Why are my belts not staying full?

[darkSol]

I'm really struggling to see why my builds that are fully optimized with the exact correct amount of machines are not maintaining perfectly full belts. I can cut off the output space belt, let everything fill up, reconnect the output and watch over a matter of minutes as gradually some input belts lose compression, while others get stutters due to being momentarily jammed.

I've carefully looked at all possible explanations, there's just no reason this should occur other than the game simply not being capable of keeping perfect timing between the various cutters and stackers.

I'm curious if everyone is running into this issue. Is it simply intended that we have to overproduce every build to maintain the max throughput? It's causing me no end of frustration.

https://steamcommunity.com/sharedfiles/filedetails/?id=3312707275
This is a screenshot of one of my stacker set-ups, you can see all the gaps in the belts that indicate loss. Every spacebelt that leads into this is optimized and should be maintaining max throughput.

[ciel]

hello fellow ultrawide gamer o7

the trick is that even if your ratio of machines is correct, often times minor asymmetries in belt length on production lines including cutters causes one of the cutters' outputs to backfill before the other, blocking the ability for that platform to continue producing cut shapes at 100% efficiency. thus creating gaps in the item flow. thus downstream saturation problems!

Gotta build a platform to balance your belts in between. something that can allow any input lane to flow to any output lane at 100% saturation. this means that after a little bit, the backed up lines will be distributed among all potential belts, solving the problem of backfill blocking the cutters, and eventually achieving 100% throughput.

[Krypter]

It looks like you have 4 stackers per belt. Assuming similar upgrades on all machines/belts, you need 6 stackers to support a full belt of items

[ciel]

Originally posted by Krypter:

It looks like you have 4 stackers per belt. Assuming similar upgrades on all machines/belts, you need 6 stackers to support a full belt of items

bent stackers

[byakurō]

I'm running into the same thing. Individually each of my platforms can run at 100% throughput, but then when I start combining them I get these random stutters for no apparent reason, usually only at one or two belts. Frustrating indeed.

Originally posted by ciel:

Gotta build a platform to balance your belts in between. something that can allow any input lane to flow to any output lane at 100% saturation. this means that after a little bit, the backed up lines will be distributed among all potential belts, solving the problem of backfill blocking the cutters, and eventually achieving 100% throughput.

Gonna try this next.

[Krypter]

Originally posted by ciel:

Originally posted by Krypter:

It looks like you have 4 stackers per belt. Assuming similar upgrades on all machines/belts, you need 6 stackers to support a full belt of items

bent stackers

i did not realize bent stackers had higher throughput until now! lol

[darkSol]

Originally posted by ciel:

hello fellow ultrawide gamer o7

the trick is that even if your ratio of machines is correct, often times minor asymmetries in belt length on production lines including cutters causes one of the cutters' outputs to backfill before the other, blocking the ability for that platform to continue producing cut shapes at 100% efficiency. thus creating gaps in the item flow. thus downstream saturation problems!

Gotta build a platform to balance your belts in between. something that can allow any input lane to flow to any output lane at 100% saturation. this means that after a little bit, the backed up lines will be distributed among all potential belts, solving the problem of backfill blocking the cutters, and eventually achieving 100% throughput.

I am aware of this issue, but it should be irrelevant with the way I've constructed this production line. Every space belt input is capable of maintaining perfectly compressed belts individually (you can check this by sending them to a trash platform and individually monitoring the belts). By cutting off the final output to the void, all of the stacker arrays will get fully loaded which ensures that arrival to the upper and lower stacker is perfectly timed so nothing is arriving ahead of anything else, and this should never change once you reconnect the output. But it does, and there's no explanation that I could find other than the game simply isn't capable of maintaining "perfection" in the timing.
Now I think you're suggesting using load balancers, and I'm already ahead of you on that if that's the case. I added them into my build after every functional array, and while it does reduce the gaps and increases throughput, it still won't maintain perfect flow.
Here's the balancer I'm using, designed by Theorigin:
https://steamcommunity.com/sharedfiles/filedetails/?id=3312949192
Here's my stacker set up with them incorporated:
https://steamcommunity.com/sharedfiles/filedetails/?id=3312949238
I also have a load balancer on the input lines, just further upstream at the point they are being produced.

[Daerious]

Originally posted by byakurō:

I'm running into the same thing. Individually each of my platforms can run at 100% throughput, but then when I start combining them I get these random stutters for no apparent reason, usually only at one or two belts. Frustrating indeed.

It is due to the timing of items when joining onto merged conveyor belts -- the orange barriers prevent them from moving for a fraction of a second which eventually ends up to the staggered distribution.

[TheOrigin]

Originally posted by darkSol:

Here's the balancer I'm using, designed by Theorigin:
https://steamcommunity.com/sharedfiles/filedetails/?id=3312949192
Here's my stacker set up with them incorporated:
https://steamcommunity.com/sharedfiles/filedetails/?id=3312949238
I also have a load balancer on the input lines, just further upstream at the point they are being produced.

👋

[byakurō]

Been fiddling around some more, and I even have issues with basic shape mining now.

https://steamcommunity.com/id/byakuro/screenshot/2403319584251032900/

3 identical shape miners that go straight into the trash, no processing of any kind. Two of them run at 100%, but one (left in the screenshot) somehow doesn't. Feels more like a bug in the game at this point as opposed to some timing issue.

Originally posted by darkSol:

By cutting off the final output to the void, all of the stacker arrays will get fully loaded which ensures that arrival to the upper and lower stacker is perfectly timed so nothing is arriving ahead of anything else, and this should never change once you reconnect the output.

That would be my intuition as well.

[Khalani]

Originally posted by byakurō:

Been fiddling around some more, and I even have issues with basic shape mining now.

https://steamcommunity.com/id/byakuro/screenshot/2403319584251032900/

3 identical shape miners that go straight into the trash, no processing of any kind. Two of them run at 100%, but one (left in the screenshot) somehow doesn't. Feels more like a bug in the game at this point as opposed to some timing issue.

Originally posted by darkSol:

By cutting off the final output to the void, all of the stacker arrays will get fully loaded which ensures that arrival to the upper and lower stacker is perfectly timed so nothing is arriving ahead of anything else, and this should never change once you reconnect the output.

That would be my intuition as well.

I'm not 100% sure, but how "fresh" is this setup?
Maybe there is an initial gap because of the corner and you never got to a point where it stabalized.

Can you try to cut the belt at the end or at least after the corner, wait until it filled up, and place it again?
See if there are still gaps after.

[byakurō]

I just read on the Shapez discord that there is a known bug that causes these kinds of issues.

I was able to fix the throughput issue from my screenshot simply by deleting and recreating the space belt.

[darkSol]

Originally posted by byakurō:

I just read on the Shapez discord that there is a known bug that causes these kinds of issues.

I was able to fix the throughput issue from my screenshot simply by deleting and recreating the space belt.

I saw mention of a bug with the throwers, and I did isolate one of my production lines was having one of these issues, and that went away after I replaced the throwers on the faulty belt line. But that bug isn't the cause of this problem either because the loss isn't occurring until the shapes are getting combined. I'm in the middle of doing some testing atm, where I've reconstructed my stacker array but only using raw shapes that don't have any pre-processing of any kind. I let that run for a really long time, and while I could occasionally see dips down to 99%, it mostly maintained 100% throughput on the one belt I monitored. So my next step is to test my theory that it's a fluids issue, so I'm coloring just one of the raw shape lines (8 belts worth) and I'm going to see if that leads to the drop in throughput.
https://steamcommunity.com/sharedfiles/filedetails/?id=3313140605
One interesting thing I learned from this build is that 5 layers is the max you can have in a shape, any more get lost when stacking.

[Hassel]

it is currently a bug with the shape launchers, sometimes they arent running at 100%, but will be fixed in the next update which is in the beta branch currently

[darkSol]

Originally posted by Hassel:

it is currently a bug with the shape launchers, sometimes they arent running at 100%, but will be fixed in the next update which is in the beta branch currently

Yes, but it's not what's causing the problem I'm having based on what I read, as replacing the launchers fixes that issue. This problem is coming from somewhere else, and I'm leaning towards fluids or the painter machines.

One thing worth mentioning in regards to that is that I'm currently one upgrade stage lower (lvl 2) in my paint tech vs. my belts/cutters/stackers (lvl 3). The build accomodates for that however because I'm using 4 painters per lane which is over the 3.2 that is required. I'm providing enough fluid that every set of 4 painters has a fluid lane to supply it.

If you look closely at this build you can see some small gaps in the belts, which is why I've been leaning towards this being the culprit.
https://steamcommunity.com/sharedfiles/filedetails/?id=3313147130

[TheOrigin]

Originally posted by darkSol:

One interesting thing I learned from this build is that 5 layers is the max you can have in a shape, any more get lost when stacking.

You mean 4 right? The 5th gets scrapped upon stacking..