FTB Infinity Evolved -- Applied Energistics 2 Autocrafting -- Channels Problem

[Brotuulaan]

I'm new to AE2, and I'm having a hard time understanding how exactly the system connects and uses channels.

I know that the system looks at attached devices and assigns each one a channel, and the smart/dense cables show how many channels are traveling through any given block of that cable.

I also know that AE2 blocks transmit power and channels through each other if they're touching--at least to some extent.

I'm trying to figure out how people make those autocrafting blocks with the molecular assemblers and interfaces, and this is where I'm running into problems.

The MindCrafters made a nice tut series on AE2, and in part 3 (on autocrafting) at about 8 mins, he says that putting a single cable on the 3x3x3 cube "lights up" the whole cube. I wondered if he's talking only about the assemblers and ignoring the interfaces, as the one uses no channels while the other does per device. A 3x3x3 cube as he has it set up has 14 interfaces and would require 14 channels. But this is where I discovered my problem.

Dense cables don't seem to transmit their full 32 channels to blocks, only to other dense cables. When I connect a dense cable to the center of one face of the cube, the connection shrinks to a smart cable size/display scheme and shows only 8/32 channels used. I have to connect extra cables to reach all 14 interfaces. But even a second dense cable placed on a bottom corner only picks up 2 more channels instead of the other 6. If I instead add the second cable above the first (center top of face), then it picks up the rest of the 6 channels for a total of 14.

I also tried stacking a column of 4 interfaces/5 assemblers in 8 layers, making a full 32 interfaces. But right now, I can't seem to get all 32 channels to work no matter what I do. Running dense cables around the column on the ground then pillaring dense cables up all four sides against the interfaces doesn't help because all the channels are concentrating on the closest side, which sort of makes sense that the cables wouldn't use unnecessary paths--except that it's topping out at 18 channels for the whole structure.

So what are the quirks of attaching cables to devices and transmitting channels through devices and b=directly through blocks? None of this seems to be behaving as it should, and it's very frustrating when things don't work.

***EDIT***
Breaking and replacing several of the cables in the active face updated them, and I'm getting all 32 channels now.

But that still leaves my question about why dense cables don't transmit 32 channels to structures and what the quirks are. I just happened to find one for myself.

Also, breaking and replacing cables on my initial 3x3x3 cube has left me with only 13 channels functioning instead of 14, even with the same connection points as before.
***EDIT***

 

[Brotuulaan]

Ugh. After breaking and replacing and snaking and encircling and all sorts of other things, the cube is finally showing 14 channels on its dense cable. And it has the exact same cabling pattern as it did earlier. Is this a common problem with the mod, that it doesn't calculate channels correctly until the cables are replaced multiple times?

 

[Brotuulaan]

Though despite this frustration, I will say that the P2P tunnel is freakin' awesome.

 

[Reddis]

The best advise I can give with AE2 channels is to designate a specific color to a specific task. I usually use dense white for generic cable that I need to run around my base, my home network is you will. Then I use red dense cable for anything that collects to power, green dense cable that connects to auto crafting, and so on. The thing about AE2 networks is if you are over your channel limit, SOMETHING will stop working SOMEWHERE. You have no idea what machine it will be without research. Using the above method of cabling, if my entire network looses power, I know something went wrong in the red cables and I only have to troubleshoot that section.

 

[adamich]

"about why dense cables don't transmit 32 channels to structures"
dense cables transmit 32 channels beetwin dense cables, any over cables or AE blocks (like constructors or interfaces) transmit only 8 channels.

 

[Brotuulaan]

Reddis, that's a fantastic idea for troubleshooting systems! I knew about the colors and separating networks, but it hadn't occurred to me that it would give that benefit as well!

Adamich, that's good to know, that dense cables only give 8 channels per connection to any blocks. Kinda stinks that you have to break them out, but that's at least a minor issue that can be accounted for. Do you know anything about why breaking and replacing cables worked only after several tries and reconfigurations, or why some connection positions didn't give the full channels? Perhaps there's a range limit to block-transmitted channels or something?

 

[adamich]

We need fotos of Your machinery, but I already can say that AE have a problem with distribution channels in Your cable network.

 

[Brotuulaan]

Here are a few pics of my multiblock structure where the 13-14 channels thing happens. Every picture uses ONLY dense cables.

With two connections to the structure (one assembler, one interface), 13 channels:

When I add a dense cable to the interface on the bottom right, it doesn't add the 14th channel:

When I add the extra cable on top instead of the bottom corner, it adds the 14th channel:

But when I pull the cable split back a ways and pull it off the face of the assembler and add the connection on the bottom right again, it adds the full 14 channels:

And just now when I went back to the fourth version from the first, it topped out at 10 channels. I broke nearly every one of the gray cables and replaced it, with the number jumping back up to only 10, but then when I broke the gray cable connected to the white and reset it, it finally jumped all the way to 14.

So all in all, very confusing to me as to how it calculates channels. I wondered originally if there was a range limit on the block transmission, like +/-1 Y from the cable (because of #2 and #3), but #4 clearly eliminates that possibility.

 

[Brotuulaan]

It looks like using the bottom corners tops out at 10 unless you break the connection to the middle and let the corner jump from 2-8, then add the middle connection back in for the last 6. Bottom corner only grabs 2 if 8 are already running on the middle.... Top corners w/o middle also tops out at 13 channels, even after letting them reset when breaking connection. I'm sure there are more quirks besides these.

Do these quirks jive with what you all have experienced? Is there any way to predict these quirks so I can know if a setup is going to be impossible vs just requiring some breaking and resetting until it works right?

 

[adamich]

As i wrote "any over cables or AE blocks (like constructors or interfaces) transmit only 8 channels."
Ur cube of interfaces 3x3x3 is not optimal - game can't connect it correctly.

 

[Brotuulaan]

Are you talking about the connection from the cable into the block itself, or the cable itself as distinguished from the connection? By that, I mean the big block of the cable rather than the small pipeline that extends into the block.

I'm following that the cable connection that extends into the block is only able to transmit 8 channels (I think it's silly, but at least the idea seems consistent), but it wouldn't make sense to me if the connection also stopped the cable from transmitting its regular channel count to other connected cables.

I would think a connection doesn't limit the cable's transmission into connected cables, as the third pic above shows a full 14 are passing that first connection into the multi-block structure and getting to the second and third connections.

Instead of using the large-ish multi-block structure like many people seem to do, do you prefer to connect groups of 8 blocks (2x2x2) in an array, so they're not touching other groups and taking advantage of the block-to-block transmission? I've seen at least one person split his connection into a 3x3 tower into the four corners with plain fluix cables. Perhaps he did that to make it symmetrical despite these glitchy behaviors.

 

[Brotuulaan]

The thing that really gets me is the second picture above. Adding a cable on one side as a Y makes no difference to the system's channels but adding it on a different spot as a line does. That and the system recognizing *almost* every single channel with just two connections, but not quite. Seems strange to me

 

[GamerwithnoGame]

This probably doesn't help, but: when I was setting up my AE molecular assembly chambers and interfaces, I had it set up so that the first time the dense cable connected to the bulk of them, there was no question about which 8 it would be going to, and same for the second 8. I'm not explaining this very well...

If you can picture this, I had a checkerboard of 2 wide, 4 tall and 8 deep. That made 32 interfaces and 32 MACs. I then had the dense cable running along the top from one end to the other. The first interface it ran across it connected to, and obviously the next block was an MAC. I set up a cable anchor there to make sure it didn't connect, and the point after that was another interface, where it connected. I did the same all the way along. What that meant was the first connection point only dealt with the first 2x4x2 set of 8 interfaces and 8 MACs, and the second point had the next 8 and so on; that way there was no confusion by the system - at least that's how it seemed to work?

I honestly don't know what the rules are for connections and stuff, but It might be worth a try. I should also mention I ran the cable across after placing the anchors in the relevant places, so I wonder if it was just doing it in a certain way that got it working? How sold are you on the 3x3s?

 

[Brotuulaan]

Yeah, that makes sense. So each time the cable connected, it would grab the next 8 interfaces from the next 2 rows back. It seems like there is some sort of range that the cables search, and they can get in each other's way somehow. I wonder if they default to the furthest cable taking the closest connections it can, and the closer ones can't grab channels that would pass through blocks being occupied by the further cables. Or maybe they try to distribute the channels available between them and count the average as the per-connection max. That would limit the first couple connections and keep the last cable from grabbing more than the other connections have.

Just experimented some with your 2x4 setup, and it was consistent that when I removed the anchor, the incoming cable dropped 4 channels, as if the connections were blocking each other from going down to the end or something.

That would make sense, given that something seems to be blocking the cables from getting the full number of blocks. Though the average wouldn't speak to a 13-channel setup like I documented earlier in the thread.

I'm not tied to the 3x3x3 setup, but condensing the blocks does make them more efficient for covering interfaces with assemblers.

It seems pretty efficient also to make a 3x3 tower with 4 interfaces in the cross pattern all the way up, then make connections every other time. That seems to distribute the channels properly and keeps a pretty high interface efficiency. Probably better than what I had before anyway. I don't know which setup I'll try to use in my survival world, but this at least gives me some ideas and options.

 

[adamich]

Are you talking about the connection from the cable into the block itself, or the cable itself as distinguished from the connection?

I am talking about 14 interfaces in cube 3x3x3. AE cant connect it correctly, U need chop it in two pieces.