Sometimes using lower tier belts can be useful. Everyone knows 1 red belt equals 2 yellow belts and 1 blue belt equals 3 yellow belts, but what does that really mean? The image below shows some of the things you can do if you don't limit yourself to one belt tier.
Screen Shot 2019-08-24 at 1.28.51 AM.png (7.7 MiB) Viewed 14708 times
The 3 blue balancer I've posted before (viewtopic.php?f=202&t=62326) and a 6 read belt balancer (viewtopic.php?f=202&t=63638). But check out the 5 blue belt balancer using an 8 red belt balancer (not mine) with out any loops. Walls added for emphasis.
Edit: Updated pic to use proper 4 belt balancer in 6 belt balancer.
Re: Belt math, using lower tier belts to effect.
Posted: Wed Aug 21, 2019 7:22 am
by ThreePounds
Will these work with less than 100% on the inputs? Will these work if the outputs are backing up?
Re: Belt math, using lower tier belts to effect.
Posted: Thu Aug 22, 2019 3:03 am
by Trebor
ThreePounds wrote: ↑Wed Aug 21, 2019 7:22 am
Will these work with less than 100% on the inputs? Will these work if the outputs are backing up?
I don't know what you mean by work. These are just offered as an alternative way to do things. They may or may not fit with your play style.
That said here is a test of my 3 belt balancer against the traditional 4 lane belt balancer with the extra output looped back to the unused input to form a 3 belt balancer. The top two are input starved (by placing a yellow belt on the input), the middle run at full speed, and the bottom are backing up (again with a yellow belt but on the output).
Screen Shot 2019-08-24 at 1.10.03 AM.png (4.31 MiB) Viewed 14710 times
Edit: Removed pic of old test setup. Updated pic because I had an improper 4 belt balancer.
Re: Belt math, using lower tier belts to effect.
Posted: Fri Aug 23, 2019 1:30 am
by Trebor
Here's the 6 red belt comparison:
Screen Shot 2019-08-24 at 1.19.06 AM.png (4.06 MiB) Viewed 14710 times
The loaders/infinity chests made it much easer to test.
Edit: Updated pic with proper 4 belt balancer.
Re: Belt math, using lower tier belts to effect.
Posted: Fri Aug 23, 2019 1:36 am
by Trebor
3 belt redone using loaders/infinity chests:
Screen Shot 2019-08-22 at 9.35.10 PM.png (4.81 MiB) Viewed 14812 times
Re: Belt math, using lower tier belts to effect.
Posted: Fri Aug 23, 2019 2:23 am
by Trebor
And for completeness here is the 5 belt:
Screen Shot 2019-08-22 at 10.16.35 PM.png (5.19 MiB) Viewed 14810 times
So, as I said earlier, I don't know what you mean by "do they work". It really all depends on what you require and your play style. It's just a different way to build balancers.
Re: Belt math, using lower tier belts to effect.
Posted: Fri Aug 23, 2019 11:31 pm
by Trebor
Just for grins and giggles:
Screen Shot 2019-08-23 at 7.16.21 PM.png (6.36 MiB) Viewed 14752 times
Note by stopping and starting the outputs at different times the pattern can change:
Screen Shot 2019-08-23 at 7.28.51 PM.png (2.02 MiB) Viewed 14752 times
Not a recommendation, just an observation.
Re: Belt math, using lower tier belts to effect.
Posted: Fri Aug 23, 2019 11:54 pm
by Impatient
Trebor, I don't understand what you try to show us. Are you trying to show ratio splitting?
By ratio splitting I mean to put the contents of one belt in a certain ratio on other belts. eg the contents of one belt can be split 50/50 on two belts by simply using a splitter. But what do I have to do if I want to split them 25/75 (or any other ratio and number of belts)?
The following are examples for 50/50 and 25/75 splitting.
factorio.example.ratio.splitting.png (141.06 KiB) Viewed 14749 times
Is ratio splitting what you are trying to achieve here, by using belts of various speeds?
Re: Belt math, using lower tier belts to effect.
Posted: Fri Aug 23, 2019 11:59 pm
by Oktokolo
Impatient wrote: ↑Fri Aug 23, 2019 11:54 pm
Trebor, I am a little bit stupid and I don't understand what you try to show us.
He showed how to build low footprint belt balancers using lower tier belts.
Impatient wrote: ↑Fri Aug 23, 2019 11:54 pm
But what do I have to do if I want to split them 25/75 (or any other ratio and number of belts)?
25/75 is achieved by splitting the belt twice so you have four belts - then take one of it for 25 and merge the other three back together to get your 75.
Re: Belt math, using lower tier belts to effect.
Posted: Sat Aug 24, 2019 12:01 am
by Impatient
Trebor wrote: ↑Wed Aug 21, 2019 5:18 am
... But check out the 5 blue belt balancer using an 8 red belt ... with out any loops.
Ok, I think I got it now. No Loops.
Re: Belt math, using lower tier belts to effect.
Posted: Sat Aug 24, 2019 12:03 am
by Impatient
Oktokolo wrote: ↑Fri Aug 23, 2019 11:59 pm
... 25/75 is achieved by splitting the belt twice so you have four belts - then take one of it for 25 and merge the other three back together to get your 75.
No no no. Look at the example above. No need for 4 belts.
Oktokolo wrote: ↑Fri Aug 23, 2019 11:59 pm
... low footprint belt balancers ...
The following is the standard design for a 5-5 balancer. I don't see the low footprint, when comparing them.
factorio.5-5balancer.png (599 KiB) Viewed 14741 times
Re: Belt math, using lower tier belts to effect.
Posted: Sat Aug 24, 2019 1:17 am
by Oktokolo
Impatient wrote: ↑Sat Aug 24, 2019 12:01 am
Ok, I think I got it now. No Loops.
Impatient wrote: ↑Sat Aug 24, 2019 12:01 am
Ok, I think I got it now. No Loops.
What is the benefit of having no loops?
Whenever there is a belt loop Factorio doesn't know where the start of the loop is, so when it updates the items it starts at an arbitrary location.
Re: Belt math, using lower tier belts to effect.
Posted: Sat Aug 24, 2019 2:22 am
by Trebor
Here's a 25/75 split:
Screen Shot 2019-08-23 at 10.06.46 PM.png (485.56 KiB) Viewed 14735 times
50% of a blue belt is 1.5 yellow belts, which a red belt and splitter can easily handle.
It's basically the same as you used just more compact. But because it uses red belts it's cheeper yet it has the same throughput when there is no back pressure.
The two are different when back pressure is present, yours can break down to a 0/100 or 100/0, but the worst my can do is 66/0 or 0/66. So which you use may depend on cost or how you want to handle back pressure. (For completeness, replacing the red belt breaks down to 66/0 or 0/100, while replacing the red splitter results in 100/0 or 0/66.)
Re: Belt math, using lower tier belts to effect.
Posted: Sat Aug 24, 2019 2:27 am
by Trebor
Note the standard 5 belt balancer above has one loop.
Re: Belt math, using lower tier belts to effect.
Posted: Sat Aug 24, 2019 3:06 am
by Trebor
Impatient wrote: ↑Fri Aug 23, 2019 11:54 pm
Trebor, ... I don't understand what you try to show us. Are you trying to show ratio splitting?
No not ratio splitting. Some balancers are easier to build, especially those with powers of 2 inputs. So if you have x inputs (and x outputs) you can use a power of 2 balancer and loop back the excess outputs back to the unused inputs. But when you create a belt loop an artifact is created where the start and end of the loop meet and you have no control over where that will be.
What I'm suggesting is to break down the inputs into lower tiers until you have enough to cover all balancer inputs, then combine the outputs back to the higher tiers. If all your inputs/outputs become lower tier the main balancer can become lower tier.
So if you needed to balance 3 blue belts you could use a 4 belt blue balancer by turning turning 2 of the inputs into 3 red and routing the 3 red and extra blue to the balancer then taking 3 of the outputs as red belts back to 2 blue with the extra output to be the third output. (Or you could use the 3 blue to 9 yellow to 3 blue from above.)
Screen Shot 2019-08-24 at 1.32.39 AM.png (1.53 MiB) Viewed 14708 times
Edit: Updated to correct 4 belt balancer.
Re: Belt math, using lower tier belts to effect.
Posted: Sat Aug 24, 2019 3:28 am
by Oktokolo
Trebor wrote: ↑Sat Aug 24, 2019 3:06 am
No not ratio splitting. Some balancers are easier to build, especially those with powers of 2 inputs. So if you have x inputs (and x outputs) you can use a power of 2 balancer and loop back the excess outputs back to the unused inputs. But when you create a belt loop an artifact is created where the start and end of the loop meet and you have no control over where that will be.
Couldn't you just leave unneeded outputs open instead of looping them back. An open output should behave like a blocked one - making the balancer balance his inputs to the other outputs.
Re: Belt math, using lower tier belts to effect.
Posted: Sat Aug 24, 2019 4:20 am
by Impatient
I took "pen and paper" (GIMP and a screenshot) and did the math on your last example. If the input belts are full, it does not balance correctly. It backlogs at the splitter marked in red.
balancer.noloop.1.backlogging.jpg (471.84 KiB) Viewed 14719 times
If you were to replace the red belts/splitter on the right side with blue, it still would not balance evenly.
Is just this design flawed or is this a fundamental flaw in the noloop balancing concept?
Re: Belt math, using lower tier belts to effect.
Posted: Sat Aug 24, 2019 4:29 am
by Impatient
Oh yeah, now I see it. The blue 4-4 in the middle is not a proper 4-4. It misses a splitter for the u-belts. That is a proper 4-4:
balancer.4-4.proper.jpg (71.57 KiB) Viewed 14715 times