Changing cable length power losses?

[Stevolution]

OK... highly unlikely, but its worth asking.

After building my spanky nuc reactor and power station... it doesn't work due to the (ridiculous) amount of power loss in a HV cable.

My MFSU's are not charging and its because the HV cable is too long. I can't shorten it anymore, and moving the HV transformers nearer the reactor basically means trashing the power station and starting again.

Can't go glass fibre... doesn't take the power.

The cable losses/length are not something that can be changed are they?

The output from the reactor feeds 20x MFSU units (they also are charged by wind and solar). Will one HV transformer be enough on the output of the reactor anyway?

[Hydra]

How long is your cable? Did you transform the power up to HV?

[Stevolution]

Transform up? My HV cable leaves the reactor and goes to the HV transformer that is approx - 50/60 blocks away (yes... I know... not ideal)

You mean reverse connect a HV transformer in reverse (and apply redstone signal I think) and then reverse it again at the other end? Does that overcome the EU drop?

I knew I would suffer large losses, but I didn't think it would cancel out the entire output of the reactor.

If I break the 3x HV cable at various points along its length and stick a HV transformer+MFSU onto the end - that charges, so it must be cable length

[Stevolution]

Hydra - thank you for that. Reversed a HV transformer + lever and that appears to be working.

Didn't know that was possible.

[Avalkiz]

Reactor ->HV cable -> HV Transformer -> Glass fiber cable -> MFSUs -> Power lines (Glass fiber).

In our nuclear power plant we use this system and it works fine.

HV cable can take up to 2048EU/t in UHV. Glass fiber cables fry with UHV, but they work fine with HV. I was testing that system and i was able to send over 50000EU/t in HV through a single glass fiber cable. Don't transport HV voltages in HV wire, it's not worth it. HV wire's loss with UHV is 1EU/t per block, and for HV the loss is around 70EU/t per block!

[Stevolution]

Mmm OK.

My setup now is Reactor -> 3xins cable-> step up HV transformer -> long 3xins HV cable -> step down HV transformer -> MFSU -> output to various items via glass fibre

Seems to work, but I have no idea what the losses are.

Oh, and wow.... getting a transformer back to front with the machines still connected..... makes a pretty damn good mess of your control room and your pants (must turn the sound down a bit)

[Watchful11]

Interesting,

HV cable loses 1 eu every block, 1 eu every 1.25 blocks for 3xins cable.

If you lose all the power in 50/60 blocks, you would probably only be outputting 50/60 eu's.

So straight into fiber cable should be fine.

Do you know the power output of your reactor configuration?

[gavjenks]

HV cable does not lose very much power at all. (Pro tip: Do not take everything that some random dude wrote as commentary on the wiki as gospel)

With your 50-60 (let's say 55) block long cable, operating at HV voltage, you would lose:

55 / 1.25 Eus lost per block per packet = 44 EUs lost per packet.

Packet size = 512, so 44/512 = 8.6% of energy should be lost. I would not call those "ridiculous" losses. But worrisome losses, yes.

If you step up the voltage to Extreme Voltage (by having two HV transformers, the first with a redstone current applied), then it becomes

44 / 2048 Eus per packet = 2.1% of energy lost. Totally reasonable.

With glass fiber cable, you'd lose

1 Eu lost (for the first 44 blocks) / 512 = 0.2% of energy lost

However, the cost of that cable would be 10 diamonds and 20 silver ingots, versus 15 iron. In my opinion, the extra diamonds would very rarely be worth a 2% difference in power savings. In terms of UU, you'd have to save 8*9*1,000,000 EU to fabricate those diamonds = 72 million EU, which means your reactor would have to run long enough to produce about 3.5 billion EU (72 million x 50, since the savings are only 2%) to make up for the cable cost. This comes out to 24 hours of runtime for a 2000 Eu/tick reactor, or almost 7 full batches of uranium = 370 uranium cells burned before you make back the extra cable cost of using glass fiber...

For comparison, if you were to step the power down and use gold or copper cable (which is entirely possible), you would lose:

GOLD: (55/3)/128 = 14.3% of energy lost

COPPER: (55/5)/32 = 34.4% of energy lost

Again, unlike the wiki claims, HV cable is actually much more efficient than copper or gold, like you would expect (and like in real life for high voltages).

[Watchful11]

Good writeup, i hadn't thought about it that way.

However,

Reactors can be an exception to this rule.

If it was a low power reactor and only put out 50 eu packets, you would lose all the power in 62.5 blocks.

[gavjenks]

Good writeup, i hadn't thought about it that way.

However,

Reactors can be an exception to this rule.

If it was a low power reactor and only put out 50 eu packets, you would lose all the power in 62.5 blocks.

No, this is incorrect. It would still send 512 or 2048 EU packets.

If a reactor is hooked up to an HV cable, it will always send 512 EU packets. If it produces 50 EU / tick, it will simply wait for 11 ticks until it has built up 512 Eu before it sends one packet.

If a reactor is hooked up to an HV cable stepped up to EV, then the transformer will collect EUs until it has a total of at least 2048, before it sends out a packet along the EV line. Even a single solar panel + HV transofmer system would wait over a minute and a half before sending a single 2048 EU packet over an EV line.

Thus,if you have a power system it hooked up to any of the types of lines I described above, the power losses would be exactly what I have written, no matter what the power source or its output is.

In fact, I don't even think there is such a thing as a 50 EU packet. Don't quote me on this, but I am guessing that the only possible packet sizes are:

5

32

128

512

2048

And that a machine will send packets in whatever the smallest packet size is that is capable of containing its entire output for that one tick possibly after dividing by the number of different energy destinations (thus, a 2400 EU/tick reactor would look first to see if more than one cable is attached to it, and then divide by that many before determining packet size, which makes it possible to still hook up two separate HV cables without them exploding).

I also tested this last rule just now on SSP with a MV solar array attached to 2 batboxes. It did NOT explode, because it divided its output by 2 prior to determining packet size. 64/2 =32. The smallest valid packet capable of holding 32 Eu/tick is a 32 size packet, which batboxes can handle, so they don't blow up, even though there's no transformer. One batbox attached to an MV array will blow up, though.

Similarly, 4 mass fabricators attached to one MV solar array with copper cables will not blow up, either.

[gavjenks]

Lol, I just also ran a test where I hooked up an HV solar array to a shitload of batboxes (over 110) with tin cable, and they did not explode. But the tin cable was carrying so many 5 EU packets through it per tick that it still almost electrocuted me to death when I walked near it.

[Watchful11]

I agree with you about the packets dividing.

But i just tried this.

Had a reactor with 2 cells in it. Produces 40 eu/t.

Clicking as fast as possible with a eu reader, i got 40 eu/t averaged over 4 ticks.

If it built up to 128, taking 3.2 ticks, then it would have read 32 eu/t averaged over 4 ticks.

But if I let it average over more ticks it would read the correct value.

I'm not saying your wrong, if your reading the code i bet your right, but I'm curious where your source is for this.

[gavjenks]

The Eu reader is horribly broken, I've found. I wouldn't trust it at all. It frequently gives negative numbers and all kinds of crazy things.

You should run the same experiment, except use like a low voltage array and HV cable, and then hook it up to an MFSU, then actually watch the MFSU to visually see if it comes in bunches of 512 or trickles in 8 at a time. This would be more reliable and would not suffer from the possibility that the Eu reader calculates things differently than the cables do (which I think it does).

Also, fun screenshot of a new test I just did:

With 4 destinations for power, the geos split 20 Eu/tick into 4 packets of 5, thus not burning the tin cable. Even when there are multiples of them. So every tick, there's actually 20 EU moving through each tin cable there, in 4 packets of 5. This could be very useful to use in actual bases, because tin cable:

1) Only loses 1 EU/packet/40 blocks

2) Is carrying about as much as a copper cable here (and could carry more)

3) is very very cheap and uses no rubber.

So you can actually save more energy over distances, for fewer resources, and no rubber

Edit: this is a better setup that also works. That single tin cable is carrying 120 EU/tick, and can go as far as you want, though you would need to add 4 batboxes every 40 blocks to not lose energy (instead of one)

Edit #2: A long long tin cable with no energy losses carrying 240 EU/tick (high voltage) to a mass fabricator. Example of system in use:

This would be especially useful for a wind farm. Attach every windmill to a central tin line, and just have 2 batboxes for every booster and at the end (ground). Thus, even if a windmill gets above 5 EU/tick during a thunderstorm, it will split it into 5 Eu/t packets, and not fry your cable. And you can get all the energy to the ground with no losses and very very minimal expense.

[Watchful11]

The Eu reader is horribly broken, I've found. I wouldn't trust it at all. It frequently gives negative numbers and all kinds of crazy things.

You should run the same experiment, except use like a low voltage array and HV cable, and then hook it up to an MFSU, then actually watch the MFSU to visually see if it comes in bunches of 512 or trickles in 8 at a time. This would be more reliable and would not suffer from the possibility that the Eu reader calculates things differently than the cables do (which I think it does).

Just did that.

Watching the mfsu it is clear that there is not a 512 eu packet coming every 3.2 seconds.

While it is hard to be precise, it did look like it was 8 eu packet every tick.

On the tin cable.

That is extremely useful and very helpful. I'll be using that a lot.

[Stevolution]

Oh... now I am lost!

Don't quite understand (yet) all this EU/packet stuff.

My reactor is currently used at a pathetic, yet totally safe 120EU output. However, its built like fort Knox with a computerised system for altering its configuration + ice generation etc.

I intend to wind that output up far higher than it is now.

My warehouse of future machinery isn't built yet, so I don't need the high output.

I basically only ever use glass fibre cable or 3x ins (occasionally 2xins) The HV cables are only ever on the reactor side of things.

Using a reversed HV transformer has boosted the juice high enough to reach the sub-station (68 blocks away). And if you click the supply cable at the sub-station end with the glitchy EU meter, I get a pathetic EU per tick count (averages about 2 or 3).

But... at least it gets there.

If you look inside the 10 MFSU units its feeding, they are charging at a rate of 512 EU every 2 seconds.

The increase count is exactly 512

I don't understand the need t max out something like tin cable, when a glass fibre would just be far better (unless resources are short I suppose)

How do I get images up on here guys?

[gavjenks]

Actually, I made a mistake with my description of the wind farm.

Batboxes output 32 Eu/tick, so once you have the tin go through the first set of batboxes, every relay station after that would actually need SEVEN batboxes, unfortunately, to split the 32 from each of the first batboxes into 5 Eu/tick packets.

So, not as terribly great as I thought at first. But still better than glass and probably slightly cheaper than HV cable.

[Watchful11]

How do I get images up on here guys?

Use something like, http://postimage.org/, then hit the insert image button and put in the link.

If your reactor is outputting 120 eu/t you should be just fine using fiber the whole way. However, if you will be ramping up the power soon you should do something like this.

The reactor splits its power to 4 transformers. Each one sends HV to the MFSU.

gavjenks, the one problem i see is,

You would have to be very careful that the batboxes fill up at the same time. If one gets full, then the packet size would increase and blow the cable.

[gavjenks]

gavjenks, the one problem i see is,

You would have to be very careful that the batboxes fill up at the same time. If one gets full, then the packet size would increase and blow the cable.

That is a good point... In the mass fab example above, it never fills up (as long as you dont let it fill up with 64 UU matter). But otherwise yes. You could probably use one extra batbox at the final end of your setup, and then have a detector cable hooked to a wireless transmitter and a splitter cable with a receiver back at the beginning?

Or you could possibly use ccSensors to figure out when the MFSU they all flow into at the very end is almost full (95%), then turn off the splitter at the beginning until it drains energy again, so nothing will ever fill.

I'm not sure why you would really want to move power over hundreds of blocks anyway, though =P

[Stevolution]

I think I am set up fine then. Watchful11 - that is basically my setup

http://postimage.org/image/t1xilrfed/

http://postimage.org/image/93bygxgld/

http://postimage.org/image/wbau3ierl/

http://postimage.org/image/uquxr7n0f/

My 3xins HV cable is stepped up as it leaves the reactor, then is stepped back down at the sub-station.

The substation consists of 10x HV transformers (converting the power back down from being stepped up), and each HV transformer feeds its own MFSU.

Then in turn via galss fibre, they are stepped down further.

All seems to be working fine now, with provision for future ramping up of the reactor power.

Oh - and in the pics, there are actually three substations. One is wind, one is solar and one is nuclear.

[Watchful11]

stevolution, sorry, i cant make anything out in that pic, its too dark. If you havn't, try setting brightness up to bright.

gavjenks, agreed, i happen to have a couple 100 block plus wires, but its a kinda extreme case.

Except possible in the example you used of the windfarm.

[Stevolution]

Added a couple more piccies.. but still hard to work out what is what.

Anyway, not to worry Watchful, a solution has been found and I'm a happy bunny.

[Stevolution]

This might be better..

http://postimage.org/image/u4ai0h9kx/

http://postimage.org/image/mysurxkrf/

Bottom is the 3x ins cable. Above that are the HV transformers, then glass fibre up into the bottom of the MFSU's

Colour coded galss fibre out to the MV transformers etc

Thanks for everyone's assistance