You should 100% make it into a cross. The huge travel from layer 4 > 3 and 3 > 2 and 2 > 1 is pretty much eliminated. The deliverysystem will be longer, but it doesn't add to any delay, because it's pre-MAM
I really enjoyed the series and learned quite a bit to put in my MAM. A series with viewer MAMs would be great. I think there are quite a few interesting approaches out there. I am finishing mine at the moment with 5 input shapes, 3 input colors and hard wires -- no global wire transmitters. I like to separate as much logic at possible and want to have the option to use as many MAMs as I like without being limited by the channels of the transmitters. I also implemented that my 4 lane MAM can run either 12 Lanes of 3 or 4 level shapes, 24 Lanes of 2 level shapes and 48 Lanes of single level shapes just for the "fun" of it.
No need for a 9 wagon fluid train. The train is going back and forth so rapidly, that it will easily fill up all unloaders. I'm training in all separate colors (7) and there's 3 unload section (one for the shape coloring and 2 for the crystals) and they are all filled up, because of what I said in the beginning.
Great videos im not sure if you will continue improving it or move on one way or another. There are still improvements to make. But that true for everything other that paper clips
17:00 hmm, maybe getting the 2nd mam and pressing f helps?. to join 12 to 3 then 4 it's going to take some space but between just 2 of them a flip is ez.
I had never thought about building a MAM, as a person who knows a little but not enough to do it my self I followed all the videos and now have the same set up for my game, it was a fun journey and even more fun build, thanks for the videos
Yeah definitely a great upgrade. If you can condense the form so it doesn't create too much extra belt travel distance for the original MAM it should be a great improvement. Although remember your MAM is only as fast as your slowest layer.. so if your layer 2 is being optimised by this thing but your layer 4 is all different shape parts, then while layer 2 is complete and backed up it will still be waiting for layer for to build. Just another consideration
well it is great and honestly used yours with some "updates" but still have not got the full crystal layer parts working LOL ... but love this and have started making mine based on like 4 other MAM's and also from my experience in shapez 1 :P ... I wish you had a discord or something to reach out and pick your brain or at least be part of little group love the way you think and make this :)
I too have been using his MAM as a base and making my own modifications. To get the full crystal layer I added a row of platforms under his shape input matrix (below the Pin platforms) that checked for a full crystal layer by using a series of AND gates. If it is a full crystal layer, I then send the entire belt of Pins up to the Stacker and Crystal Generator platform. This splits the throughput and does require an additional Pin shape Unloader to keep everything compact. I originally had the 'full crystal layer check' logic drive the existing Pin platforms to send the Pins up through the matrix using a couple of gates to accept either a full crystal layer or actual pins as input signals. However, as the pins are painted first, before being quartered and re-formed, they were failing all the shape / color checks that he'd built at the Painting platforms. I'm going to have a another look at this later tonight and see if I can't disable some of the checks. It will mean less flushing of the line, but I don't mind that. I think he's overdone it with the shape and color checks as his MAM is so compact, there's little wastage. The next challenge is trying to work out how to do crystal layers that have two colors. I know he's mentioned that it is out of scope, but I'd like to build in that feature. I've not seen any shapes with more than two colors for a full crystal layer and the ones I have seen are always in halves and not in a one-quarter-three-quarter configuration. That might be beyond my capabilities and will never be as elegant as his. The other thing I'd like to play around with is putting in a bypass of the Crystal Generator platform for shapes that have a full pin layer as the base.
Level 51 of random operator shape #2 finally has a full crystal layer... I have some logic that can handle the shape picking (if the entire layer is crystal, it sends a full pin layer instead) but the 1x1 painting platform will be an issue. I'm almost tempted to check the full crystal layer and then send a space belt up to bypass all of the shape selection, cutting, stacking, etc. If I bring it in at the "next layer up" null check then it will work. There's so many filters throughout the MaM that just bypassing the entire thing sounds easier.
I put a pin where the shape needs a crystal. From what I understood from the ROS (Random Operator Shape) they work like this. ROS1 can have pins, shapes either uncoloured or coloured. ROS2 will only have single colored crystals pr. layer, normal shapes (in colours or not) but no pins. Hence the reason I put a pin where I need a crystal. Makes it very easy to check
Like I wondered in an earlier episode on the main series, have you ever played any Zachtronics games? where you program solutions to puzzles? there are very nifty to optimize especialy Opus Magnum is the "easiest" but their other games like Exapunks and TIS-100 have more Litteral Programing, which might be too much like data manipulation
I think you can optimize a lot in the shape selection process. Ur total time of changing a shape is not from the beginning to the end it is from the first decision to delivering the shape to the hole. So i suggest that you build a new platform for selecting the shape. With 5 shapes you can still have 4 from the bottom and one from the side. The selection happens then in the last platform. You would probably safe another minute or so
I love to watch your MaM Series. But i am wondering everyone is building a MaM and Stack "full" Layers, isn't it possible to stack a Quater and then Combine the four Quaters to the Final Shape?
@@DataEngineerPlays I got it to work, thank you, this series taught me so much, once I finish the mode, I'm going to do everything without internet assistance in challenge mode now that I understand that game more
You're correct it doesn't. I did say that in the crystal episode, that there was 2 assumptions 1. Crystals will be the same color 2. There's no gap in the crystal layer
Hey there, I used your blueprint and it works great! The only thing that it DOESN'T do (unless I'm missing something somehow) is a layer that's a crystal layer of 2 colors. (Example Shape code for reference: cmcrcmcr:cwcwcwcw). Essentially, if layer 1 is a crystal shape in all 4 quadrants it doesn't work though. From what I see on your logic I think you'd have make something to allow pins to be on the bottom layer then crystalize it but right now it doesn't do that. Pretty sure the only way would be to tell Layer 1 all 4 are crystals then use pins to make a crystal layer.. the problem I think mainly comes if you have a crystal layer that's 2 colors because you have to do a funky thing with swappers to get them to mix lol.
I one of the earlier videos he mentions that the crystalizer he made was made with assumptions and a layer of crystal only being one color is one of those.
Yes I did say the multicolor crystal is out of scope, but one thing my MAM doesn't do is make full crystal layers. There needs to be a logic check at the pins platform such that if its a full crystal layer to push out a full layer of pins
![Cursed Belt Weaving, Curseder Belt Tapestry & Cursedest Legitimately Obtained "INFINITY CHEST"! [4K]](http://i.ytimg.com/vi/2NKK_2v4jiE/mqdefault.jpg)
You should 100% make it into a cross. The huge travel from layer 4 > 3 and 3 > 2 and 2 > 1 is pretty much eliminated. The deliverysystem will be longer, but it doesn't add to any delay, because it's pre-MAM
I was working through the same thing. Using the space between as alleys for data signals and transport.
Yeah definitely
I really enjoyed the series and learned quite a bit to put in my MAM. A series with viewer MAMs would be great. I think there are quite a few interesting approaches out there. I am finishing mine at the moment with 5 input shapes, 3 input colors and hard wires -- no global wire transmitters. I like to separate as much logic at possible and want to have the option to use as many MAMs as I like without being limited by the channels of the transmitters. I also implemented that my 4 lane MAM can run either 12 Lanes of 3 or 4 level shapes, 24 Lanes of 2 level shapes and 48 Lanes of single level shapes just for the "fun" of it.
If you was to flip layer 2 you could get the MAM's closer and eliminate that long belt. Just a thought, enjoyed the videos.
Is there a chance to update the blueprints for MAM you share? I'm using your basic blueprints and I'm delighted :)
Imo compacting is the most interesting problem I’ve had to tackle in this game. Logic is a close second, like how to implement Else.
Bravo! 4 layer test shape?
EDIT
______
You caught it!
No need for a 9 wagon fluid train. The train is going back and forth so rapidly, that it will easily fill up all unloaders. I'm training in all separate colors (7) and there's 3 unload section (one for the shape coloring and 2 for the crystals) and they are all filled up, because of what I said in the beginning.
Great videos im not sure if you will continue improving it or move on one way or another. There are still improvements to make. But that true for everything other that paper clips
@@blockshift758 endless improvements. I might come back to enhance it after the hard factory series to accommodate those shapes
17:00 hmm, maybe getting the 2nd mam and pressing f helps?. to join 12 to 3 then 4 it's going to take some space but between just 2 of them a flip is ez.
Next phase: Nevermind efficiency, make the MAM spell out "MAM" with it's shape and belts. 😂 I've enjoyed the series!
I had never thought about building a MAM, as a person who knows a little but not enough to do it my self I followed all the videos and now have the same set up for my game, it was a fun journey and even more fun build, thanks for the videos
Yeah definitely a great upgrade. If you can condense the form so it doesn't create too much extra belt travel distance for the original MAM it should be a great improvement.
Although remember your MAM is only as fast as your slowest layer.. so if your layer 2 is being optimised by this thing but your layer 4 is all different shape parts, then while layer 2 is complete and backed up it will still be waiting for layer for to build. Just another consideration
well it is great and honestly used yours with some "updates" but still have not got the full crystal layer parts working LOL ... but love this and have started making mine based on like 4 other MAM's and also from my experience in shapez 1 :P ... I wish you had a discord or something to reach out and pick your brain or at least be part of little group love the way you think and make this :)
I too have been using his MAM as a base and making my own modifications. To get the full crystal layer I added a row of platforms under his shape input matrix (below the Pin platforms) that checked for a full crystal layer by using a series of AND gates. If it is a full crystal layer, I then send the entire belt of Pins up to the Stacker and Crystal Generator platform. This splits the throughput and does require an additional Pin shape Unloader to keep everything compact.
I originally had the 'full crystal layer check' logic drive the existing Pin platforms to send the Pins up through the matrix using a couple of gates to accept either a full crystal layer or actual pins as input signals. However, as the pins are painted first, before being quartered and re-formed, they were failing all the shape / color checks that he'd built at the Painting platforms.
I'm going to have a another look at this later tonight and see if I can't disable some of the checks. It will mean less flushing of the line, but I don't mind that. I think he's overdone it with the shape and color checks as his MAM is so compact, there's little wastage.
The next challenge is trying to work out how to do crystal layers that have two colors. I know he's mentioned that it is out of scope, but I'd like to build in that feature.
I've not seen any shapes with more than two colors for a full crystal layer and the ones I have seen are always in halves and not in a one-quarter-three-quarter configuration.
That might be beyond my capabilities and will never be as elegant as his.
The other thing I'd like to play around with is putting in a bypass of the Crystal Generator platform for shapes that have a full pin layer as the base.
Level 51 of random operator shape #2 finally has a full crystal layer... I have some logic that can handle the shape picking (if the entire layer is crystal, it sends a full pin layer instead) but the 1x1 painting platform will be an issue. I'm almost tempted to check the full crystal layer and then send a space belt up to bypass all of the shape selection, cutting, stacking, etc. If I bring it in at the "next layer up" null check then it will work. There's so many filters throughout the MaM that just bypassing the entire thing sounds easier.
I put a pin where the shape needs a crystal. From what I understood from the ROS (Random Operator Shape) they work like this. ROS1 can have pins, shapes either uncoloured or coloured. ROS2 will only have single colored crystals pr. layer, normal shapes (in colours or not) but no pins. Hence the reason I put a pin where I need a crystal. Makes it very easy to check
Like I wondered in an earlier episode on the main series, have you ever played any Zachtronics games? where you program solutions to puzzles? there are very nifty to optimize especialy Opus Magnum is the "easiest" but their other games like Exapunks and TIS-100 have more Litteral Programing, which might be too much like data manipulation
sounds fun!
I think you can optimize a lot in the shape selection process. Ur total time of changing a shape is not from the beginning to the end it is from the first decision to delivering the shape to the hole. So i suggest that you build a new platform for selecting the shape. With 5 shapes you can still have 4 from the bottom and one from the side. The selection happens then in the last platform. You would probably safe another minute or so
yeah definitely a great idea
I love to watch your MaM Series. But i am wondering everyone is building a MaM and Stack "full" Layers, isn't it possible to stack a Quater and then Combine the four Quaters to the Final Shape?
I dont that works with crystals, because they break when dropped or cut.
You should just flip the new mam layer so the crystal systems are facing each other
on the MAM Final v2.0.spz2bp file of blueprints there are some conflicts popping up do you wanna correct it ?
@@darkysparrow I think it happened after one of the updates
So using your blueprints on drive but the final mam blueprint does use pins, where would you put your pin blueprint
@@davidaustin2087 you mean to get a single layer of pins? You can make that anywhere
@@DataEngineerPlays I got it to work, thank you, this series taught me so much, once I finish the mode, I'm going to do everything without internet assistance in challenge mode now that I understand that game more
I've connected to your discord i can't seem to get the MaM to create a shape so there is a flaw in this design
Couldn't you in theory have a crystal layer that is crcgcbcw? Doesn't seem like this MAM is handling that?
no. There will only be one color crystal pr. layer
@@DenHvideHund on normal mode, but not on hard/insane
You're correct it doesn't. I did say that in the crystal episode, that there was 2 assumptions
1. Crystals will be the same color
2. There's no gap in the crystal layer
Hey there, I used your blueprint and it works great! The only thing that it DOESN'T do (unless I'm missing something somehow) is a layer that's a crystal layer of 2 colors. (Example Shape code for reference: cmcrcmcr:cwcwcwcw). Essentially, if layer 1 is a crystal shape in all 4 quadrants it doesn't work though. From what I see on your logic I think you'd have make something to allow pins to be on the bottom layer then crystalize it but right now it doesn't do that. Pretty sure the only way would be to tell Layer 1 all 4 are crystals then use pins to make a crystal layer.. the problem I think mainly comes if you have a crystal layer that's 2 colors because you have to do a funky thing with swappers to get them to mix lol.
I one of the earlier videos he mentions that the crystalizer he made was made with assumptions and a layer of crystal only being one color is one of those.
Yes I did say the multicolor crystal is out of scope, but one thing my MAM doesn't do is make full crystal layers. There needs to be a logic check at the pins platform such that if its a full crystal layer to push out a full layer of pins
@DataEngineerPlays I fully admit I may have missed that portion. Sound like it's working as intended then (with the exception you mentioned)!