Assambler output blocking input with clocked inserters

[J o]

Dear factorians,

while preparing for my next project i come across some unexpected assembler behavior.

So i tried to exactly calculate the item flow for a certain spm through every assembler and clock the inserters to those cycle times. Thats fairly easy but then i noticed something. A first test run showed that the produced spm is far off the calculated ones.

A quick investigation showed: The input inserter wont work when the assambling machine's status is "full output". Ok, fair enough.

Lets take numbers to show the point:
Of course the input cycle time is not a multiple of the output cylce time because of prod modules.

Input cycle time of 12 copper plates and 12 gears are 516 ticks
output cycle time for 12 red science packs are 369 ticks
for a production of 234.5 SPM in 2 Assemblers.

Scince the cycle times don't match some input cycles will be skipped and the machine is not producing as calculated because the status of the assambling machine is "full output".

How to be tricky and get around that behavior? And yes, i dont want to not clock the output inserter and the output stack have to be a multiple of 4.
I thought about having 2 output inserter for every machine but that destoyes the optimization thought behind the inserter clocking.

Thanks!!

[DaveMcW]

You could limit the input inserters to a stack size of 8, and you don't even need to clock them. This produces an output of 11.2 science packs in 344 ticks.

[gGeorg]

You can read the current needs then adapt. Make dynamic self controled system.
Make one output inserter free (unclocked), set it on read hand content & pulse. Then make simple combinator latch to count to 12 (auto-reset on 12) then set all other output inserters in the section (blueprint) to activate on value 12.
Some assemblers might get status full sooner, if it the case then set the activation value not to 12 but lower.

For optimization reason, you would use stack filter inserter, becouse with empty filter they fall imediately into deep sleep. So you set filter for one tick to wake up, and start swing arm. Rest of their time, they are deep sleeping.

[J o]

Hey,

Thanks for the responses!

You could limit the input inserters to a stack size of 8, and you don't even need to clock them. This produces an output of 11.2 science packs in 344 ticks.

Yeah, but then the logistics network manager go crazy for all those bots who fly for a not full cargo.

You can read the current needs then adapt. Make dynamic self controled system.
Make one output inserter free (unclocked), set it on read hand content & pulse. Then make simple combinator latch to count to 12 (auto-reset on 12) then set all other output inserters in the section (blueprint) to activate on value 12.

Right, nice idea! But I wanted to keep the amount of combinators as low as possible. And same for the amount of seperated networks.

I might just put an extra inserter plus box between the assambler and the output inserter, maybe thats a somewhat neat idea, with that i could reduce the amount of logistic boxes as well.

And now i know why i didn´t thought about that behavior, the furnances dont have this status and can be filled with ingredients up to the stack size of the output item.

So long, the factory must grow!

[gGeorg]

Right, nice idea! But I wanted to keep the amount of combinators as low as possible. And same for the amount of seperated networks.

I might just put an extra inserter plus box between the assambler and the output inserter, maybe thats a somewhat neat idea, with that i could reduce the amount of logistic boxes as well.

And now i know why i didn´t thought about that behavior, the furnances dont have this status and can be filled with ingredients up to the stack size of the output item.

So long, the factory must grow!

You probably didnt catch the idea. In case of a blueprint of a smelter, 2 rows of 24 asseblers, eg. 48 assemblers. You only need 2 combinators to make it all clocked. On top, they are all filter inserters so max UPS friendly.

Compare to Adding an extra inserter to each assebler means 48 inserter and 48 boxes. and all are regular inserters, It might work but is pretty wastefull, by resource and also UPS.

[J o]

Right, nice idea! But I wanted to keep the amount of combinators as low as possible. And same for the amount of seperated networks.

You probably didnt catch the idea. In case of a blueprint of a smelter, 2 rows of 24 asseblers, eg. 48 assemblers. You only need 2 combinators to make it all clocked. On top, they are all filter inserters so max UPS friendly.

Compare to Adding an extra inserter to each assebler means 48 inserter and 48 boxes. and all are regular inserters, It might work but is pretty wastefull, by resource and also UPS.

No i totally get your point, but lets say i need to do that for only 2 assamblers, then an inserter and a buffer chest is maybe more practical. Sure for 48 (or even more) its fine, my question would be here for how many assamblers is yours more ups friendly then my idea?

[gGeorg]

Right, nice idea! But I wanted to keep the amount of combinators as low as possible. And same for the amount of seperated networks.

You probably didnt catch the idea. In case of a blueprint of a smelter, 2 rows of 24 asseblers, eg. 48 assemblers. You only need 2 combinators to make it all clocked. On top, they are all filter inserters so max UPS friendly.

Compare to Adding an extra inserter to each assebler means 48 inserter and 48 boxes. and all are regular inserters, It might work but is pretty wastefull, by resource and also UPS.

No i totally get your point, but lets say i need to do that for only 2 assamblers, then an inserter and a buffer chest is maybe more practical. Sure for 48 (or even more) its fine, my question would be here for how many assamblers is yours more ups friendly then my idea?

A project who need 2 smelter assemblers ? well you can use whatever you want.

[J o]

A project who need 2 smelter assemblers ? well you can use whatever you want.

Of course not^^. More generally speaking.

[Guenni7]

Is using a clocked stack filter inserter still ups-friendlier than a normal stack inserter? I thought that had changed?