Sweepy fun, but useless?

[Zakery]

Hi all

I try to test last update with sweepy - its funny, but for my opinion is useless.

Only specific use that i find for them - store food from farm in case if we use 2 cells hight farms.

In all other cases loader+ sweeper more effective.

Anyone find useful sweepy use scenario? 

[Zarquan]

I think it could be useful in an ethanol plant, where it can sweep up all the polluted dirt without having to allocate space for multiple sweepers and loaders.  This reasoning can apply to almost any building with solid outputs, including the water sieve, the fertilizer synthesizer, polymer press, farms (as you said) etc.

Also, in the mess hall, dupes regularly get interrupted while eating and leave food on the floor.  A sweepy can pick up after the messy dupes and prevent food from rotting.

It can essentially act as a weaker but longer range sweeper arm in builds, which is good.

[Saturnus]

40 minutes ago, Zarquan said:

I think it could be useful in an ethanol plant, where it can sweep up all the polluted dirt without having to allocate space for multiple sweepers and loaders.  This reasoning can apply to almost any building with solid outputs, including the water sieve, the fertilizer synthesizer, polymer press, farms (as you said) etc.

For anything that has a solid input as well it's not useful as you'll already have a sweeper arm to feed the equipment/plant. So water sieves, fertilizer synths, ethanol distillers, and most farm set ups you have no need for a sweepy. At the same time polymer press output way too much mass for the sweepy to keep up with so it's not useful for that either.

40 minutes ago, Zarquan said:

Also, in the mess hall, dupes regularly get interrupted while eating and leave food on the floor.  A sweepy can pick up after the messy dupes and prevent food from rotting.

Every game you start with at least one hatch and that does a much better job at this task than the sweepy and doesn't take up any extra room.

As the OP I struggle to find any useful applications where a sweepy is a better choice than existing methods (other than the one he specifically mentioned).

[Zarquan]

1 hour ago, Saturnus said:

For anything that has a solid input as well it's not useful as you'll already have a sweeper arm to feed the equipment/plant. So water sieves, fertilizer synths, most farm set ups. At the same time polymer press output way too much mass for the sweepy to keep up with so it's not useful for that either.

I haven't had too much first hand experience with sweepies, as my current base has 80 dupes and the lag would not be pleasant to play with on the beta branch, so I haven't experimented very much.  I do know the numbers.  I believe you on the plastics.  I forget the sheer mass they produce. 

My thought process is as follows:  A conveyor loader can only serve to sweep in a fixed area.  Though its throughput will always be greater than a sweepy, its area is not.  Therefore, if there is a fairly sparse amount of mass appearing per unit area, then perhaps a sweepy could be more space efficient. 

My theory is that they could save space by reducing the number of conveyor loaders in an ethanol distiller setup.  The space is because one sweepy might be able to support more distilleries than a single conveyor loader could.  This could allow for more distilleries per unit area.  I know they pick up 10 kg of stuff when it moves over it, so it should be able to handle many more distilleries than 3 conveyor loaders.  (I say 3 because you still need a conveyor loader to move the stuff away from the ethanol plant).

I think they are useful for farms, at least the way I farm.  I usually have dupes manually deliver to farms so my farms can be more compact, as usable and easily reachable space is normally a larger issue in my colonies than dupe labor.  Plus, some plants don't require solid deliveries, including bristle blossom.  Some people also grow food naturally, so they don't need deliveries.  Since all my farms are as compact as I can make them, sweepies can make sense.  They can deliver the food to one place to be picked up by a sweeper arm later, which could save on dupe labor in a reasonable amount of space. 

I don't like hatches in my great hall.  Hatches eat very small amounts of food, so unless you want to have a lot of hatches in your great halls, they won't be able to handle it if you have a large great hall.  Also, I don't want to waste my food on hatches.  I would rather have a sweepy deal with it, and we would only need one sweepy per great hall.  The space taken by the sweepy is payed back in farm and ranch space saved by having to produce less food.

Also, the sweepy could take the food from the great hall and put it in a sterile atmosphere, as they can move through one tile high areas, so the food wouldn't go bad.

Since they are low powered and have a very good reach, so maybe they can be used as a less laggy and more power efficient space solution than sweepers and rails.  Of course, that is assuming that sweepies are less laggy than conveyor rails.  This part is just a theory, but I think it may be worthy of an experiment.

Also, keep in mind that it is a fairly early game building.  A youtuber managed to build it fairly easily before cycle 10.  Much like the oil refinery, it can be useful early and mid game and become obsolete later.

[Saturnus]

15 minutes ago, Zarquan said:

My theory is that they could save space by reducing the number of conveyor loaders in an ethanol distiller setup.  

Ethanol distillers needs to be loaded with wood so you would already have sweeper arms to cover all the ethanol distillers so a sweepy isn't useful in an ethanol distiller set up.

15 minutes ago, Zarquan said:

I think they are useful for farms, at least the way I farm.  I usually have dupes manually deliver to farms so my farms can be more compact, as usable space and easily reachable space is normally a larger issue in my colonies than dupe labor. 

That is the specific use case mentioned by the OP so no need to mention it again.

15 minutes ago, Zarquan said:

I don't like hatches in my great hall.  Hatches eat very small amounts of food...

That's your choice naturally. They give decor and clean up the small left overs the dupes leave behind. And that's just the point, one wild (sage) hatch easily support the leftovers from 8 dupes because dupes literally only leaves grams on the floor. The only real objection toward hatches in the mess/great hall is the egg shells they leave behind but you can't really ever get too many egg shells so I don't see it as a problem either.

[Zarquan]

32 minutes ago, Saturnus said:

Ethanol distillers needs to be loaded with wood so you would already have sweeper arms to cover all the ethanol distillers so a sweepy isn't useful in an ethanol distiller set up.

I am aware that sweeper arms are still required.  I didn't say it saves space on sweeper arms, I said it potentially saves space on conveyor loaders for polluted dirt.  Not much, but it could be something. 

Though now that I think of it, it could potentially save a little bit of space on sweeper arms too, as the sweeper arms would no longer have to cover the polluted dirt output tile of the ethanol distillery.  Therefore, perhaps they could be rearranged to cover more distilleries per sweeper arm.

I would like to point out that the space gains that could be gotten from this are very small.  This is where the power benefits can be taken in too account.  If sweepy can make the build smaller and more power efficient, then I think it is worth while.

32 minutes ago, Saturnus said:

That is the specific use case mentioned by the OP so no need to mention it again.

True, but I was agreeing with the OP and expanding on his idea.  Along with saying this is a good use case if you didn't want sweepers in your farm, I also expanded on the idea by saying that centralizing the food storage within the farm allowed for the use of a single sweeper arm to handle the output of the farm.  I also added natural farms as a related use case, and natural farms are quite popular these days and they have no need for supplying sweeper arms. 

32 minutes ago, Saturnus said:

That's your choice naturally. They give decor and clean up the small left overs the dupes leave behind. And that's just the point, one wild (sage) hatch easily support the leftovers from 8 dupes because dupes literally only leaves grams on the floor. The only real objection toward hatches in the mess/great hall is the egg shells they leave behind but you can't really ever get too many egg shells so I don't see it as a problem either.

I think the number is also a reasonable objection.  I tend to have perfect decor in my great halls, so the decor is a non-factor.  I tend to have 32 dupes per great hall and multiple great halls, so that is more pathfinding for the hatches.  I use unpowered incubators to manage the eggs in my hatch hatchery, so dupes might mess up and take eggs from the great halls if they were there.  Plus, I don't get more hatches for having some in the great hall, as I put all my eggs in a 3 stage population maintaining pen.  Any hatches I would have in the great hall would be there instead.

Anyway, I think sweepies can help handle early game cleaning as well as a few niche scenarios in the late game.

I wonder what the lag testers will discover about sweepy. 

[Saturnus]

56 minutes ago, Zarquan said:

Anyway, I think sweepies can help handle early game cleaning as well as a few niche scenarios in the late game.

I completely and utterly disagree. Sweepy has in my opinion absolutely no early game use at all. None.

And as I said in the first post, the whole point of my post really but let me just reiterate as it didn't seem like it came across:

2 hours ago, Saturnus said:

As the OP I struggle to find any useful applications where a sweepy is a better choice than existing methods (other than the one he specifically mentioned).

But early game then no, it has no uses what-so-ever.

56 minutes ago, Zarquan said:

I would like to point out that the space gains that could be gotten from this are very small.  This is where the power benefits can be taken in too account.  If sweepy can make the build smaller and more power efficient, then I think it is worth while.

What power benefit? Sweepy uses way more power than sweeper arms and loaders because it's a continuous power drain whereas sweepers and loaders only use power when they are active. It's the same reason sweepers and loaders are way more efficient than dupe labour to do the same task because power-wise you'd get much more power from using a hamster wheel instead. 

[SackMaggie]

Maybe in front of exosuit dock in-case someone peed in their suit.

Would be cool if we have a drone version like Fatorio that can help building/deliver. Or do work like a dupe while eating electricity instead of calories.

[Zarquan]

37 minutes ago, Saturnus said:

I completely and utterly disagree. Sweepy has in my opinion absolutely no early game use at all. None.

And as I said in the first post, the whole point of my post really but let me just reiterate as it didn't seem like it came across:

But early game then no, it has no uses what-so-ever.

I am a little confused.  You usually seem reasonable, but what you say here is that since you struggled to find uses for sweepies, that my examples of potential uses are invalid, and that I should accept that.  The flaw here is that your words are not gospel; they are not true simply because they are your opinion.  And you failing to find a use case does not mean that use cases do not exist.  It reminds me of a quote from Zapp Brannigan from Futurama, "If anyone can move it, I can. [attempts to lift the tree but fails] No one can move it."  (In-A-Gadda-Da-Leela)

I just thought of another use case.  Rather than sending dupes in to pip ranches to pick up little dirt packets, just send a sweepy.  It will pick up the tiny packets and consolidate them in one storage container that the dupes can draw from.  And it can do it earlier in the game than sweeper arms, cheaper than sweeper arms, and more space efficient than sweeper arms.  Sweeper arms would spend large amounts of energy to consolidate the dirt packets, but the sweepy can do it all for a fraction of the power cost. 

A similar principle applies to dreckos and hatches. 

[Saturnus]

54 minutes ago, SackMaggie said:

Maybe in front of exosuit dock in-case someone peed in their suit.

Yeah, I thought about but using a hydrosensor and an notifier that just pauses and zooms in on the location the rare times it happens seem like a better option.

38 minutes ago, Zarquan said:

I am a little confused.  You usually seem reasonable, but what you say here is that since you struggled to find uses for sweepies, that my examples of potential uses are invalid, and that I should accept that.  The flaw here is that your words are not gospel; they are not true simply because they are your opinion.  And you failing to find a use case does not mean that use cases do not exist.  It reminds me of a quote from Zapp Brannigan from Futurama, "If anyone can move it, I can. [attempts to lift the tree but fails] No one can move it."  (In-A-Gadda-Da-Leela)

I have no idea how you can come to this conclusion and I frankly do not care to respond to such false allegations.

However, it does seem like you've not played the test branch and don't realize just how much power the sweepy uses. It uses 240W the entire night cycle, so 18000J. Or 30W on average. Generally sweeper/loaders can be/is time gated when you're dealing with very low amounts that needs to be swept so the total power of a sweeper/loader set up that completely replaces a sweepy is maximum 2400J or 4W on average, and that is setting it very high. Typical set ups would be well under 2W on average.

Again I feel the need to reiterate:

3 hours ago, Saturnus said:

As the OP I struggle to find any useful applications where a sweepy is a better choice than existing methods (other than the one he specifically mentioned).

It's not that I can't find uses cases. It's just that the uses cases it can be used in can be solved better/more efficiently by other methods. 

[Zarquan]

1 hour ago, Saturnus said:

I have no idea how you can come to this conclusion and I frankly do not care to respond to such false allegations.

The point is that you are saying that because you didn't find use cases means that they do not exist as a reason in your previous post that I was wrong.  If I am misunderstanding, then I apologize, but I can't see any other meaning.

1 hour ago, Saturnus said:

However, it does seem like you've not played the test branch and don't realize just how much power the sweepy uses. It uses 240W the entire night cycle, so 18000J. Or 30W on average. Generally sweeper/loaders can be/is time gated when you're dealing with very low amounts that needs to be swept so the total power of a sweeper/loader set up that completely replaces a sweepy is maximum 2400J or 4W on average, and that is setting it very high. Typical set ups would be well under 2W on average.

You don't need to run sweepy every day (assuming you have a dirt stockpile).  You can put it on a timer sensor to only run every 10 cycles and it should be able to pick up all the dirt in a pip ranch. 

On the other hand, to get the 2 W you propose from sweeper arms for the pip ranch use case, you would have to wait until the debris was at least 1000 kg per packet.  This is not a realistic scenario, as the dupes will probably run in to grab dirt long before 8 pips manage to create that much dirt.  Additionally, the dirt won't be in multiples of 1000 kg and you can't stop the sweeper from loading the wrong packets.  Therefore, the sweeper arm will use more power per kg dirt swept than you estimate.  But this workload of spread out small packets is perfect for the sweepy, who only needs to run at once every 10 cycles to clean up the entire ranch. 

Additionally, you don't want to use automation on the sweeper arm because that would mean you can't use the sweeper arm to pick up eggs quickly to remove the pips' cramped debuff as fast as possible unless you have extra sweeper arms specifically for the eggs that can't sweep the dirt.  And sweepies aren't good for eggs as they take longer to remove the egg from the room.

[Saturnus]

4 minutes ago, Zarquan said:

The point is that you are saying that because you didn't find use cases means that they do not exist as a reason in your previous post that I was wrong.

I have no idea what you're talking about.

5 minutes ago, Zarquan said:

This is not a realistic scenario, as the dupes will probably run in to grab dirt long before 8 pips manage to create that much dirt. 

And they wouldn't if you had a deactivate sweepy? It's the same thing. Again, I fear you're drawing conclusions without having played the test branch.

Much like i thought counters and timers were a great addition to the game before I saw the actual implementation and immediately notified the devs to make changes before release.

[Zarquan]

31 minutes ago, Saturnus said:

And they wouldn't if you had a deactivate sweepy? It's the same thing.

The dupes are less likely to take small packets if there are closer large packets.  If you arrange your pip ranch so that the sweepy docks are closer to the places that need dirt, they will generally take the larger packets from the sweepy dock.  And if a dupe decides to take a packet from inside the ranch, it won't make the sweepy that much less efficient to sweep out the rest.  However, waiting for pips to accumulate 1000 kg packets would take an extremely long time. 

Suppose you had a 4x24 pip ranch.  In order to accumulate that much dirt, it would take 8 pips a minimum 150 cycles before the sweepers would turn on to deal with it.  This is assuming the dirt is evenly distributed.  That is a significantly larger amount of time for them to be waiting, and they would still pick up many of the small packets after picking up large packets.  If you need the pips for dirt, then 150 cycles is too long to wait for the sake of power efficiency, and the dupes will spend time picking up small packets of dirt from the front of the ranch. 

This is in contrast to the sweepy, which is designed to deal with debris that is spread out.

Also, there is still the issue of the excess eggs which need to be swept away as quickly as possible.  If you want a ranch which produced the right number of eggs, then you can't turn off the sweeper unless you have fewer animals than the ranch can hold.  If the sweepers are always on, then they are consuming significantly more power picking up every tiny piece of dirt.

31 minutes ago, Saturnus said:

Much like i thought counters and timers were a great addition to the game before I saw the actual implementation and immediately notified the devs to make changes before release.

All of this said, I would be happier if the sweepy swept more material per pass, say 100 kg.  That would make it significantly more powerful, as it could then clean up the debris in the early game in a reasonable amount of time, sparing dupe labor.

[Saturnus]

10 minutes ago, Zarquan said:

The dupes are less likely to take small packets if there are closer large packets.  If you arrange your pip ranch so that the sweepy docks are closer to the places that need dirt, they will generally take the larger packets from the sweepy dock.  And if a dupe decides to take a packet from inside the ranch, it won't make the sweepy that much less efficient to sweep out the rest.  However, waiting for pips to accumulate 1000 kg packets would take an extremely long time. 

Suppose you had a 4x24 pip ranch.  In order to accumulate that much dirt, it would take 8 pips a minimum 150 cycles before the sweepers would turn on to deal with it.  This is assuming the dirt is evenly distributed.  That is a significantly larger amount of time for them to be waiting, and they would still pick up many of the small packets after picking up large packets.  If you need the pips for dirt, then 150 cycles is too long to wait for the sake of power efficiency, and the dupes will spend time picking up small packets of dirt from the front of the ranch. 

This is in contrast to the sweepy, which is designed to deal with debris that is spread out.

Also, there is the issue of the excess eggs which need to be swept away as quickly as possible.  If you want a ranch which produced the right number of eggs, then you can't turn off the sweeper unless you have fewer animals than the ranch can hold.  If the sweepers are always on, then they are consuming significantly more power picking up every tiny piece of dirt.

All of this said, I would be happier if the sweepy swept more material per pass, say 100 kg.  That would make it significantly more powerful, as it could then clean up the digs in the early game in a reasonable amount of time.

The solution is time gating the dirt loader. You already know how much dirt your pips will drop per cycle. So you fill the loader, and only open up for a single 20kg packet to leave at a time. That way you don't need to time gate your sweeper because it will only run when there's room in the dirt loader, and the sweeper will pick up from multiple locations in a single run, which leaves the sweeper ready to sweep up eggs to an egg loader.

Remember, sweepy takes up 2x2 floor tiles while shipping equipment can be placed anywhere.

[Zarquan]

12 minutes ago, Saturnus said:

The solution is time gating the dirt loader. You already know how much dirt your pips will drop per cycle. So you fill the loader, and only open up for a single 20kg packet to leave at a time. That way you don't need to time gate your sweeper because it will only run when there's room in the dirt loader, and the sweeper will pick up from multiple locations in a single run, which leaves the sweeper ready to sweep up eggs to an egg loader.

The issue with this is that it completely negates the power efficiency of the sweeper per kg shipped, as the sweeper will only sweep around 20 kg packets of dirt per swing of the arm instead of 1000 kg.  That makes the 2 W average from an optimal sweeper arm irrelevant, as the sweeper arm is not behaving optimally.

Also, remember that without the sweepers, the ranch can be wider while still remaining a ranch, specifically 32x3.  That makes the ranch area taken up by the sweepers translate in to floor tiles in to around 8 new floor tiles that can house something else, like trees or food production.  If two of those are used by a sweepy, that's ok.

[Saturnus]

2 minutes ago, Zarquan said:

That makes the 2 W average from an optimal sweeper arm irrelevant, as the sweeper arm is not behaving optimally.

Not really. It's still massively more power efficient than the sweepy.

Single 20kg packet sent by loader per pip = 120J x 8 = 1600J, sweeper on for 3s+0.2s per pip, let's just say 5s to round it up so that's 600J. Total is 2200J, or 3.666(6)W on average. Still more than 8 times more efficient than sweepy.

[Zarquan]

37 minutes ago, Saturnus said:

Not really. It's still massively more power efficient than the sweepy.

Single 20kg packet sent by loader per pip = 120J x 8 = 1600J, sweeper on for 3s+0.2s per pip, let's just say 5s to round it up so that's 600J. Total is 2200J, or 3.666(6)W on average. Still more than 8 times more efficient than sweepy.

What is this 0.2 seconds per pip you speak of?  As far as I know, the sweeper doesn't deal in increments that small. 

Assuming the pips only eats once per day and produces dirt once per day, there could be up to 8 20 kg packets.  A sweeper with no automation will transfer an average of 60 kg per swing to an active loader, as it takes 3 seconds to load and it unloads at 60 kg/s.  Assuming the pips are orderly and it is all on one tile, then they produced 160 kg of dirt on that one tile.  This means that ideally, the sweeper will be on for 9 seconds, 2 loads of 60 and one of 40 kg.  But this is assuming all the dirt is in one place.  If the dirt is spread out, as pips are wont to do, then you have to add 1 second to swing to the different pile and pick up dirt from it. (as the piles will not perfectly add to multiples of 20 kg). 

Try putting 160 kg of dirt spread out across 9 tiles of debris and ask the autosweeper to sweep it in to an active full conveyor loader.  You will find that it uses a lot more power because the material is spread out.  It has to stop to pick up every piece.  Also, as a side note, you can't have an autosweeper cover more than 9 tiles.  A sweeper only covers the first 9 blocks, so that means that you need multiple sweepers to cover the ranch and I doubt they will cooperate with each other for ideal sweeping. 

The sweeper does not takes less time and power to sweep 1 kg as it does to sweep 1000 kg.  And it may well be sweeping many small chunks.

In the mean time, the sweepy managed to pick up the 20 kg in about 2 seconds at a cost of a bit more than 60 J.  I could be wrong about the timing, but the bot seems to move 1 tile per second and pick up 10 kg per tile.  And from what I can tell, pips tend to relatively evenly distribute the dirt.  Though I should say I haven't really extensively studied the patterns of pip droppings, I just see piles of approximately the same size.  Of course, the autosweeper's best case is the sweepy's worst case, but the probability of the autosweeper's best case is negligible.

[Saturnus]

16 minutes ago, Zarquan said:

It has to stop to pick up every piece. 

No. It does not.

[Gurgel]

Did anybody try using sweepies to clear the surface Regolith yet (after mining it with autominers, of course) ? If that works, it would probably be my main application. From the sheer volume, it should be enough and about 8 sweepies should cover it. 

[pether]

Won't he overheat?

[Saturnus]

12 minutes ago, Gurgel said:

Did anybody try using sweepies to clear the surface Regolith yet (after mining it with autominers, of course) ? If that works, it would probably be my main application. From the sheer volume, it should be enough and about 8 sweepies should cover it. 

It'd require at least 20 sweepies to keep up with the amount of regolith. Again, sweeper arms combined with automatic dispensers are a much better option.

[Gurgel]

6 minutes ago, Saturnus said:

It'd require at least 20 sweepies to keep up with the amount of regolith. Again, sweeper arms combined with automatic dispensers are a much better option.

I tried that. It was a complete fail because too many rails were needed. So, no, they are not "better", they can, at best, be just as bad. 

Also, I found an estimate of about 20t of Regolith per cycle. 8 sweepies cover the map width, and, at 10kg/tick need to be actually sweeping, less than 50% of their time to remove that. No idea where you get that "at least 20" figure.

21 minutes ago, pether said:

Won't he overheat?

No idea. In the dev-video they had a sweepy mopping up lava.

[Saturnus]

6 minutes ago, Gurgel said:

I tried that. It was a complete fail because too many rails were needed.

You didn't read it right then.

I never mentioned rails or loaders.

Automatic dispensers is the key.

8 minutes ago, Gurgel said:

No idea where you get that "at least 20" figure.

Sweepies sweep about 500kg per cycle

[Gurgel]

2 minutes ago, Saturnus said:

You didn't read it right then.

I never mentioned rails or loaders.

Automatic dispensers is the key.

Nooo, I did read it right. My base-design does not support doing this with automated dispensers and autosweepers. Yours may be different, but that does not solve if for my approach.

So, what about your made-up numbers? Or are there some actual experiments behind them?

Just to be clear: I am not asking to copy your approach. I have zero interest in that. I am asking whether anybody tried whether around 8 sweepies are suitable to clear the Regolith. 

[Saturnus]

11 minutes ago, Gurgel said:

So, what about your made-up numbers?

Ok. I think the discussion ends there. I'm doing this from a standpoint of having tested it. While you've clearly not even tried the testing branch yet.