Water Tower: Will this cause lag?

[ShneekeyTheLost]

Aye it is. Watermills are a 3x3x3 cube, so the water in the diagonal corners count as water.

It would still significantly increase the efficiency of your design to use the standard four mill surrounding cable design.

You will practically double the power output of each tower if you have a layer of water between 'segments', because then you'll have the 3 x 3 (except the 1 in the middle for cable) above and below the water tower also generating power. This is far more energy efficient than spamming them on every level.

As regarding lag, it has a good chance to generate lag. Every packet has to be tracked. So if you have 1,000 towers (5-255 x 4), that will be 1,000 packets the computer will be trying to keep track of. THAT many packets CAN cause lag.

You are using MFSU's as relays? That's not really an optimal use of resources. Using transformers would do the same job, and would be FAR cheaper. And using MFSU's would 'throttle' your output. They can only output 512 EU/t. So the bottom MFSU, regardless of how many towers are above it, would only output that amount. If you want to realize more than this, I suggest you use Transformers instead.

By that same token, don't forget to have more than one MFSU in serial receiving input and providing output.

 

[PoisonWolf]

It would still significantly increase the efficiency of your design to use the standard four mill surrounding cable design.

You will practically double the power output of each tower if you have a layer of water between 'segments', because then you'll have the 3 x 3 (except the 1 in the middle for cable) above and below the water tower also generating power. This is far more energy efficient than spamming them on every level.

As regarding lag, it has a good chance to generate lag. Every packet has to be tracked. So if you have 1,000 towers (5-255 x 4), that will be 1,000 packets the computer will be trying to keep track of. THAT many packets CAN cause lag.

You are using MFSU's as relays? That's not really an optimal use of resources. Using transformers would do the same job, and would be FAR cheaper. And using MFSU's would 'throttle' your output. They can only output 512 EU/t. So the bottom MFSU, regardless of how many towers are above it, would only output that amount. If you want to realize more than this, I suggest you use Transformers instead.

By that same token, don't forget to have more than one MFSU in serial receiving input and providing output.

What are the size of watermill packets anyway? 32eu/p?

 

[Loufmier]

What are the size of watermill packets anyway? 32eu/p?

i believe it`s about 1-2 eu/t

 

[ShneekeyTheLost]

What are the size of watermill packets anyway? 32eu/p?

The size of the individual packet is the same size as the amount of EU/t it is producing, so less than 1 EU/t. At least until it hits something which would consolidate that, such as an MFSU. The MFSU would collect all the multitudinous packets, then spit out one HV packet at 512 EU/t. However, by that same token, it throttles it to 512 EU/t.

My suggestion to you is to have multiple towers producing 512 EU/t, roughly half the size you are currently pondering. Each one can then feed into an MFSU, which can then output HV packets to wherever it would like.

For more information on how EU works, I suggest you read my QuickTip on the subject.

 

[AlanEsh]

What is the block that looks like a cage and has water piped into it?

 

[Numkin]

It would still significantly increase the efficiency of your design to use the standard four mill surrounding cable design.

I'll give the math i've done on this, it is quite possible i've done the math incorrectly.


~ ~ ~ ~ ~ ~ ~ ~ ~
~ ~ ~ ~ w ~ ~ ~ ~
~ ~ ~ ~ o ~ ~ ~ ~
~ ~ w o w o w ~ ~
~ ~ ~ ~ o ~ ~ ~ ~
~ ~ ~ ~ w ~ ~ ~ ~
~ ~ ~ ~ ~ ~ ~ ~ ~

Assuming this is on 20 levels and this is continued one level above and below the 20 levels ( but will not include Eu/T)

For outside Mills:
7 Water blocks per level
3 levels per mill
21 water blocks per mill
.01 eu/t per water block

.21 eu/t per mill
4 outside mills:
outside mill total: .84eu/t

For inside Mill:
4 Water blocks per level
.12 Eu/T per inside mill:

.84 + .12 = .96eu/t per layer

.96*20 = 19.2 Eu/t for 20 layers
_________________________________________________________________________________________
~ ~ ~ ~ ~ | |~ ~ ~ ~ ~
~ ~ w ~ ~ | |~ ~ ~ ~ ~
~ w o w ~ | |~ ~ o ~ ~
~ ~ w ~ ~ | |~ ~ ~ ~ ~
~ ~ ~ ~ ~ | |~ ~ ~ ~ ~

Assuming this is repeated every 20 levels and continued one level above and below 20 levels (but will not include Eu/T)

On layer with mills:
5 blocks of water = .05 Eu/t

On layers with no mills:

9 x 2 blocks of water = .18 Eu/t

.05 + .18 = .23 * 4 = .92 Eu/t every 2 layers

.92*20*.5(only generates eu every other layer) = 9.2Eu/t for 20 layers

Recap:

My current way 19.2 Eu/t for 20 layers
Other way 9.2 Eu/t for 20 layers.

Again I might be missing something on this.

You are using MFSU's as relays? That's not really an optimal use of resources. Using transformers would do the same job, and would be FAR cheaper. And using MFSU's would 'throttle' your output. They can only output 512 EU/t. So the bottom MFSU, regardless of how many towers are above it, would only output that amount. If you want to realize more than this, I suggest you use Transformers instead.

Thought about using transformers, but wasn't sure how to link those since they constantly upgrading the signal, and trying to put redstone would just be annoying, but I would like to use them if there is a trick to using them.

As far as it being a bottle neck, I'm not sure I fully understand the Packets, but MFSU does a max a 512 Eu/t, a single water tower won't generate that much Eu/T at max it could generate under 250eu/t. There are also 4 lines, which means the Eu/T is split roughly 4 ways. Way I understood it I technically could have used an MFE and still had no bottle neck. But then does a water mill transmit in .12Eu packets or does it wait till it has a 32Eu packet? If it waits that adds a bit of complexity to the system I had not thought.

By that same token, don't forget to have more than one MFSU in serial receiving input and providing output.

Not quite sure what you mean by this. Have the MFSU stacked? or have more than 1 for the entire tower?
Thanks for all your input, gives me ways to think of building it differently.[DOUBLEPOST=1375291011][/DOUBLEPOST]

What is the block that looks like a cage and has water piped into it?

That is a redpower 2 grate. Amazing for filling up towers!

 

[ShneekeyTheLost]

Your problem is that all those cables are going to be vertically-connecting on every layer. This is a Very Bad Thing that will fry your setup. Which means you need a gap between each layer anyways. Which means, if you are *insisting* on doing it that way, to do this instead:

~~~~~~~|
~~~W~~~|
~~~c~~~|
~WctcW~|
~~~c~~~|
~~~W~~~|

W=Water Mill
c = Cable
t=Transformer

You still get maximum water coverage without that silly fifth wheel.

OR, if you want to QUARTER your cable usage, you go with

~~~~~
~~W~~
~WcW~
~~W~~
~~~~~

Now you only get .6 EU/t/layer, or 12 EU/20 layers, however your resource investment is SIGNIFICANTLY less. And the footprint is only 5 x 5.

 

[Bigglesworth]

For me, I typically avoid large intricate structures not out of fear of complexity, but I generally just hate unnecessary calculations when there is a clearer path that is computationally leaner.

Might not be the best way to get into a game about creativity, but hey as long as youre having fun.

Its sort of like an artist/engineer limiting himself by amount of paint/tools he can use

 

[ShneekeyTheLost]

Also, instead of building one huge tower, you might get a lot more use out of several smaller ones. Let me show you what I mean...

Take, for example, the tower I just posted with .6EU/t/layer. Say you run that thing 38 blocks. At this point, you can use TIN cabling, because everything is producing less than 1 EU/t.That's 22.8 EU/t per 'micro tower'.

Now then, you build a micro-tower feeding UP into an MFE, then another micro-tower feeding DOWN into the MFE. You've got an 80 block structure which is producing you 45.6 EU/t.

Right, now duplicate that three more times, one on each cardinal direction from an MFSU. Now we have 182.4 EU/t generation. You can also build one above the MFSU, with the 'output' facing down, for a total of 228 EU/t.

If you don't mind horizontal footprint, each of the MFE's can double their output by having a pair of micro-towers on either side perpendicular to the side feeding into the MFSU.

Now then, let's look at resources:

For 228 EU/t, you have 5 MFE's and one MFSU. You've got about twelve glass fibre cable, the rest is just tin.That's four diamonds per MFE, and another 10 for the MFSU, and two for the Glass Fibre Cable. That's a total of 16 diamonds.

Just running your four Glass Fibre Cable lines would cost you more diamonds than that, not even counting the MFSU's you'll be periodically dropping in.

 

[Zelfana]

But then does a water mill transmit in .12Eu packets or does it wait till it has a 32Eu packet? If it waits that adds a bit of complexity to the system I had not thought.

Water mills only output 1-2 EU packets or else you wouldn't be able to use tin cables that only support 5 EU/p.

 

[ShneekeyTheLost]

Water mills only output 1-2 EU packets or else you wouldn't be able to use tin cables that only support 5 EU/p.

In this case, water mills are only outputting 0.15 EU/t, so that isn't an issue. You're using Fibre Cabling from the MFE's to the MFSU.

However, if he is using MFSU's as repeaters, he's going to have to use Glass Fibre Cable 'downstream' because they output HV at 512 EU/t.

 

[namiasdf]

Unless you route the power out of that tower every 39 blocks, you will lose all power from watermills further than that length. They generate 1 EU/t and the loss after 40 blocks using glass fibre is 1 EU.

 

[ShneekeyTheLost]

Unless you route the power out of that tower every 39 blocks, you will lose all power from watermills further than that length. They generate 1 EU/t and the loss after 40 blocks using glass fibre is 1 EU.

Which is why my 'micro towers' are 38 blocks tall

I think his plan is to place an MFSU every 40 blocks facing down to act as an energy repeater so EU output is not lost.

 

[Saice]

Every block that ticks and injects power into an IC2 network introduces tick lag. As you build more, the lag goes up. So yes, this COULD cause lag. Make sure to monitor this with a ticklag measurement mod like BechelliTPS.

This man speaks the truth.

But there more to it. IC2 power can cause lag in other ways to. But there are ways to avoid this lag.

First you have to know every time you have a branch in wiring IC2 does some math to decided where to send the packet. Next you need to now that IC2 sends a packet every time the generator has something to send. So you have 200 Water mills that is 200 packets every time they are ready (normally this will be every tick). This is vary important as it is these root cause of much mystery lag from IC2 power systems. And here is a small list of ways to avoid this.

1. Replace jointed wiring with transformers or storage(like a bat box). I prefer using storage near the power system because it adds a buffer. The idea is you want to gather up all them tiny packets and make them one bigger packet. Less total packets but the same power. This means less math for IC2 down the line.
2. NO LOOPS. You never want a wire to loop back around in cases where there is no place for the power to go this can cause a tiny lag machine.
3. Step up your power when you can. This goes back to 1 when you have enough more to go to the next voltage tier it is normally a good idea to do so. Less total packets less possible lag.

Now in most small systems none of this will really be noticeable but something like your build could start to induce lag so might as well get into the habit of building IC2 power grids right.

 

[Numkin]

Also, instead of building one huge tower, you might get a lot more use out of several smaller ones

Aye, this is true. The half of the tower that's underground no one sees. I could drop the height down to only about 65 layers tall and have 16 visible towers instead of 4 253 layers. I don't really want the towers to be shorter than my pyramid. I could do 1 tower on each corner and 2 in the middle walls where my doors are.

As far as resources go I'm not all to worried about those. I have over 900 diamonds right now, and have 4 quarrys running at 60 MJ/t each. All the mats are piped into a gold/diamond transport pipes sorting system using an item tessaracts. The biggest slow down I have is sand and tin (don't know why tin is so hard to get) and copper, but thats just until we are done with the nuclear reactor. Aluminum isn't bad, and I already have enough iron/redstone to make all of the mills I need.

1. Replace jointed wiring with transformers or storage(like a bat box). I prefer using storage near the power system because it adds a buffer. The idea is you want to gather up all them tiny packets and make them one bigger packet. Less total packets but the same power. This means less math for IC2 down the line.
2. NO LOOPS. You never want a wire to loop back around in cases where there is no place for the power to go this can cause a tiny lag machine.
3. Step up your power when you can. This goes back to 1 when you have enough more to go to the next voltage tier it is normally a good idea to do so. Less total packets less possible lag.

Thanks! I will follow these guide lines.

I think his plan is to place an MFSU every 40 blocks facing down to act as an energy repeater so EU output is not lost.

Aye, though its also an energy buffer to some degree. Since I will be using a teleport system, from what I have read, they use a lot of power really fast, though it won't be used that often, I want to make sure I have enough power to run the other things in the building. I plan to have the teleport system on a power sub-net grid, so if it happens to run low it won't kill the lights or other things.

 

[Saice]

I'm a big fan of passive power gen on server play. Chunk loaded means you getting power even when your not around. I've built some fairly rediculus systems using thermals, solars, and wind based system. All low power but you move that to a batery bank of some sort and you can build up a good amount of power. Tie that into a smaller on demaind system and you can have a fairly good power grid.

Just remember to turn off power consumers before you log out.

 

[KirinDave]

Never forget the humble MFSU cart. It can transfer a ton of power an infinite distance with 0 loss and even service multiple endpoints if you set it up right.

 

[Numkin]

I'm a big fan of passive power gen on server play. Chunk loaded means you getting power even when your not around. I've built some fairly rediculus systems using thermals, solars, and wind based system. All low power but you move that to a batery bank of some sort and you can build up a good amount of power. Tie that into a smaller on demaind system and you can have a fairly good power grid.

I like this way also. I'm really the only technical involved person on the server, so simplicity is a must. Lava is great imo, because I can give them a liquid tessaract/magmatic engine and they are set, until the lava runs out if I haven't been on in a day or 2. With this system, they can make their home around my building, tap into the grid and Never have to worry about power and the big towers just are nice bonus to the area.

Just remember to turn off power consumers before you log out.

Ha, I have my Matter Fab running constantly right now, but I might just leave it on my old power grid and let it use all of that energy, hmmm.

Never forget the humble MFSU cart. It can transfer a ton of power an infinite distance with 0 loss and even service multiple endpoints if you set it up right.

I have never tried or really thought about this.... interesting, I'm guessing using portals of some kind.

 

[Zelfana]

Inspired by the discussion I decided to make a quick example of maximum efficiency and the least lag possible, just add water:

 

[Numkin]

Cannot argue with that, more eu per water mill and least expensive mats for job