Page 1 of 1

Integrated smart inserter production

Posted: Sun Mar 12, 2017 4:45 am
by vanatteveldt
For a 'there is no spoon' attempt I wanted to make a 'low tech' setup for producing smart inserters from plates, consuming a full yellow belt of iron:

Image

This gets very close to the 67/min you would expect from a yellow belt of iron (see helmod below). To get the maximum blue plant output of 90/min, you would need more than a yellow belt of iron. It should go up to 75 with a bit more iron (e.g. second yellow belt or red section for the gears), to get 90 you would have to add more circuit production as well (and possible add more inserters, haven't tested). However, 67/min is enough to saturate 20 (blue) blue science plants, which should probably be enough for a speedrun (and to at least get you up to stack inserters / beaconed setups if you really feel like going gung ho on the blue science).

I'm not a big fan of looping back the circuits like this, but otherwise you have to make a really complicated setup to get enough circuits and iron fed into the plants. I also tried setting something up with direct insertion of circuits (1 plant to yellow, 2 to blue, and 3 to purple), but that is just horribly complicated so in a speedrun it is certainly not worth it. With stack inserters or inserter bonus it should be much easier, but that's not worth researching, certainly not early.

Note that the red inserters grabbing circuits before the loop are needed because the single yellow belt lane doesn't have enough capacity for the circuits. Possibly you can do with less, since total circuit production at 469 is only just above yellow lane capacity of 400, but that would require a fully compressed belt. Of course you can also use the full yellow belt by side inserting some production of splitting/balancing it, but that seemed more complicated than just adding some red inserters.

(Also Note that there's enough room to place chests between the yellow, blue, and smart inserters so any unused capacity is stored there, which is presumably more useful than buffering the smart inserters; but I prefer to have the basic stuff I need in a separate central location so I don't have to run all over the map to collect goodies)
Production stats
Helmod plan
blueprint string

Re: Integrated smart inserter production

Posted: Sun Mar 12, 2017 8:50 pm
by lachanfall
How about simply changing direction of the yellow belt?
would require a relocation of the inserters too though..
Unbenannt.png
Unbenannt.png (2.79 MiB) Viewed 8504 times

Re: Integrated smart inserter production

Posted: Mon Mar 13, 2017 1:22 am
by vanatteveldt
lachanfall wrote:How about simply changing direction of the yellow belt?
would require a relocation of the inserters too though..
Thanks for the suggestion!

The problem is that the fast inserter needs 3 circuits per second, which I think a single red (or blue) inserter can't deliver.

Re: Integrated smart inserter production

Posted: Mon Mar 13, 2017 8:54 am
by torne
A blue inserter with +1 stack bonus can keep up with 3/s easily. A red inserter with +2 stack bonus should do it too but might not quite keep up perfectly.

Re: Integrated smart inserter production

Posted: Mon Mar 13, 2017 12:30 pm
by lachanfall
Ah you're right.
I just checked the wiki on that one: the fast inserters get 2.3 items/sec. :shock:

Re: Integrated smart inserter production

Posted: Mon Mar 13, 2017 1:46 pm
by mrvn
You can also use underground belts on the iron to create gaps. Then loop the electronic circuit belt through those gaps and use fast inserters. You don't even need to interleave red and yellow underground belts.

Re: Integrated smart inserter production

Posted: Mon Mar 13, 2017 2:10 pm
by vanatteveldt
torne wrote:A blue inserter with +1 stack bonus can keep up with 3/s easily. A red inserter with +2 stack bonus should do it too but might not quite keep up perfectly.
Yes that's absolutely right... but for a speedrun I'm not sure you want to go into stack inserter research, which is 600x30sec red and green to get +1 to normal inserters. I won't be using trains or beacons (except around the silo), so I don't think I need more than blue inserters anywhere...?
mrvn wrote:You can also use underground belts on the iron to create gaps. Then loop the electronic circuit belt through those gaps and use fast inserters. You don't even need to interleave red and yellow underground belts.
I actually built a design like that first. The problem is that the fast inserter needs four inserters to get enough material, as a blue inserter cannot keep up with the 3/s demand for both iron and circuits. So, you need to use more than the three access points on the side. An alternative could be to loop the iron down the side and insert a chest between the fast and smart inserter, so you can pick up excess fast inserters if the smart inserters aren't consumed. But I figured that all things considered the loopback solution was easier to build...

(edit3: maybe you could loop the iron back onto the circuits belt going back left, since 3 blues can get more than 6/s, lemme try :))

Re: Integrated smart inserter production

Posted: Mon Mar 13, 2017 4:12 pm
by torne
Oh, sorry, didn't notice it was a spoon run. Yeah, you won't want the stack bonus research then ;)

Re: Integrated smart inserter production

Posted: Fri Mar 17, 2017 1:25 pm
by mrvn
vanatteveldt wrote:
mrvn wrote:You can also use underground belts on the iron to create gaps. Then loop the electronic circuit belt through those gaps and use fast inserters. You don't even need to interleave red and yellow underground belts.
I actually built a design like that first. The problem is that the fast inserter needs four inserters to get enough material, as a blue inserter cannot keep up with the 3/s demand for both iron and circuits. So, you need to use more than the three access points on the side.
I was talking about the red long inserters. Surely replacing them with blue fast inserters will have higher throughput.

Re: Integrated smart inserter production

Posted: Fri Mar 17, 2017 2:16 pm
by vanatteveldt
mrvn wrote: I was talking about the red long inserters. Surely replacing them with blue fast inserters will have higher throughput.
Right, sure. The problem is you need 3/s of both iron and circuits, which requires >1 blue inserter. So unless you share a belt, you need four inserters to satisfy demand, which is obviously more than you can put on one side of an assembler. With a single blue inserter you can produce 69/s (assuming the 2.3/s quoted above is correct), so this would become the bottleneck.

Re: Integrated smart inserter production

Posted: Sat Mar 18, 2017 6:02 pm
by Case
Hey mate,
I played a bit with your idea and came to this ..
Not sure if it fits for a speed run but it has a more constant output ... compared to yours with is more periodic it seams.
Stats
Build
If you play a bit (or do the math) for the inserters (yellow/blue - to balance both belts) you should reach the cap ... and maybe you find a better way to feed the one sided belt.

Case.

Re: Integrated smart inserter production

Posted: Sun Mar 19, 2017 4:49 am
by vanatteveldt
Thanks!

I had a look at your build, and I think it's a good idea to just use a second (half) iron belt to solve the iron shortage.

Problem is though that 6 circuit assemblers require 60*6*2*0.75=503 iron/minute, while a single yellow lane only transports 400, so the last assemblers are iron starved. This problem is aggregated by the inserter assemblers 'stealing' some of the iron. You alleviate the problem with the inserters that insert the iron back on the top belt, but that's pretty ugly :)

So I figured I could keep the second iron belt parallel with the first until the assemblers, which gives 75/min, consuming the full belt of copper:
more iron
This can also be extended to get the full 90/min by adding a bit of copper (I made the first stretch of belt red) and adding 1 more circuit assembler. Since the yellow lane would not be able to carry enough circuits, I place the extra assembler at the end and loop around the filter inserter:
max production
I have to admit I never realized fast inserter production consumed that many circuits, which probably explains why I'm always surprised to be copper/green starved when I ramp up blue science :)

The other take home message is that with a single red belt shared iron/circuits you also get max production:
simple linear design
(edit: never mind the doulbe blue inserters between the plants, but I'm too lazy to make a new screenshot)

While a single shared yellow belt gets you about 60/min, which is enough to feed 16 "blue" blue science assemblers, which is probably plenty in most cases:
yellow belt circuits+iron
I think I'm done with filter inserter production analysis now :-)