Are there other functional smart furnace setups?

[MeduSalem]

If you like "simple solutions" just dont make steel in smart furnaces, make steel on-site. its much easier, but i guess its not about easier but just to overengineer something and have fun

Well that's what it boils down to anyways in my map because I'm already stuck with the wrong Inserter Stack Sizes (if I would have known earlier I wouldn't have upgraded that far altogether). I had to build a dedicated setup just for Steel Plates since my Smart Furnace array got obviously stuck with residuals after updating to 0.13 and I couldn't find an easy way to adapt the existing contraption so I had to remove Steel from the cycle.

But I guess you are also somehow right... Smart Furnaces are a little bit over-engineered in the first place.

Though XKnight takes it to an entirely new level. If something seems near to impossible it's sure he'll find a way, even if it's out of scope of practical application. That really deserves respect.

[afk2minute]

If you like "simple solutions" just dont make steel in smart furnaces, make steel on-site. its much easier, but i guess its not about easier but just to overengineer something and have fun

Well that's what it boils down to anyways in my map because I'm already stuck with the wrong Inserter Stack Sizes (if I would have known earlier I wouldn't have upgraded that far altogether). I had to build a dedicated setup just for Steel Plates since my Smart Furnace array got obviously stuck with residuals after updating to 0.13 and I couldn't find an easy way to adapt the existing contraption so I had to remove Steel from the cycle.

But I guess you are also somehow right... Smart Furnaces are a little bit over-engineered in the first place.

Though XKnight takes it to an entirely new level. If something seems near to impossible it's sure he'll find a way, even if it's out of scope of practical application. That really deserves respect.

Well if it were possible to request demand from assemblers and chemical plants (and other buildings) via network (read content: demand/min. if its not prodcing because it cant extract produced things, it stops giving that demand) smart furnaces actually would be a really good solution.
You can place constant combinator near every factory with this demand and let furnaces to calculate correct ratio (and switch off with power switch unneeded ones if there are any), that is a bruteforce, but it looks ugly again.

[Lilly]

The solution I have is to melt stuff in batches of 30 (assuming a stack size of 3) and only start melting when those are available in a chest next to the furnace. Additionally, I let each furnace decide for itself what it does. There is no central control circuit that assigns recipes to furnaces, which makes it extendible by pasting the blueprint a few more times. A downside is that the furnaces have to be rebuild if the stack size changes (or if a brown-out happens). They also turn off for a short time period when choosing a new recipe. If the system runs dry, then small amounts of ore/stone/iron will be stuck in the system's chests, but none of the furnaces will contain residual ore/stone/plates.

smart furnace in action

Smart furnaces in action

furnace-alt.png (829.32 KiB) Viewed 5411 times

Note that the fifth furnace was melting iron into steel, but did not receive enough iron to keep going. So now it is accepting all 4 types of ingredients. Nothing is being fed into the furnace however, that only happens when it collected 30 or more of one of the 4 ingredients in its chest. The other 5 furnaces all have chosen a recipe and are currently only accepting more input for their recipe.

how it works

Why 30? because it is the lowest common denominator of 1 (copper & iron), 2 (stone bricks), 3 (stack size) and 5 (steel). This ensures that there are no left-overs:

• the inserter (with stack size 3) moves exactly 30 ore/plates in 10 steps;
• If stone is baked in to stone bricks, then 30 stone will yield 15 stone bricks;
• If iron is melted into steel, then 30 iron will yield 6 steel.

Smart furnaces with labels and without 'alt'.

frunace-labeled.png (223.15 KiB) Viewed 5411 times

• The red wire at A contains the number of Available ingredients. It is what's in the chest minus the amount that is reserved for melting.
• The green wire at B contains the negative number of ingredients reserved for melting. If it is -6 iron, that means that the inserter at A needs to insert 6 more iron. It loops back into itself, so it is updated using pulses.
• The green wire at C is a Clock signal that sends -10k for every ingredient during 2 of 3 clock ticks.
• The combinator block at D Determines whether there are 30 or more of any ingredient in A. In which case it sends a pulse of -30 of that ingredient to B. The clock signal ensures that this is propagated to A before a new check is made.
• The combinator at E checks whether a filter has been set for a specific recipe. If not, it will pass all 4 ingredients as filter to the inserter at F. Those 4 ingredients are set in the constant combinator to the value 1. The combinator is set to 'everything < 2'.
• If a recipe is chosen then F contains the Filter for the ingredient of that recipe. It is computed as B*-1.
• Whenever the inserter moves something from the chest into the furnace, a pulse with the inserter's hand contents is added to B.

The chests should be set to allow only 4 inventory slots to be used.

[XKnight]

Whenever the inserter moves something from the chest into the furnace, a pulse with the inserter's hand contents is added to B.

I've already pointed in two different topics, that inserter's pulse mode is quite dangerous, because combinator might miss this pulse due to low energy. And here the effects will be very painful (furnace will stuck).

Also, I've just measured filter inserter operation time:
top-down straight-belt-to-chest operation from belt's nearest side with stack size=3 requires 33 ticks on express belts, and this is in case of compressed belt filled with single item type.
So, looks like you definitely should use stack filter inserter to have efficient furnaces, otherwise this inserter will be a bottleneck.

[ssilk]

viewtopic.php?f=8&t=31766 One more Smart Furnace®

[Lilly]

I've already pointed in two different topics, that inserter's pulse mode is quite dangerous, because combinator might miss this pulse due to low energy. And here the effects will be very painful (furnace will stuck).

Also, I've just measured filter inserter operation time:
top-down straight-belt-to-chest operation from belt's nearest side with stack size=3 requires 33 ticks on express belts, and this is in case of compressed belt filled with single item type.
So, looks like you definitely should use stack filter inserter to have efficient furnaces, otherwise this inserter will be a bottleneck.

To make sure it doesn't break on low energy, a kill-switch can be added (based on the charge level of an accumulator). Either to shut-down the entire furnace block, or to only clear the inserter filters. During low-energy conditions, there are more important problems than not having the furnaces running.

As for using stack inserters, once those at fully upgraded, they have a stack size of 12. So in that case, use a batch size of 60 (5 insertions). To get around the problem that a stack inserter allows only 1 filter, use the contents of the belt instead of the constant combinator.

[XKnight]

To make sure it doesn't break on low energy, a kill-switch can be added (based on the charge level of an accumulator). Either to shut-down the entire furnace block, or to only clear the inserter filters. During low-energy conditions, there are more important problems than not having the furnaces running.

Yeah, I also thought in the same way, but there is one problem: accumulators may be almost full and the entire system may not have enough energy. How? Your outpost is under attack and your laser turrets are firing... because lasers have higher priority than combinators they will consume all energy. At that moment inserter may pulse a signal (because it has energy buffer) and our combinator is disabled.

As for using stack inserters, once those at fully upgraded, they have a stack size of 12.

Small note: they have stack size 3, 4, 5, 6, 7, 9, 11, 13.

[afk2minute]

Provide a full separate grid for furnaces then?

[XKnight]

Provide a full separate grid for furnaces then?

It is possible, but it is quite expensive.
Also, in this case you should be able to dump extra energy from furnace's grid into the outter world (in case if not all furnaces are running).
But the main disadvantage of this solution is your decision that furnaces are more important than main base defense, and this is bad.
Who cares of some iron plates produced 30 seconds earlier when your outpost is destroyed?

[afk2minute]

maybe its possible to move structures around and provide this grid only to inserters/combinators then?
Sadly blueprints dont preserve wire connections, so even if its possible that is going to be long time setting this up... Again bad.

[MeduSalem]

Provide a full separate grid for furnaces then?

It is possible, but it is quite expensive.
Also, in this case you should be able to dump extra energy from furnace's grid into the outter world (in case if not all furnaces are running).
But the main disadvantage of this solution is your decision that furnaces are more important than main base defense, and this is bad.
Who cares of some iron plates produced 30 seconds earlier when your outpost is destroyed?

In my factory I'm doing it this way:

• Power Supply -> Accumulators/Laser Turrets ... basically getting always power.
• Power Supply -> Power Switch -> Smart Furnace

As long as the Accumlators in the defense grid are at 100% you turn the Power Switch on, if the accumulators in the main grid are less than a certain threshold because your Power Supply can't keep up with the demand for some reason you turn the Power Switch off.

The Power Switch can actually be multiple switches independent switches, each individually controlling a certain amount of furnace/Beacon rows or whatever. Or even completely different parts of your factory altogether.

That's basically how I dynamically re-allocate power among my individual factory departments using priority overrides. But the individual factory deparments can also shut down themselves, for example the smart furnace will shut down if the Storage is full.

But if someone is going to do that you really need a safe way for the Combinators to recover from shutting down the power... that or make at least the combinators a part of the "always-on" grid.

[XKnight]

maybe its possible to move structures around and provide this grid only to inserters/combinators then?
Sadly blueprints dont preserve wire connections, so even if its possible that is going to be long time setting this up... Again bad.

That was my second attempt for belt-based solution. Basically, I tried to fill belt with 10 identical items, and then activate predefined inserters using knowledge about belt configuration in time (assuming belt is not jammed) and desired items for each inserters (this is possible because stack inserter is able to get 10 consecutive items from the express belt without missing one)

This solution is quite robust, unless the grid is low powered, that's why I started to think about separate network, and end up with inability to check is furnace ready to accept new items, and with low efficiency of this build.

[afk2minute]

use buffer chest between belt and furnace and checking its content (and letting the furnace catch up if needed)?
Ah sorry didnt see the gif file, power shortage will not allow inserters to grab fast enough from this kind of belt.

[XKnight]

use buffer chest between belt and furnace and checking its content (and letting the furnace catch up if needed)?

Additional chest... Yes, this is the latest stable solution.
But in previous attempt I wanted to use direct loading from belt, so it should have been more compact.

Ah sorry didnt see the gif file, power shortage will not allow inserters to grab fast enough from this kind of belt.

...that's why I started to think about separate network...

[afk2minute]

use buffer chest between belt and furnace and checking its content (and letting the furnace catch up if needed)?

Additional chest... Yes, this is the latest stable solution.
But in previous attempt I wanted to use direct loading from belt, so it should have been more compact.

Ah sorry didnt see the gif file, power shortage will not allow inserters to grab fast enough from this kind of belt.

...that's why I started to think about separate network...

So if im thinking of this design right way it will be only 4 tiles wider, thats not a big problem at all (10 vs 14 tiles). (its even less as substations can be used for 2 rows at the same time).

[XKnight]

So if im thinking of this design right way it will be only 4 tiles wider, thats not a big problem at all (10 vs 14 tiles). (its even less as substations can be used for 2 rows at the same time).

Strange numbers, why 10? and why 14?
Optimal layout is 6, and new will have 8.
Some pics?

[afk2minute]

So if im thinking of this design right way it will be only 4 tiles wider, thats not a big problem at all (10 vs 14 tiles). (its even less as substations can be used for 2 rows at the same time).

Strange numbers, why 10? and why 14?
Optimal layout is 6, and new will have 9.5.
Some pics?

Just thought of the simpliest thing that i wouldve done there (came first to my mind) had no game loaded at that so maybe im mistaken.
Here is screenshot of reproducing what i thought of.



Substations form 1 grid for furnaces and beacons, wooden poles form another for inserters only. (yes the right inserters are powered from substation too, but it doesnt matter as if power drops there the wooden pole power kicks in and fills the gap, atleast it should).

(i dont know how long should this "section" be so i used 3 furnaces just to show).
2 (substation) + 3 (furnace) + 3 beacon + 4 (belt + inserting system) = 14 (thought that beacon will not reach second column before i loaded the game so i said this number)

With two rows thats 19 rows per 2 furnaces wide. 9.5 just like in your design...

[XKnight]

Substations form 1 grid for furnaces and beacons, wooden poles form another for inserters only. (yes the right inserters are powered from substation too, but it doesnt matter as if power drops there the wooden pole power kicks in and fills the gap, atleast it should).

Wow, you misunderstood what I said before...
If input inserters are powered from the separate grid (this is your case) than we can spawn items in groups by 10 consecutive identical items. BUT, this is possible to do only if you are sure that furnace is empty (otherwise inserter might ignore several items in the begining of the group because furnace is not empty). Intermediate output chest might solve this problem, but as I see this is not your case. So, I really don't understand why do you need this separate energy network? you get nothing from it, just waste your space.

Also, you can place substation above underground belt and save 2 tiles (and merge your networks )

Ops, I get it. Yeah, this solution seems to be stable, and I definitely will try it.

[afk2minute]

Substations form 1 grid for furnaces and beacons, wooden poles form another for inserters only. (yes the right inserters are powered from substation too, but it doesnt matter as if power drops there the wooden pole power kicks in and fills the gap, atleast it should).

Wow, you misunderstood what I said before...
If input inserters are powered from the separate grid (this is your case) than we can spawn items in groups by 10 consecutive identical items. BUT, this is possible to do only if you are sure that furnace is empty (otherwise inserter might ignore several items in the begining of the group because furnace is not empty). Intermediate output chest might solve this problem, but as I see this is not your case. So, I really don't understand why do you need this separate enerrgy network? you get nothing from it, just waste your space.

Also, you can place substation above underground belt and save 2 tiles.

I was thinking that we need "input" chest to not care if there is anything inside the furnace. It will chew through it some time later then, need just to check contents of the input chest and see if its filling up or not and let furnaces catch up if that is happening. So the idea is to have separate grid for inserters (not furnaces), inserters consume much less energy than beaconed furnaces and its cheaper to create a power grid for them, not caring about furnace power.
Yep thats complete different thought. Are there any fundamental flaw that will not allow it to work?

And how output chest will guarantee that there is no power shortage on my furnaces and they just melt too slow and unable to catchup with what im sending them? By checking output chest i suppose?

[XKnight]

This discussion is very active, and we both are editting our posts simultaneosly, so it is quite hard to understand what we are talking about.
I will come back to this thread tomorrow with the fresh mind and new ideas (also it is a deep night here).