ULTIMATE 2-Full-Belt MAM and MASSIVE Delivery System! Shapez.io 1.4.4

This is the MAM to end all MAMs. It is small, efficient, and can produce TWO full belts of product. This is MAM 3.0 for me.

Check out MAM 2.0!

Check out Joseph Agnew's design!

[sonalita_]

For anyone watching in 2022, this is a CRIPPLED machine. It is an elegant design and Joe has done an amazing job - and it does indeed produce 2 belts of product BUT he crippled it at the cutters. It could easily be a 4 belt machine. Why? Joe goes into detail at 6:12 to explain that 1 slicer can fill a belt. He has 4 slicers here so that's 4 full belts - he even says that when he's discussing the cutters at 18:03 "actually I think it's FOUR!", he kinda stops for a brief moment as he realises that but carries on anyway - he then goes and merges the 4 belts into 2, effectively throwing half the product away!! With a bit more space between the slicers and 1st stackers and obviously doubling the number of stackers, this can easily be a 4 belt machine. The whole point pf pre-painting whole shapes is to produce multiple belts of output. I'm building my MAM based on this design but with 4 belts out. I've prototyped the belting from cutters to stackers and it is quite easy to route the 8 lanes of cutter pair outputs into the 4 stackers - using the same double tunnel system on the cutter outputs as Joe does here.. I had to move the trashcan filter to after the belt routing but before it hits the cutter bypass filters so there's an extra few tiles of propagation delay between cutter output and trashcan filter but now it can run at it's full potential of 4 belts out.

[dannybaromen]

This machine is a BEAST! Took me from level 30-odd to over 400... However, I have noticed now that it periodically runs out of stored shapes.... I'm busy re-working it to have two inputs for the shapes to keep it fed...

[Captain_DeSync]

It is truly amazing what is possible in a game like this. They really should add a creative mode so we can test out new things instead of having to go through the whole game again just to test a small thing. Thanks for sharing!

[jaydenblake]

Love everything about this! Your wires are so clean, and I've gotta say that having every colour and full shapes stored at the start of each lane is brilliant. I stupidly delivered belts of already cut quarters, then painted them. And even worse, only had one set of storage boxes for every 4 lanes (ie per layer).


Couple small suggestions to improve efficiency.
Mainly, your machine has a lot of 'left over clutter' after you level up. You can massively reduce your 'flush time'

1) In each of your combiners, there are shapes stuck on one side, waiting for a partner to be stacked with. This means that the first ~40 shapes you produce will be duds. And the problem is multiplied at each step (ie ~40 shapes lost when stacking NE and SE quadrants together, then ~40 shapes lost when combining the NESE with NWSW halves, then ~40 when combining layers 1&2, etc.)

I know you have a filter gate + trash, after the combiners, to get rid of 'misshapes'. But you can also add a filter + trash at the end of each belt, on each side of the stacker array. That would flush out any left over shapes that have not yet gone through your splitters, into the one belt segment between the splitter and the stacker.

You also could add one at the end of your quad cutter array input belts. At 19.00 mins, we watch the level up happen, and it takes over 30 seconds before the blue squares are flushed out, and white squares can get to the cutters. This is because your current setup only deletes unwanted blue quarter squares AFTER they get cut. Why not add a small circuit that takes all the leftover whole blue squares in the belt and deletes them rather than forcing them through the cutter unnecessarily?


2) Similar minor efficiency gains in your delivery system. You noted that the MAM storage boxes will keep delivering until they are completely empty, and only then do the gates switch, and your upgrade shapes resume their delivery.

Why not have a gate+trash at the front of each of the 16 MAM storage boxes that would only allow the Hub's requested shape to pass through. Then, as soon as you level up, any shapes left in the storage boxes would be diverted to a trash can rather than taking up your Hub's input lines. Wouldn't be too hard to do, just put a transistor in between the clock and the existing filter gates. Link the side to your clock, and bottom input linked to the Hub request, then top output to the gates. This way, the filter gates are controlled by the Hub's requested shape signal, rather than directly by a 1 or 0 from your clock. (If the clock is a 1, the Hub shape will be passed to the filter gates. As soon as the level up happens, the signal coming through will be the 'new' shape so your surplus 'old' shapes will be deleted. If the clock is a 0, there is no signal coming to the filter gates, so they stay closed like you have currently)

4) I also had an idea for a clock that doesn't involve a big loop system. My clock was an item 'counter' rather than a timer. You can create a simple binary counter that counts the items coming from your MAM's output.

On the belt leading from your MAM to your MAM storage, place 8 compact splitters side by side. Their top outputs are merged together using compact mergers and sent on to MAM storage.

The first item to come down this belt will be diverted by the first compact splitter, and go to storage. The second item will pass through the first splitter, but be diverted by the second splitter. The third item is diverted by the first compact splitter. This simple system is counting in binary. 8 splitters is the equivalent of 2^8 = 256. Add a sensor after the final splitter and that's your signal to release the flood.

The sensor will only trip when the 256th item has come through the counter, as this is the first item that will be able to pass through the 7 splitters without being diverted up to storage. So you know you have exactly 256 shapes in your storages. That's enough to win a level. When you get to higher levels, you can add an 9th splitter so you store 512 shapes before the trigger is pulled.

(Instead of adding a 9th splitter, you can actually tweak the exact number you want to count to by inserting a splitter that points down, not up, and looping the down output back into the the counter. Eg With 5 splitters, every 32nd (2^5) item gets through the last splitter. If you add a down splitter between the 4th and 5th up splitters, then loop that back and merge it between the 2nd and 3rd splitters, you will count to 56 instead of 32.
(This is because you are making it go through the 3rd and 4th bits again. 3rd bit is every 8th count, 4th bit is every 16. 8+16 = 24. 24+32 = 56).

This saves having to make massive colour clock loops

[SomeOne-vfrs]

I made the mistake of watching this before anything else wrt to a MAM and this is all completely insane to me. Congrats on making something so amazing and refining it to this degree.

[Xudmud]

I think the only way I could see an improvement on efficiency would be integrating mixers into the design. Only take the three primary colors in and use mixers to get the requested color before sending it off to the rest of the machine. It might extend the construction path a bit but you save a lot of space by not having to pre-mix the colors.

[XtianDH]

I found out I only need 12 lanes of Product because the max delivery is 200 per second
I have enjoyed your Designs and have use some of them in my machine
With 3 MAM 12 lines of delivery I’m now going through levels every five minutes getting ready to hit 1000 soon

[JosephAgnew]

Hey super awesome work! Thanks for the credit also! My wires are very similar yes, but not optimized as i was kinda winging it at the time and need to go back and simplify but... other games got in the way lol.

[audreyplaysnice]

Your previous version would be good if we often need massive quantities (which is what I was thinking we needed); however, 16 sets of 7 colors is daunting. if most of the higher levels concentrate on parts per second, this is perfect. Only one set of colors is so much cleaner. Really a game changer. No wonder I saw some vids at level 600. just set it and forget it leaving the game running 24/7.

[killman]

I basically build a carbon copy of this one. The one thing i did add though is a filter and a trashcan at the end of the belt feeding the painters to clear out the old color faster when the machine switches colors.

[AngeloMiguel]

Can you make a Tutorial for that delivering sistem? Other people videos show delivering sistems only for the HUB shape but yours show the shape plus the upgrads, i need that :)

[MaxiveLegend]

You could double the storage of shapes and colors, since you have a 2x2 space available above the storages when you move the supply line up 2 cells. This gives you a 10k buffer instead of 5k, for basically no cost. That, or you could make this mam a lot more compact by optimizing out that extra space.

[jacobjarvis]

Any chance you can give a tutorial on the build? I was able to get some of it but not all. Just trying to get a better understanding of the complete build. Thanks! Love the content.

[quads]

i did this :
step 1 make a single full belt mam
step 2 copy and paste four times
step 3 enjoy your four belt mam

[AngeloMiguel]

You have a tutorial to How build this machine? I understand zero about the wires. Watching this video tour i barely understood.

[CosminVarlan]

Try to remove the last stacker from the last 3 stackers and put a garbage collector there.... they will clean your lines so fast, and even if you loose some shapes, the result is better (in my case).

[Luna]

i built my mam approximately 650 tiles away, which means it takes me around 32 seconds for the belt to get to my automatic thingy that lets it go automatically, and it only makes 3/s so i need to wait literally like 5 minutes for my automatic system to fill and for the level to pass, which is REALLY bad, but am too lazy to rebuild the entire MAM

[PerTrygveMyhrer]

Another idea is devlivery of raw shapes to mam and have random shapes in and cut and put together to get what mam needs

[HR_DL]

You actually got into my recommended gg

[mrxmry]

1:00 you could achieve the same result with half as many storage units, just by using both output lines.

with a small modification you could make that MAM (or FAF or whatever you want to call it) give you not 2 but FOUR full belts of 1- and 2-layer shapes.