Biochamber nutrient consumption discrepancy.

[nononame]

100kw.png (2.94 MiB) Viewed 291 times

Biochamber has 1827 recipe finished. With efficiency modules in beacons -80 % nutrient consumption where should be 200kw of nutrients left. 100 of those 200 were spent on almost making nutrients.So somehow 100kw of nutrients dissapeared.I have 2 hypothesis.
1) Increased power draw reduced effiency of beakons.
2) Threre was a floating point error accumulation.
I've checked power stats and the power consumption was below production.
The production scheme worked for more than 20 hours without trouble and there was nothing suspicious before it broke.
So, i suspect some cumultive interaction.
Does anyone has any ideas what happened?

[Hurkyl]

I just did a test. A single biochamber making nutrients without any modules, a fast inserter removing nutrients.

Drop in bioflux. Drop in one nutrients.

This setup is consistently unable to complete four crafting cycles; the machine goes dead when the fourth one is at 99%.

Change the fast inserter to a stack inserter and it has no problem.

So I posit that crafting machines will consume a hair extra power if a craft is blocked due to the output being full.

[its a me]

It also might just be floating point imprecision, like in viewtopic.php?p=671543

[Hurkyl]

I did another experiment to amplify the effect I observed.

Still doing a biochamber with no modules. I let the biochamber consume 60 nutrients. This should allow for 360 crafts. (4 crafts per nutrient + 50% productivity)

With a fast inserter (or two!), I complete 358 crafts and halt at 73% completion. (the last crafting cycle would have given be two crafts)

With a stack inserter, I complete 358 crafts and halt at 80% completion. (the last crafting cycle would have given be two crafts)

With three stack inserters, I complete 360 crafts.

The single stack inserter seemed to follow a pattern of three blocked cycles to one unblocked cycle, and 20 is roughly 75% of 27.

There were no blocked crafting cycles when I used three stack inserters.

So I am fairly convinced blocked crafting cycles can causally produce this result, whatever the underlying mechanism may be, although of course I can't know if these can happen in your setup. 16 frames out of 240 crafting cycles is a weird number to lose, though.

[Rseding91]

So I posit that crafting machines will consume a hair extra power if a craft is blocked due to the output being full.

Correct. The energy potential consumed that would have gone to starting the next craft is instead lost when the craft fails to output.

[NineNine]

So I posit that crafting machines will consume a hair extra power if a craft is blocked due to the output being full.

Correct. The energy potential consumed that would have gone to starting the next craft is instead lost when the craft fails to output.

I'm perpetually impressed by the detail in this game.