Cable splicing/packet splicing?

velusip · Mar 13th 2012

Hello. I have some questions about how cable splicing works in IC2.

Does a transformer always send maximum "voltage"? Will a transformer outputting 128"V" ever send a partially filled packet? Even when cable splitting is next in line?

Does the EU within the packet get divided at a cable split? e.g. Does a
128"V" packet containing 128 EU get split into two 128"V" packets each
containing 64EU? (128EU/128"V" -> split -> 64EU/128"V" + 64EU/128"V")

Must you always place an EU storage machine before a cable splice to ensure only appropriately filled packets are sent to requesting machines? e.g. If two machines want different levels of power after a splice, will the cable split offer precisely enough EU to each machine so long as the origin is a storage block? Or will both machines get equally filled packets with wastage on one line?

What happens when a transformer is before the split and only one machine of the the split lines requests power? Does the large packet get split up evenly and the remaining EU is lost on the non-requesting machine? Or do the split packets contain variable amounts of EU, effectively not splitting the EU down the non-requesting line? Should a feedback loop be installed to counter any loss after splits?

How does power pass through a split when each cable connects to a machine that requires different amounts of EU? Are the EU amounts variable after a split?

I'm just trying to account for a lot of missing power even after all cable loss is calculated in a very massive grid. I was trying to use as few storage devices as possible to simulate brownouts in an active grid, but the power usage is unreal and I have a feeling the cable splices after transformers are to blame.

FenixR · Mar 13th 2012

tl;dr...

Anyway from what little i actually read, transformers actually had some internal buffer, so if the machines on the other side of the lv transformer only ask for 8 eu, the remaining 120 eu stays inside the transformer until its fully depleted... i think? Forgot when i tested this.

velusip · Mar 13th 2012

Thanks FenixR, your reply is much appreciated.

One more I forgot to include on the topic of your reply... if a requesting machine is connected to a LV transformer which is fed 128 and outputting 32"V" packets, is there ever a time when four 32"V" packets are not sent? e.g. will the LV Transformer hold three 32"V" packets and just send the one until more requests are made? Or will 128"V" packets always output as four 32"V" packets leaving 3 of them to disappear if not needed that tick?

velusip · Mar 13th 2012

I built this auto shutoff circuit to prevent large packets of EU from being wasted to fill that last little bit of an EU storage block as it is being slowly drained from another consumer block. Perhaps it is unecessary?

Altaree · Mar 14th 2012

the shut off circuit will save energy if you have loss on the cable. Loss is charged per packet. If you lose 4 eu on the cable, you will lose 4 eu if you are sending a packet of 5eu or a packet of 512 eu....

velusip · Mar 14th 2012

Quote from Altaree

the shut off circuit will save energy if you have loss on the cable. Loss is charged per packet. If you lose 4 eu on the cable, you will lose 4 eu if you are sending a packet of 5eu or a packet of 512 eu....

Yes, the circuit is intended to prevent EU wastage when transfering large, full packets of EU out of a transformer. This is assuming that transformers always send full packets, unlike storage blocks which can send partially filled packets.

Assumptions:
- HV Transformer (configured EV to HV) transmits 512"V" packets containing 512EU ALWAYS. (yes? no?)
- MFSU transmits 512"V" packets containing anywhere from 1EU to 512EU depending on the request. (yes? no?)

An example would be a Battbox sends a 32"V" packet containing 1EU to an induction furnace to keep it hot. The Battbox is now not full and makes a request up it's supplying cable for more power. Since it is not connected directly to another storage block -- which would have sent a single EU of power this tick -- it instead requests a full 32"V" packet containing 32EU from a transformer. 31EU is wasted in this request. Again, this is the assumption about transformers that I need confirmed.

The topic my thread was initially asking has many prerequisites, so I'll try to get them answered in order.

Meanwhile, my testing is not going well because it's so very miniscule and requires huge samples to determine.

biomatter · Nov 28th 2012

Necroposting because this guy asks a lot of really good posts about cable splitting in IC2. I'm very familiar with IC1 cables, but these newfangled systems have me beguiled, the wiki has very little information, and I've already read nearly a dozen topics that don't even come close to answering the tough questions.

To add some fuel to the fire: If an Induction furnace needs only 1 EU/t to keep warm, where the heck am I going to get a machine that outputs a measly 1 EU? Off the top of my head all machines output at least 32 EU/t. The way I see this playing out is: Induction Furnace drops from max energy to (max - 1) and then requests more energy, gets sent a large packet, and bam. Wasted energy. Please tell me MFE's/etc. can output less, through some arcane means, than their displayed output.

To restate some of TC's questions in english:

If there is a split in a cable and only one machine on one end requests power, does the packet that gets sent out in response get split in half, or does it all go down the 'right' path?

Actually, the more I think about it, the more that question gets to the heart of it all. If the e-net is 'smart' and it always sends the right amount of EUP down the right path, well, life will be just dandy! I bet we could all get rid of a lot of buffer boxes and other shenanigans. But if the e-net isn't 'smart'... Well... I've got a hell of a lot of work ahead of me getting a small windmill farm to supply energy to over a dozen different machines/areas as efficiently as possible. I'd *really* rather not have to stuff transformers, MFEs, and more transformers at every single split and output.

Any takers?

narc · Nov 28th 2012

Quote from biomatter

Induction Furnace drops from max energy to (max - 1) and then requests more energy, gets sent a large packet, and bam. Wasted energy.

No, no. Induction Furnace drops from maxEnergy to (maxEnergy - 1), states that it can accept power, receives a 32 EU packet, and stores (maxEnergy + 31) EUs. Or, if it got a 128 EU packet, it stores (maxEnergy + 127). The furnace will not accept power anymore until it drops below maxEnergy again. All electric blocks that can accept power work this way, and this carries over correctly to machines with transformer upgrades (yes, maxEnergy + 511 is perfectly possible. I'm not sure if you can force EV into a machine, but if it were, it could go to maxEnergy + 2047).

Quote from biomatter

If there is a split in a cable and only one machine on one end requests power, does the packet that gets sent out in response get split in half, or does it all go down the 'right' path?

The packet goes to the machine that flagged itself as accepting power. I'm not entirely sure what happens when there are multiple machines accepting power, but I think (from glancing at some nigh-unreadable decompiled source) that multiple packets get sent, as many as necessary to fulfill demand.

Share