[0.16.51] Bob's miners with high productivity lose resources

[DaleStan]

It appears that, at high speeds and/or productivities, mining drills lose resources when unloading onto belts or splitters.

In the attached save, there are two of Bob's Large Area MK5 mining drills, each with 5 god modules. The mining areas do not overlap, so the two should not affect each other. The north mine has 227k expected iron ore, and the south mine has 96k expected iron ore. Both have a +650% productivity bonus.

Turn the switch to the east on. When the chest by the south mine fills up, turn it off again.

When I do this, the two mines now have expected production of 225k and 94k. (Both have gone down by 2k.)
The south mine shows the expected behavior. The chest by the south mine contains 16000 iron ore, which comes to 2,133 real resources and 13,867 bonus resources.
The north mine shows the unexpected behavior. When I got to the switch, the chests contained 645 iron ore each. 2,580 iron ore means the mine produced 344 real resources and 2,236 bonus resources. I think this mine should have lost 344 (0k or 1k) real resources, not 2k. If I turn the switch back on for a few more minutes, until the four north chests have 6k each, the pattern continues. I expect 24k actual output to be 3,200 real iron ore and 20,800 bonus iron ore, but the north mine is down to 207k expected, a consumption of 20k real iron ore.


I'm using Factorio 0.16.51, and the latest (as I post this) version of all of Bob's mods. Also Creative Mode, to make it easier to build and power the reproducer.

EDIT: To add a screenshot of the two miners.

[Rseding91]

Please upload the mods you're using as well.

[DaleStan]

Sorry about that. Here you go. I also included mod-settings.dat.

[Jap2.0]

That's... odd. I'm getting the same amount in both cases (in the upper one in the video, ~131*16 = ~2k), however in the first case, the belt somehow gets overcompressed, so I don't know if the splitter at the end can't keep up, the mining drill is still outputting resources despite the GUI showing otherwise, or what, but in any case I'm not losing resources.

[DaleStan]

Interestinger and interestinger. It seems to start out OK, but gets worse as time goes on. I added a circuit controlled switch, to make this easier to reproduce. Also, I wasn't clear about which behavior of the north mine was unexpected.

In mining test 3, I doubled the output of the south mine, to 32k. Turn on the manual switch and wait until the circuit-controlled switch turns off and the inserters stop moving. This time, the north chests contain 6292. This part is expected. I added some circuitry next to the chests, and it counts 6318 ticks when at least one inserter is holding at least 1 iron ore. The belt moves 1 item per tick, so I'd say that matches up pretty well. But I expect that 6292 output to be 839 real resources and 5,453 bonus resources. That means the north mine should most likely have gone from 227k expected to to 226k expected. It might stay at 227k expected, but it certainly should not drop to 223k expected.

Good catch on the overcompressed belt; I didn't notice that at all. That also looks like a bug, but isn't what I'm reporting here.

[Jap2.0]

Your first post says that you're getting less from the top mine than the bottom, but it shows the same amount of resources depleted; in my tests, I get the same amount from both mines.

[TruePikachu]

By my math, each of these miners (assuming they're set up identically) should be getting 50.325/s ore from the ground, and 327.1125/s from productivity bonuses. These, changed to per-tick counts, are 0.83875/t from the ground and 5.451875/t from bonuses, for a total expected belt sideloading input of 6.290625/t. I'm pretty certain the overcompression is related to the issue.

[DaleStan]

Your first post says that you're getting less from the top mine than the bottom, but it shows the same amount of resources depleted

Yes, that's what I'm seeing. How long are you running your tests? Less than about 2k, I do see same-to-similar output between the two mines; it's only obviously different once the south mine produces 4k or more. If you are seeing similar output at that production level, are there other settings I haven't posted that could be making a difference?

[Zavian]

5.451875/t from bonuses

I remember reading that assembler type entities can only get a max of 1 bonus recipe from productivity per tick. I'm guessing that applies to other entities as well. I haven't done the math, but could that be causing problems here?

Also I wonder what that means for really high mining productivity bonus from the infinite productivity research in a vanilla game?

[TruePikachu]

I remember reading that assembler type entities can only get a max of 1 bonus recipe from productivity per tick. I'm guessing that applies to other entities as well. I haven't done the math, but could that be causing problems here?

Most likely not; if the miner outputting to a chest is giving significantly different resource counts from outputting to a belt, it won't be the source of issues. All that it would change is the expected output being 1.83875/t in total, for each of the two miners.