it was mentioned that painters are now working at 100% even when not looking at them but I was still having problems with them only working at 80%. after watching for some time I would see the fluid pipes suddenly empty. I tracked it back to the fluid launcher if you have a chain of fluid launchers with a pipe to a painter (fluid catcher>painter>fluid launcher) the next launcher will consume all of the paint from the paint catcher before the painter can use it whether it has some thing to send the paint to or not. causing the painter to reset.
Now by the numbers this doesn't make sense. the painter is only consuming 1/4 of the fluid catcher and the 2 painters down the line should be getting backlogged. but for some reason the launcher gets the fluid priority and will launch even if the next catcher can't handle the full load.
#[0.0.9-rc7] fluid launchers stealing fluids
twilit lynx
my current work around is one painter per fluid launcher chain but that can't be intentional.
waxen isle
What’s the input of the green paint onto the island that you’re working on? Is it just 2 launchers of input?
While paint fluid pipes can hold infinitely through itself, 1 single launcher can really only support 4 painters, if my understanding of any of this is correct, you have 16 painters being supplied with 2 launchers
Could you try increasing the input to 4 launchers on the island and see if the problem persists?
twilit lynx
to clear up confusion this is a bit more zoomed out
also something to note the glitch happens more constantly if the paint is attached second or the game loads save
this is the painter blueprint
and before you ask pipe throughput is 4 launchers per at max level
I am using half of that
wary hull
Sorry for the late response, are you also able to share a savegame for easier reproduction? 🙂
twilit lynx
Oddly if I don't merge the fluid launcher source pipes the flow evens out better
stable lily
@twilit lynx is your problem that youre not getting 100% efficiency from theoretically efficient designs that use fluid launchers?
twilit lynx
The pipes are not letting through the throughput they claim they do when you have a chain of (catcher>painter>launcher) if the part where it attaches to the space pipe is connected together.
stable lily
would you say the problem is similar to this? each launcher can supply 4 mixers and in this design there is no point at which it bottle necks to 1 launcher feeding more than 4 mixers and yet the output is not 100%
for testing ive got the white feeding into the theoretically correct amount of painters (a module that ive been using for a while with 100% efficiency) and yet there's points every 15-20 secs where the paint supply is insufficient and the belts from the painters drop below 100% cap
twilit lynx
That dose look like the same problem yes.
When I watch mine 1 or 2 pipes empty completely. That seems to be where the loss is happening. While it dose recover immediately the time spent empty seems to cancel production.
obtuse fable
I've noticed a trend in my crystallizer modules as well, it seems the intended flowrate decreases when there are too many fluid lauchers/catchers in a sprawling system, relying on fluids being evenly distributed more than simply carried from point a to b.
This might be due to how unlike pipes, launchers/catchers act as a one-way valve which hampers the ideal spread of fluids.
Not sure how representative the visuals of fluids are for the code that works behind it, I've tried looking at pipes/launchers/catchers but couldn't really notice a clear pattern. It wasn't present before 0.0.9 and I assume it could be due to the recent improvements on performance.
twilit lynx
I had a large mam that was doing this before 0.0.9 I had to delete fluid launchers between platforms because they were all taking paint even if the other side was too full to accept it.
wary hull
thank you! It'll take a good amount of time until we can look into it, but with a savegame it should be easy to reproduce 🙂
runic sorrel
I think im also experiencing this issue as well, it seems when all the output launchers are connected together the fluid backs up, while if they arent the tanks dont back up
amber quest
I believe I am having the same issue, as each of the 4 segments shown here should be supplied with one fluid launcher worth of paint
Also on a side note, pipe outlines are visible when the floor they are on is not
mental plover
Here's a work-around...
The 2 input fluid launchers on the bottom feed 4 fluid consumers (2 painters and 2 fluid launchers). Only one of those launchers supply paint to the painters on the bottom right. It is those painters that are being starved of paint and making gaps.
The painters on the top right use 2 launchers to feed them. They are producing a full belt.
mental plover
Here's BP of the factory.
The factory in the original post has a different configuration. That factory has a daisy chain of launchers, where each step balances between one painter and one launcher. It looks like the launchers are getting preferential service (over the painters), as the launchers down the chain are always full and the ones at the beginning are starved.
mental plover
Here's a simple example...
With the button off, both belts are full. With the button on, the bottom one has gaps.
This is understandable, but probably not expected by most players.
There's no fluid that is being deleted, it is just being balanced / split to two consumers.
The painter on top is consuming from both the fluid receiver on top and the pipe gate in the middle. This consumes some paint from the bottom fluid receiver, which causes the bottom painter to starve.
It's not clear to me if the original post is the same situation or not.
mental plover
I would agree that the fluid launchers are "stealing" fluid. They are hungrier than the painters, and the painters sometimes starve. All the launchers need to be full before the painters can work at full speed, and this never happens in the original posted factory, because of the way the game splits / balances fluid.
amber quest
so would the issue be resolved if i just connect the paint before everything else and wait for everything to fill up, then connect the shapes?
mental plover
The photo I posted shows how I "fixed" it. I think the important part is having one receiver supply only one launcher. It is still very slow to come up to speed. It is like trying to fill a bucket that has holes in it.
mental plover
so would the issue be resolved if i just connect the paint before everything els...
To your question about filling the paint first, I think that would work for a while, but would eventually revert back to the behavior seen in the original post.
obtuse fable
It's not clear to me if the original post is the same situation or not.
It is, you hit the nail on the head man, thank you.
Indeed, if paint is distributed through a one-way valve (say a fluid launcher/catcher or pipe filter) it prioritizes balancing the fluid before and after it, regardless of whether the part before has a higher demand, or the part after is already saturated.
So this 'obstinate' behaviour causes the fluid to spread unevenly even if there's enough to go round, to fix this make sure you avoid one-way valves since just like when merging two belts, they will obstinately do so even if the one being merged into is fully saturated. So in the end, even if there's a free path, it just lowers the throughput of the other.
I wouldn't really call this a bug, since it's consistent, albeit counterintuitive.