant combinator.

What I wonder about is how autotrash values will be handled now. The interface pictures showed only a single value for each request, so are autotrash values handled as separate logistic groups now instead? Since constant combinators can only have a output value for each signal.

And what happens if I couple my personal requests to a constant combinator and another player changes the constant combinators with the same group?

[Waity5]

1 landing pad per world seems fine for non-mega bases, and mega bases are the only ones that can afford infinite research. I'd say it makes perfect sense for the limit of landing pads to be increasable via infinite research, as earlier in the game you're forced to use just one, and in the late game you still need to ration them

[TheVeteraNoob]

I'm here to join the anti 1 platform club. Intentionally limiting throughout feels really anti the spirit of factorio. Perhaps we could be allowed to have multiple space platforns and then 1 landing pad per platform. But one per planet feels like we're going to be forced to rely on having bots doing short travels in and out delivering to a bunch of chests.

This gives me an idea though. I've always thought it was kinda strange for utilizing how bots can occupy the same space at the same time allows for stuff like this. It would be kinda interesting if there was a slight loading time delay for chests and stuff. So you can't just unload a whole chest instantly. Might make the bots vs belts challenge easier. Larger buildings could have more loading docks.

[FuryoftheStars]

I'm here to join the anti 1 platform club. Intentionally limiting throughout feels really anti the spirit of factorio. Perhaps we could be allowed to have multiple space platforns and then 1 landing pad per platform.

Just putting this out there as your wording makes me think you thought otherwise: as far as I can recall, the devs did not say there was going to be a limit of 1 platform per planet. Just the landing pads. There was even a couple points where they used platforms, plural, so we'll be unloading multiple platforms to 1 landing pad. The landing pads will also be extendable with additions, meaning they can get bigger (and perhaps different shapes) and this support more inserters coming off from them (assuming they don't just support logi bots pulling directly from them).

[computeraddict]

1 landing pad per world seems fine for non-mega bases, and mega bases are the only ones that can afford infinite research. I'd say it makes perfect sense for the limit of landing pads to be increasable via infinite research, as earlier in the game you're forced to use just one, and in the late game you still need to ration them

+1, really like this. Having it locked behind the victory condition like current space science techs are makes a nice amount of sense. Or even just a tech that removes the limit, depending on how tall the new tech tree is.

[XT-248]

Just putting this out there as your wording makes me think you thought otherwise: as far as I can recall, the devs did not say there was going to be a limit of 1 platform per planet. Just the landing pads. There was even a couple points where they used platforms, plural, so we'll be unloading multiple platforms to 1 landing pad. The landing pads will also be extendable with additions, meaning they can get bigger (and perhaps different shapes) and this support more inserters coming off from them (assuming they don't just support logi bots pulling directly from them).

Wait a second, do landing pads have cargo bay extensions? Aren't those cargo bay extensions limited to Space Platform Hubs?

The inventory of the space platform hub is limited, but so called cargo bays can be constructed next to the hub to increase the capacity.

I don't think those cargo bays are mentioned with a landing pad at any point, or did I miss something?

[Tohim]

I'm here to join the anti 1 platform club. Intentionally limiting throughout feels really anti the spirit of factorio. Perhaps we could be allowed to have multiple space platforns and then 1 landing pad per platform.

Just putting this out there as your wording makes me think you thought otherwise: as far as I can recall, the devs did not say there was going to be a limit of 1 platform per planet. Just the landing pads. There was even a couple points where they used platforms, plural, so we'll be unloading multiple platforms to 1 landing pad. The landing pads will also be extendable with additions, meaning they can get bigger (and perhaps different shapes) and this support more inserters coming off from them (assuming they don't just support logi bots pulling directly from them).

I was just about to suggest making the landing pads extendable to prevent throughput limitation. But how do you know this to already be true?

[Skellitor301]

This is such a great change! Though I am curious about something and trying to figure this out, even though I know it won't matter come 2.0 I'm still curious to know why. I am still figuring out how people do circuits and I have a basic grasp on how they work, but I for the life of me cannot figure out why in Klonans example of a remote base has the passive supply chest have it's contents be multiplied by -2 in that network. Can someone explain what the purpose of doing that is vs just multiplying the signals by -1 and applying the constant combinator to that signal so if anything is below the amount required it'll send that as a negative signal?

[Tricorius]

This is such a great change! Though I am curious about something and trying to figure this out, even though I know it won't matter come 2.0 I'm still curious to know why. I am still figuring out how people do circuits and I have a basic grasp on how they work, but I for the life of me cannot figure out why in Klonans example of a remote base has the passive supply chest have it's contents be multiplied by -2 in that network. Can someone explain what the purpose of doing that is vs just multiplying the signals by -1 and applying the constant combinator to that signal so if anything is below the amount required it'll send that as a negative signal?

I invert things within an isolated resupply network with a -1. So I’m not sure why he would use -2 (my main thought is that he uses the “read train contents” option on his outposts). I’d have to play around with it, but I’m assuming if he does have the read train contents setup for some reason it would double the supplies, which would maybe necessitate the -2. However, he also mentions using the circuit to call the train (enable / disable the station). And I’m always just using any item is positive. And I think the -2 would make that not work.

Edit: my only other theory is that he is somehow adding that requester chest *and* the network contents (doesn’t look like this sample screenshot is hooked up to a roboport either) back in for some reason. Maybe to help offset when logistics bots pick something up the counts get messed up and the train adds more than it needs? But I don’t think that would fix that issue. But that brings me back to a request I’ve often had is that items bots are carrying are still calculated as “within network” until they are placed into the world. Currently, even if bots pick up some items to quickly shuffle them elsewhere in the network (for example, an ammo chest feeding turrets or a supply belt) they are “removed” from the network and the train then immediately replaces them. Not a huge deal for me, but for high-cost things, it could be.

But basically, if you have a remote outpost and setup a roboport network with a roboport set to display network contents (instead of bot stats) it will report everything available in the chests. You can then setup a constant combinator with all the stuff you want to have on hand in that network. You can invert that (with a -1) and send it to the inserter unloading a train. It will then only unload enough to bring those counts back to positive. You then know you have at least your desire restock count on hand.

Edit: sorry the description above was inverted. Hopefully this image explains it better. (More explanation: viewtopic.php?p=196162#p196162)

So I think, you’re doing the same thing I’m doing.

It’s actually a pretty simple circuit:

IMG_0867.png (3.08 MiB) Viewed 2232 times

[computeraddict]

Time to do the math on what 1 landing pad per surface means for megabasing.

Depending on the size of the building, we'll have some number of spots to stick stack inserters around it. With current max research, stack inserters can move 27.69 items per second. With legendary quality, this increases to ~69.23 items per second. Such a blazing fast inserter can only operate chest-to-chest, so we'll just assume that there are snaked chains of steel chests leading away far enough to where mortal inserters and belts can keep up.

Not all packs need to be imported, though. Nauvis can produce red, green, black, and blue science on its own. One of the other planets may be able to produce all four of the homeworld packs in addition to its local product, too. This leaves the landing pad to import space science and one pack type per other planet, for a total of 4-5 packs of imports. This acts as a divisor on how many SPM a single inserter can support, dropping our whopping 4154 item per minute inserter to supporting a "mere" 831-1038 SPM.

Given that the landing pad appears to be about a 6x6 building like the platform core, this puts us at just under 20k SPM that a single landing pad on Nauvis (importing 5 packs) can support. And given that you can have one of these on up to five planets, we're looking at a best case of 120k SPM if the other planets can also produce red/green/black/blue. If they have to import everything other than their local pack it drops their pads from supporting 25k SPM each to "only" supporting 12k SPM each, for a system total of "only" 68k SPM instead of 120k.

In conclusion, I take back my gripes. Legendary stack inserters are more than capable of making sure the landing pad isn't a bottleneck for any realistic level of production.

And I didn't even have to explore the "roaming labs on space platforms" idea, where you just have a bunch of laboratory platforms running about to pick up science from each of the worlds and analyze it in space since export to platforms is unlimited by virtue of platform count being (so far as we know) unlimited.

Edit: and looking again at the art and previews, I may have gotten the size wrong and we might be looking at a 8x8 or 9x9 for landing pad size, which would bump all those numbers up by 30-50%.

[mmmPI]

And I didn't even have to explore the "roaming labs on space platforms" idea, where you just have a bunch of laboratory platforms running about to pick up science from each of the worlds and analyze it in space since export to platforms is unlimited by virtue of platform count being (so far as we know) unlimited.

Maybe the science pack are super heavy but cheap, giving incentive to use high quality science pack, to reduce the cost of rocket per science researched

[Neutronium]

...

Remember that the landing pad acts as a provider chest as well, opening up the possibility for crazy bot throughput. There will probably be crazy train station loading designs where the pad is surrounded by roboports to load nearby requester chests at stations.

[computeraddict]

Remember that the landing pad acts as a provider chest as well, opening up the possibility for crazy bot throughput. There will probably be crazy train station loading designs where the pad is surrounded by roboports to load nearby requester chests at stations.

I was going to write a section on how high you could get with logibot throughput but it turned out to be entirely unneeded. Stack inserters, especially legendary ones, are just wildly OP when swinging chest-to-chest. Depending on the size of the landing pad you could theoretically have it just unloading directly into trains for 5000 peak SPM per train car you can fit next to it (which is at least 2, even if it's an odd x odd size, and even those can be parity-fixed by car shenanigans) lowered only by how fast the trains can clear the station. 10k SPM throughput for a fairly braindead setup is still wild. 6 legendary stack inserters can fill a train car with 8000 science packs in under 20 seconds.

The numbers with roboports, especially legendary roboports, are faster, but the inserter speed was already so insane I just didn't bother even looking at the bot case.

[TheRaph]

I invert things within an isolated resupply network with a -1. So I’m not sure why he would use -2 (my main thought is that he uses the “read train contents” option on his outposts).

(More explanation: viewtopic.php?p=196162#p196162)

This setup has an general problem. As far I've understand it, it has no protection against an empty train.

So imagine you extend your outpost by one artillery. It immediately start to fire the bug-housings in range and eat up all its shells.
The network start to order new shells and the train are already there to deliver it.
While the train gets unloaded the artillery keep on firing.
Eventually the train are empty in case of shells and the network is empty too.
As far as I understand you picture (and linked post) this will lead to stop the train forever, to being unloaded (what doesn't happen because of empty train).

So to prevent this, the train inventory should be taken in account too for "hold the train until it isn't able to fulfill the needs anymore" ...

or not?

[Tricorius]

I invert things within an isolated resupply network with a -1. So I’m not sure why he would use -2 (my main thought is that he uses the “read train contents” option on his outposts).

(More explanation: viewtopic.php?p=196162#p196162)

This setup has an general problem. As far I've understand it, it has no protection against an empty train.

So imagine you extend your outpost by one artillery. It immediately start to fire the bug-housings in range and eat up all its shells.
The network start to order new shells and the train are already there to deliver it.
While the train gets unloaded the artillery keep on firing.
Eventually the train are empty in case of shells and the network is empty too.
As far as I understand you picture (and linked post) this will lead to stop the train forever, to being unloaded (what doesn't happen because of empty train).

So to prevent this, the train inventory should be taken in account too for "hold the train until it isn't able to fulfill the needs anymore" ...

or not?

That makes sense. I always just bypass that with an “or” seconds have passed condition on the resupply station. But reading train contents is probably a more elegant option.

Edit: sorry, I meant “or” inactive for five seconds. This also handles the case where it can’t fulfill a filter inserter’s requests so it skips back to the until full. It can lead to some unnecessary travel if it can’t load the needed items though.

As an aside, I deliver artillery shells with an artillery train since they stack way higher. But I know that was just an example and this situation could happen with basically any item.

[FuryoftheStars]

Wait a second, do landing pads have cargo bay extensions? Aren't those cargo bay extensions limited to Space Platform Hubs?

[...]

I don't think those cargo bays are mentioned with a landing pad at any point, or did I miss something?

I was just about to suggest making the landing pads extendable to prevent throughput limitation. But how do you know this to already be true?

Because it says it in the same paragraph as it mentions the 1 per planet limit:

Overall logistics

The last non-explained piece of the puzzle is the way items are transported from orbit back to the surface.

This is done with the cargo landing pad, which is a special building that can be extended with cargo bays like the space platform hub, and more importantly, you can have only one per planet!

But also, as someone else mentioned and I've had the chance to reread to confirm, they do support logistic bots pulling directly from them, too, so....

The landing pad has logistic requests, which are satisfied by platforms in orbit. Inserters can pull items from it directly, and it also works as a provider chest when in a logistic network on the surface.

[gGeorg]

Can we modify list of items in the Logi group by combinators ?

[Klonan]

why in Klonans example of a remote base has the passive supply chest have it's contents be multiplied by -2 in that network. Can someone explain what the purpose of doing that is vs just multiplying the signals by -1 and applying the constant combinator to that signal so if anything is below the amount required it'll send that as a negative signal?

It gives a bit of a buffer, so the train will only come when the items in the chest are less than half of what is needed, instead of sending the train just to resupply a single wall or turret

[_TheTechGuy_]

I mean, I just can't get over how cool the concept art looks for the platforms and stuff. Really looking forward to this update!

[Kyralessa]

What happens to blueprints and blueprint books? Do I need to request them too, to avoid them getting trashed by trash-unrequested?

Normally, blueprints are supposted to be in the blueprint library. At this moment, blueprints are trashed as well, but we are planning to change it as it is quite annoying anyway, so blueprints-like items would be ignored by the trash-unrequested feature.

This makes sense to me, as blueprints aren't really something that would be useful to have floating around in the logistics network.

But here's another question:

Can a given item (say, yellow belts) be in more than one logistics group?

If so, what happens when two different logistics groups are active and have conflicting requirements for numbers?