Re: Is it possible to input balance a single belt's lanes?
Posted: Wed Oct 14, 2020 11:19 am
by dabalciunas
This is the only belt design I know which truly guarantees input and output balancing.
"Waste of resources" or not, for balancing 1-1 belts I would just use the 2-Belt I/O Balancer, because it doesn't use unconnected underground belts, which saves you from unwanted underground belt connection mistakes in very tight spaguetti places.
2-Belt I/O Balancer
2-belt balancer R.png (54.26 KiB) Viewed 6969 times
2-belt balancer L.png (45.26 KiB) Viewed 6969 times
4-Belt I/O Balancer
4-belt balancer.png (309.52 KiB) Viewed 6969 times
Re: Is it possible to input balance a single belt's lanes?
To my knowledge, these are the best available however the 5-5 needs a lot of work, it's a giant mess. There's a really nice 6-6 and 8-8 out there but no one really makes the odd sizes...
Re: Is it possible to input balance a single belt's lanes?
To my knowledge, these are the best available however the 5-5 needs a lot of work, it's a giant mess. There's a really nice 6-6 and 8-8 out there but no one really makes the odd sizes...
I didn't understand the "2-2 Unlimited Throughput" though, what does it improves over the 2-2 version? Maybe I am missing something here...
The ones I posted are true I/O balancers, meaning the can take from any of the 2 lanes of any input belt and then output it to any of the 2 lanes of any output belt. This explains why they look "bigger" than the compact ones.
The 3x3 version you have posted uses more elemements than your 4x4 version, so I would just prefer to use the 4x4 instead of the 3x3.
Below is another slightly different 4x4 design, which also serve as AxB where A and B can be any number between 1 and 4:
And by following the same principle of this 4x4, I also made the 8x8 Full Throughput Balancer: any input lane can be redirected to any of its output lane. Due to the size of the build, minor gaps may occur temporarily until the balancer adapts itself to the new input/output.
I didn't understand the "2-2 Unlimited Throughput" though, what does it improves over the 2-2 version? Maybe I am missing something here...
The normal 2-2 version shown there only has 1 path for left lane input to get to left lane output, the unlimited one has 2 paths.
I did not quite understand yet, I think I must be really missing something here... see the attached images here, the left input belt can be diverted to the left output belt in 2 different ways: inverting lanes or not. Is there really a need for the extra elements, considering the output is limited to the same belt speed as the input?
I was so interested on the compact 4x4 desgn, that I decided to fully test it. However unfortunately I have found a throughput issue with this compact 4x4 design: it cannot fullfill the necessary throughput when only left lanes of output belts are consumed, which could lead to starvation of assembly machines. See the image below, it is self-explanatory...
Re: Is it possible to input balance a single belt's lanes?
Posted: Tue May 11, 2021 11:34 am
by Trebor
dabalciunas, I think you are making an invalid assumption about these lane balancers. Although they have a splinter on the input they should only have one belt as input not two.
Take a look at the designs by disentius in this viewtopic.php?f=202&t=74911 for some two belt lane balancers.
Re: Is it possible to input balance a single belt's lanes?
Posted: Tue May 11, 2021 3:18 pm
by disentius
Solution for single lanes.
However, this will break the input balance the moment you have one of the imput lanes starving, not ideal.
2021-05-11 17_17_02-Window.png (909.33 KiB) Viewed 5667 times
2021-05-11 17_19_36-Window.png (1.33 MiB) Viewed 5667 times
Re: Is it possible to input balance a single belt's lanes?
I didn't understand the "2-2 Unlimited Throughput" though, what does it improves over the 2-2 version? Maybe I am missing something here...
The normal 2-2 version shown there only has 1 path for left lane input to get to left lane output, the unlimited one has 2 paths.
I did not quite understand yet, I think I must be really missing something here... see the attached images here, the left input belt can be diverted to the left output belt in 2 different ways: inverting lanes or not. Is there really a need for the extra elements, considering the output is limited to the same belt speed as the input?
That’s for left belt input to left belt output, now do the same for left lane input (from both belts) to left lane output (from both belts), there is only one belt lane of internal capacity for two belt lanes of input.
Coincidentally the “problem” for the compact 4-4 belt balancer you discovered is the same problem the normal 2-2 belt balancer has, and is the exact problem that the unlimited 2-2 belt balancer solves.
Try the same test on both the 2-2 variants as you did for the 4-4. The normal will have the same problem while the unlimited works perfectly.
Basically in this nomenclature, your 4-4 balancer would be "4-4 Unlimited Throughput".
Re: Is it possible to input balance a single belt's lanes?
That’s for left belt input to left belt output, now do the same for left lane input (from both belts) to left lane output (from both belts), there is only one belt lane of internal capacity for two belt lanes of input.
Coincidentally the “problem” for the compact 4-4 belt balancer you discovered is the same problem the normal 2-2 belt balancer has, and is the exact problem that the unlimited 2-2 belt balancer solves.
Try the same test on both the 2-2 variants as you did for the 4-4. The normal will have the same problem while the unlimited works perfectly.
Basically in this nomenclature, your 4-4 balancer would be "4-4 Unlimited Throughput".
Oh... I see it now. Thanks for opening my eyes for that issue. I was actually wondering why the 2-2 was so much smaller than the 4-4. The explanation is, there is no free lunch!
Anyway, now we got a 2-2 Full-Thoughput Balancer and a 4-4 Full-Thoughput Balancer
I was not happy with the fact that both of these balancers do not guarantee the output of items to happen in an alternate way, from every available input lane. So I came up with a new 2-2 Full-Thoughput Alternator design, which I have also posted on that thread referenced by Trebor: viewtopic.php?p=545588#p545588
That’s for left belt input to left belt output, now do the same for left lane input (from both belts) to left lane output (from both belts), there is only one belt lane of internal capacity for two belt lanes of input.
Coincidentally the “problem” for the compact 4-4 belt balancer you discovered is the same problem the normal 2-2 belt balancer has, and is the exact problem that the unlimited 2-2 belt balancer solves.
Try the same test on both the 2-2 variants as you did for the 4-4. The normal will have the same problem while the unlimited works perfectly.
Basically in this nomenclature, your 4-4 balancer would be "4-4 Unlimited Throughput".
Oh... I see it now. Thanks for opening my eyes for that issue. I was actually wondering why the 2-2 was so much smaller than the 4-4. The explanation is, there is no free lunch!
Anyway, now we got a 2-2 Full-Thoughput Balancer and a 4-4 Full-Thoughput Balancer
I was not happy with the fact that both of these balancers do not guarantee the output of items to happen in an alternate way, from every available input lane. So I came up with a new 2-2 Full-Thoughput Alternator design, which I have also posted on that thread referenced by Trebor: viewtopic.php?p=545588#p545588
And a diagram showing 2 independent internal lanes of left lane to left lane capacity. Note that it branches after splitting and reconverging the lanes into belts.
Mixer-ft.jpg (98.62 KiB) Viewed 5591 times
As the other one branches before reconverging the lanes:
2-Belt Full-Throughput Alternator-annotated.jpg (376.36 KiB) Viewed 5590 times
Re: Is it possible to input balance a single belt's lanes?
It's possible to reduce that slightly, as it duplicates much of the work. For example a slightly shortened version:
Perfect, good catch! I further improved the compactness of the design, still maintaining all of its original properties, allowing it now to be crossed by other belts.
