#Preventing deadlocks with quality in all steps.
1 messages · Page 1 of 1 (latest)
Instead of parallel no quality lower priority chain for backup one can use quality recycling for EACH item.
why not smelt everything normally and abuse PU/LDS productivity to produce legendary everything?
then legendary plastic lds smuggling for copper and steel, and 1/4 pu-circuit-iron for legendary iron
this way you can "not care" for deadlocks as it will only really produce as much legendary anything as is necessary
holmium/tungsten/biter eggs/spoilage for modules will be the biggest problems however, and I don't have an answer for those rn
how do you prevent the assember with quality mods that produces uncommon green circuits getting stuck because rare green ciruits arent being used.
they arent being used because the assembler making rare red circuits doesnt have rare wire, the rare wire that gets made in the uncommon wire assembler is stuck because uncommon wire isnt being used in the uncommon green assembler, and the dedicated rare wire assembler doesnt have plate because rare copper plate is stuck waiting on uncommon plate. if you have a dedicated line, then the problem backs all the way to the miner.
you need to match ratios for quality for each step in a production chain. the wire sent to green circuits needs to be the same ratio as the iron plate.
or you recycle
void looping any excess is possible but idk if I'd like that. buffer chests for more problematic situations like uncommon wire can work, but aren't automatic unless you include circuit-based active provider chest disposal
I think some recycle voiding is inevitable with quality unless you do pure end product recycle looping which has its own costs
would my idea not be able to avoid recycle voiding?
This is a very late game thing. Before getting there there is a time frame when you want your epics, rares, even uncommons, but do not want to pay too much for it.
fair enough
One crazy idea is to use beacons with speed modules and circuits controlling logic to drop the quality instead of building a parallel setup. If you detect that you have excess quality iron ore that blocks production of common iron ore you swap speed modules in!
Suppose you have a set of buffers and define a global policy for ratio between qualities! If the buffered quality ration starts to exceed the limits you swap speed in
This requires global circuit network coverage
do note that this only works between <quality as a whole> vs common
You can require thresholds for every level for even more craziness 🙂
If you have a train base you can swap speed in if you notice that quality provider station has significantly more than one train of stuff that was not picked up, while lower quality has less than one train.
this, tho. you cannot stop X quality from appearing with speed modules. you either stop the entire thing from producing quality, or you don't
For mining - yes. For other levels you have setups for different qualities
So you can prevent assemblers that consume uncommon plates from creating rare chips
in general. the best you can do is tell the uncommon ingredient handler to only produce uncommon, and no chance at producing rare and such. idk if that'd be able to do much of anything
If you have excess of quality component X that is consumed together with component Y and there is not enough quality Y and quality Y is no longer produced because common Y is not consumed...
void looping is the best answer there
You can attack the problem from two sides. You can increase the supply of low quality X. Or you can decrease the supply of low quality Y. But not with a void looping, with quality upgrade looping
Crazy sh..t
increase the supply of low quality X? oh cuz the common Y isn't being used? sadly the supply of low quality X is unlikely to be fixed with speed modules
It can. It reduces in advance the fraction of high quality X that is not being consumed fast enough to prevent deadlocks.
But it requires some very tricky logic
but that doesn't stop the deadlock, only deals with the aftermath in the case where it was on the verge of happening but not yet
It must be at least somewhat proactive/react on a large buffer
I don't think it'd work. maybe some extra complex stuff that needs a lot of time to figure out after it's built, but eh
cannot recommend 
dw I do get the idea
I just don't think it'd work
And the other balancing act is that consumers of quality Y can consume low quality Y via a scrap with quality loop, but scrapper loops have reduced priority and kick in only when actual consumers of low quality Y are oversaturated.
so, void loops for any excess?
Semi-void
Doesn't void fully. Voids 90% and 10% are upgraded to the next tier.
For simplicity the void-upgrade loops try to upgrade to last tier (kinda wasteful)
But yes, it is more of a void loop, than quality loop. Even with high tier modules you void 90%
if you want to be more efficient in resources
make a recycler upgrader, and once that backs up, a proper void loop
Yep
But I the crazy idea is to drop the quality of stuff that is consumed with what you void loop via speed modules.
So you have double action
what
This is even crazier on Fulgora where production chain is reversed
I think with 2.0 improvements to train this can in theory work reasonably well with big buffers
Using train stop priorities to detect excess and shortages.
uh
maybe, that I can't talk about. I haven't touched 2.0 trains yet
Same, I am theoryctafting based on FFF
When I try to implement there will be surprises.
yeah, go for it
some practical experience with that kind of stuff always comes in handy
You can’t downrank quality with speed modules, they only reduce your chance to rank up
Once it’s quality, it’s stuck that way
This effectively downranks quality of products
…how?
consumes the mid-quality ingredients via recycler looping, which in the grand scheme of things, would reduce global quality index?
idk either
That’s just voiding 

I have miners that produce 8% quality ore. Low quality ore goes to foundries. Foundries are starving because quality or is not produced fast enough. By swapping in speed modules I reduce the percentage of quality ore added to the system.
foundries starving because quality ore is not consumed* fast enough?
If you want to mega fine tune yeah you can fool with micromanaging the quality chance of given steps/ores
Like my coal and stone have lower quality modules just cause I don’t need those materials for mall applications much (coal got boosted back cause modules go brr though)
But there’s still never a way to literally down rank an item
Yes
The idea is to do it with circuits by inserting speed modules into beacons
@cerulean hollow nice discord name by the way
Speed modules in beacons is an interesting idea once you have bmds going
I feel I’d just go for the on/off route instead of anything scaling though. Once I have enough quality plate, throw enough speeds in to completely negate the quality output
I liked using trains to balance stuff in SE, but with quality it is even crazier
bmds?
But at the same time I’d rather just recycle loop to rank things up
I just got done making a quality extractor recycle loop 
Big drills
oh. right
Yep, my idea was that recycle loop and quality downgrade of companion ingredient can go together. And that you can use circuits and beacons
you cannot downgrade quality
Effectively downgrade the average quality of item pool in the system 😛
That's why @unkempt yarrow said that the low quality parallel chain must start from the very bottom, from the mine.
very different from dealing with quality-related deadlocks is the issue
quality related deadlocks come from something stopping because its own quality output cannot be consumed because another ingredient's quality output is itself blocked. Reducing overall quality is not an answer, dealing with excess is
to which, well
void looping, quality recycling, are your best bets
and yes, parallel chain must start all the way from the mine. you can't leave anything with different qualities together
That wastes resources for maximum things should run on the no quality chain (with prod) whenever you don't need quality items (for example if your only making science and all your quality stuff is full)..
my best idea is to put no qual on miners, qual on foundries and have any no-qual excess be turned to quality
with priority input for what comes directly from foundries to lower deadlock possibility
this way you might lower quality item throughout if your no-quality items like science are being consistently produced, but there is less risk of deadlock
The ratio of uncommon to rare output from an assembler crafting with a normal recipe doesn't change as quality % lowers, until 0% quality. Two assemblers set to craft a normal recipe, one with 1% quality and the other with 24.8% quality, will both have 90% of their non normal output as uncommon
His idea was to use speed modules in beacons as a dynamic way to adjust the base quality chance on ore
So that the ratios of quality inputs can be actively controlled to prevent overflows
ratio doesnt change
If you speed mod your iron but not your copper, you now get a lower amount of quality iron assuming it’s skimmed off of normal stuff sent to science
Oh yeah
That’s the ratio we’re talking about, not the multi rank ratio 
thats more complicated than having non priority unquality moded miners
