? on EU Insufficient energy when below max output

[snowmirage]

I thought I had a solid understanding of EU but it seems maybe not..

I have a bank of various batteries (1 IDSU, 4 AESU, ~24 or so MFSUs) the IDSU and AESU are connected to HV transformers and it all feeds into my base power line over glass fiber cable.

For the sake of simplicity lets pretend I have 20 MFSU's each puts out 1 512 EU packet per tick (correct me if I'm wrong anywhere here) so each ticket I could be pulling form this battery back into my base a maximum of 10240 EU (512 EU x 20 MFSU's).

My current setup is capable of sending 10000+ EU per tick into my bases power grid, I've seen it do it (EU reader on the one cable leading into the base)

Here's where I'm getting lost....

I'm building a massive array of machines to supply the fuel for the new fusion reactor. (128+ industrial centrifuges and others)

I connect everything and I'm getting "Insufficient energy line" but when I check the power I'm pulling from my battery array I'm only pulling out ~4000-5000 eu per tick....

The only wrench I threw into this was instead of having to step up / down the power from my base (which is HV at 512) for the machines I used transformer upgrades to allow them all to access HV without blowing up.

From my understanding putting on a few transformer upgrades does NOT change the power consumption of the machines just allows them to be connected to a different voltage line is this correct?

I'm kind of at a loss here to explain why I'm getting Insufficient energy when I should have plenty of energy output potential left....

Any ideas / thoughts would be appreciated

 

[namiasdf]

Are you centrifuges, transformer upgraded or are you using HV->MV->LV as your strategy. Remember that LV face of the LV transformer can only output 32 EU/t. The non-upgraded energy draw of a centrifuge is around 5 EU/t, so to be safe, you can power 5 centrifuges off one face of a LV transformer.

 

[snowmirage]

They all have transformer upgrades to run directly off my bases HV power line

there is nothing between the machines and the MFSU battery bank other than glass fiber cable, wih the exception of the IDSU and a few AESUs which use HV transformers to step down 2048 EU/t to 512 EU/t. Its a long way away maybe 100 blocks or so but that shouldn't be an issue with glass fiber[DOUBLEPOST=1374914942][/DOUBLEPOST]And by the way none of them are overclocked at all just transformer upgrades

 

[namiasdf]

Well the number of 2048->512 step downs you have in parallel, will determine your output regardless of the number of MFSU's you have. The MFSU's only act as an energy buffer, but considering that you will need continuous production for the deuterium/tritium, the buffer isn't required.

Take out the MFSU's, unless you have an energy positive system (i.e. you supply more energy that your system consumes).

It should look like this:

[IDSU/AESU] -> [HVxform] -> [MFSUs] -> [centrifuges]

If you have 20 MFSUs, you will need 21 MV faces from HV-transformers, feeding into your MFSUs with everything connected in parallel.

I would just make 7 HV transformers, all in a row. The two end ones will can have four outputs (side, front, top, bottom) the five middle ones will have three outputs (front, top, bottom), giving you a grand total of 23x512 EU/t output faces connected in parallel to your 20 MFSUs which are only capable of outputting 20x512 EU/t packets.

This will ensure your energy draw from your main energy system is greater than the output capabilities of your energy buffer, which makes sense. If you still get energy insufficient messages, it means you don't have the EU/t output required to power everything.

 

[snowmirage]

Hmm I must be missing something...

What is the difference between ...

An IDSU outputting 2048 EU/t into an HV transformer which steps it down to 512 EU/t making a total of 4 512 EU packets each tick

and

4 MFSUs outputting 512 EU/t each again making it 4 512 EU packets each tick

If you fed the 4 MFSU's down into one fiber cable and compared that to the cable coming out one side of the HV transformer connected to the IDSU wouldn't it be the same thing essentially?



And for my system I do generate more power than I use. For the purposes of the discussion we can assume all the batteries start fully charged.

 

[namiasdf]

Well, you'd need four sides of the HV transformer connected to the line going to your MFSUs, but yes that should be 2048 EU/t in 4x512 EU/t packets.

 

[snowmirage]

My MFSU's and AESU and IDSU are all connected the same way so its....


[ Power generation ] connected to [ bank of all batteries ] connected to Base machines that need power


From what I've seen I don't see a need to connect multiple sides of an HV transformer. I have an AESU outputting 2048 eu /t connected to a HV transformer with 1 side connected to glass fiber cable

if I draw enough power and do a reading with the EU reader on that cable it will read up to 2048 EU/t (the HV transformer breaks the single 2048 EU packet into 4 512 EU packets and spits all 4 packets into the glass fiber cable at once)

So i'm not sure why you would ever need to connect more than one side of the HV transformer unless it was just for convenience.

Unless I'm missing something still.


The only thing I can think of here is its something to do with the Number of packets ..... but I don't understand how they are injested by the machines yet.. It can't be that each machine requires its own packet. I'm outputting like 36 or so 512 eu packets each tick and I have WAY WAY more than 36 machines going at a time....

 

[namiasdf]

Hmmm, well from what I've been doing, attaching another side, or adding another transformer fixes the insufficient energy problem.

I see where you logic is coming from and though I agree with what you're saying, in how EU works, how infinite EU/t can go through any cable as so long as the packets are the right size... It would seem that the throughput of a transformer is what causes insufficient energy messages, aside from having insufficient energy. The solution in most of my cases is to have more than one side attached to my output line.

Take for example my GT grinder. I have three of them. If I output from one face of my HV transformer, to the three GT grinders... If all three are running at the same time, I will get an insufficient energy message. Attach two more faces to the output line and voila, problem solved.

It's a conflict between knowing that infinite EU/t can go through any line and the fact that these insufficient energy problems are solved by attaching more faces.

I have yet to test whether a line coming from a transformer can exceed the transformer's energy rating, so I couldn't really suggest to you a reason. I know that I've seen from unhindered lines, EU/t sizes greater than the rating of the cable, but that's without a transformer inline.

 

[Zelfana]

When you step down in packet size with transformers you have to connect 4 sides to the lower voltage or it won't be able to work at full speed. Going the other way would work just fine with only one cable on both ends but it would be slower unless you had multiple sources of power outputting to the same line. Transforming speed is limited at the higher packet size per tick.

 

[netmc]

with transformers, you take 1 512eu packet and convert it into 4 128eu packets, which are turned in 16 32eu packets. What happens sometimes when you have lots of machines running off HV directly, is that you just don't have enough actual packets to satisfy the requests going to each machine.

so in 1 tick, if you have 20 MFSUs connected to your power line, you can only satisfy 20 machines. The next tick, you send out another 20 packets. and so on. I believe the energy net serves the closest machine first, so if you have something that can take that packet, it won't make it to machines further down the line.

when you have lots of small packets, they get distributed to all the machines, since each machine can only receive 1 packet each tick.

because of the infrequent method of receiving large 512eu packets, you my need to bump up the internal energy storage.

Industrial centrifuges only use 5eu/tick, so they don't need a lot of power, but do need it constantly.

This is at least how I understand the way eu works.

 

[snowmirage]

I think the netmc is right on point with the issue I'm having, I'm still testing but I think adding "more packets" will solve my problem.

I think the reason this didn't occur to me at first is that it must be more complicated than a simple 1 packet per machine, since as I said before I know I'm already way way over that number with the current machines that are working.

 

[Grydian2]

The HV transformer outputs 512eu/tic per side. You need 4 cables on 4 sides outputting into one main line to get 2048 out of the transformer.

 

[netmc]

I think the netmc is right on point with the issue I'm having, I'm still testing but I think adding "more packets" will solve my problem.

I think the reason this didn't occur to me at first is that it must be more complicated than a simple 1 packet per machine, since as I said before I know I'm already way way over that number with the current machines that are working.

It's 1 packet per machine, but only if there is space in the machine's internal energy buffer. So, if the machine doesn't have space for 512eu in its internal buffer, it won't accept a packet. (This is how you are able to power more than 20 machines at once.) If you only had 120 machines on the eu network, and are outputting 20 packets per tick, it would take 6 ticks to get packets to all the machines, provided none of them accepted more than 1 packet. A industrial centrifuge needs a 512eu packet every 100 ticks to keep running (about 5 seconds if at full tick rate).

I believe the more recent versions of GT have a tricorder, which will tell you everything you ever wanted to know about an IC2 machine, including the internal power storage. You can use this to see how much energy is actually making it to the machine. It should allow you to test things a bit more thoroughly.

 

[namiasdf]

Well, netmc's logic for MFSU's, only being able to output 512 EU/t out from one side, would apply to transformers as well. You guys are confusing the ability to send infinite packets, resulting in infinite EU/t down a single line with the maximum EU/t output of a block.

The EU/t rating of a block is literally it's EU/t limit or so how I understand it to be.

 

[immibis]

Last I checked you could send 4 512 EU/t packets out of one side of a transformer, you didn't need 4 sides. That was a while ago though.