Train bandwidth Questions

[Nich]

I am going to have to agree the less trains you have on your network the less chance you have of a train needing to give right of way and stop. I have not done much with trains but from my observation the clear bottle neck is the loading/unloading station. There is no way around the 12 stack inserters per wagon limit. That being said you can amortize the startup cost of rails to multiple mines. Thus if you have 4 mines running on 1 set of rails it will be cheaper then 1 yellow belt (for just rails investment i think too lazy to do the math)

[Trepidati0n]

First of all...i'm not claiming i'm an expert. I've done enough math though (2 particular posts on reddit), coupled with actual game play, to feel pretty comfortable in my opinions though One thing I've learned about trains in this game, which i'm 100% confident in, is that a real base (e.g. not custom tailored to specific goal) will never hit perfect throughput.

I feel that acceleration is very important because of how rail signals operate. It is either "this is my block, GTFO" or "dammit, I gotta stop". So while a group of trains all heading in the same direction can have amazing throughput...throw "i'm going to cross now" and meh. To me, the only way to keep trains at full speed would be to use logic if a train is within ~170 block of junction (both cross and merge) then it stops anybody from entering that junction. That seems like a lot of wiring.

Thus i feel it okay to "waste throughput" at the expense of getting up to speed ASAP.

[Keeper]

I currently play with 1-8-2 trains, with enough input trains each station can keep 8 blue belts full. If I had less carriages then it would be 2 less belts for every carriage lost. If the belts went empty it's because there's no train unloading, need more input.

I disagree with the stack inserter thing I thought the energy requirement was too little. I have stack inserter 5 or 6 can unload a train in about 20 seconds. I'm sure it took just around 30 seconds to unload after just one or two stack researches

[Frightning]

I currently play with 1-8-2 trains, with enough input trains each station can keep 8 blue belts full. If I had less carriages then it would be 2 less belts for every carriage lost. If the belts went empty it's because there's no train unloading, need more input.

I disagree with the stack inserter thing I thought the energy requirement was too little. I have stack inserter 5 or 6 can unload a train in about 20 seconds. I'm sure it took just around 30 seconds to unload after just one or two stack researches

Stack Inserters are important to maximizing train throughput, but in many applications, you don't need a high throughput, and thus can afford to use less expensive and more energy efficient Inserters (Fast Inserters or maybe even regular Inserters).

[Keeper]

That is true. When I first looked at the stack inserter I expected to see massive energy usage as the more it picks the more weight it's moving right? Needs more powerful motors etc but I digress.

[Yoyobuae]

Stack Inserters are important to maximizing train throughput, but in many applications, you don't need a high throughput, and thus can afford to use less expensive and more energy efficient Inserters (Fast Inserters or maybe even regular Inserters).

With full research, fast inserters are the second most inefficient of inserters (Filter inserters are the worst). And I'm not talking about energy use while working, I'm talking about energy use per item moved.

Stack inserters are slightly worst than fast/filter inserters when without any upgrades, or when only up to Inserter capacity bonus 2 is researched.

[Frightning]

Stack Inserters are important to maximizing train throughput, but in many applications, you don't need a high throughput, and thus can afford to use less expensive and more energy efficient Inserters (Fast Inserters or maybe even regular Inserters).

With full research, fast inserters are the second most inefficient of inserters (Filter inserters are the worst). And I'm not talking about energy use while working, I'm talking about energy use per item moved.

Stack inserters are slightly worst than fast/filter inserters when without any upgrades, or when only up to Inserter capacity bonus 2 is researched.

Yes, but their idle energy drain is also lower than Stack inserters (about half as much), so if you're application isn't one that will run continuously (e.g. oil trains are a prime example). Then it can indeed be more energy efficient to use Fast inserters. (Also much cheaper on fixed costs, which imo, is a bigger issue).

[Yoyobuae]

Then it can indeed be more energy efficient to use Fast inserters. (Also much cheaper on fixed costs, which imo, is a bigger issue).

Better use normal inserter or long-handed inserters. They move 40% more items per energy used:
http://imgur.com/a/6Puym

If throughput is a concern then could double up the number of long-handed inserters.

[Frightning]

Then it can indeed be more energy efficient to use Fast inserters. (Also much cheaper on fixed costs, which imo, is a bigger issue).

Better use normal inserter or long-handed inserters. They move 40% more items per energy used:
http://imgur.com/a/6Puym

If throughput is a concern then could double up the number of long-handed inserters.

Idle cost matters in practice (drain), if your inserters are used only infrequently, then a significant portion of the energy expended will be due to idle costs rather than actually moving items. At the same time, you want/need inserters to be fast enough for the task at hand (ex: 1 Fast inserter is enough to keep up with a single fully speed moduled Assembly machine 3 unbarreling oil with the aid of stack size bonuses, but you would need 2 regular (or Long handed) Inserters to do so, which means double idle costs. If the system is generally inactive for long periods of time, such as between oil shipments, then those idle costs will more than offset the efficiency gains during full usage times). This is a good reason to use Burner inserters to fuel boilers (especially if you use Solid fuel or Rocket fuel to feed them) as Burner inserters have 0 drain, but ofc, are rather expensive to run (188kW, but at 100% efficiency rather than 50% efficiency of Boilers).

[Yoyobuae]

Power switch...

[ratchetfreak]

Then it can indeed be more energy efficient to use Fast inserters. (Also much cheaper on fixed costs, which imo, is a bigger issue).

Better use normal inserter or long-handed inserters. They move 40% more items per energy used:
http://imgur.com/a/6Puym

If throughput is a concern then could double up the number of long-handed inserters.

Idle cost matters in practice (drain), if your inserters are used only infrequently, then a significant portion of the energy expended will be due to idle costs rather than actually moving items. At the same time, you want/need inserters to be fast enough for the task at hand (ex: 1 Fast inserter is enough to keep up with a single fully speed moduled Assembly machine 3 unbarreling oil with the aid of stack size bonuses, but you would need 2 regular (or Long handed) Inserters to do so, which means double idle costs. If the system is generally inactive for long periods of time, such as between oil shipments, then those idle costs will more than offset the efficiency gains during full usage times). This is a good reason to use Burner inserters to fuel boilers (especially if you use Solid fuel or Rocket fuel to feed them) as Burner inserters have 0 drain, but ofc, are rather expensive to run (188kW, but at 100% efficiency rather than 50% efficiency of Boilers).

If you care about idle cost that much then you shouldn't be speed moduling the unbarreling assembler, putting a few in parallel with efficiency modules cuts down the power requirement much more than changing the inserter type would.

[Frightning]

Then it can indeed be more energy efficient to use Fast inserters. (Also much cheaper on fixed costs, which imo, is a bigger issue).

Better use normal inserter or long-handed inserters. They move 40% more items per energy used:
http://imgur.com/a/6Puym

If throughput is a concern then could double up the number of long-handed inserters.

Idle cost matters in practice (drain), if your inserters are used only infrequently, then a significant portion of the energy expended will be due to idle costs rather than actually moving items. At the same time, you want/need inserters to be fast enough for the task at hand (ex: 1 Fast inserter is enough to keep up with a single fully speed moduled Assembly machine 3 unbarreling oil with the aid of stack size bonuses, but you would need 2 regular (or Long handed) Inserters to do so, which means double idle costs. If the system is generally inactive for long periods of time, such as between oil shipments, then those idle costs will more than offset the efficiency gains during full usage times). This is a good reason to use Burner inserters to fuel boilers (especially if you use Solid fuel or Rocket fuel to feed them) as Burner inserters have 0 drain, but ofc, are rather expensive to run (188kW, but at 100% efficiency rather than 50% efficiency of Boilers).

If you care about idle cost that much then you shouldn't be speed moduling the unbarreling assembler, putting a few in parallel with efficiency modules cuts down the power requirement much more than changing the inserter type would.

Was merely pointing out that there are cases where each inserter type makes sense even from a power efficiency standpoint.

[reallyLost]

The reason for double-header trains is that the stations are much smaller (and easier to build), and it turns out that smaller builds make a huge difference when you're scaling up really high.

If you're only talking about outposts then sure, but I think that the concerns of central unloading stations are more important, and for those I don't agree with this logic. For small and medium stations, yes, double-headed makes for smaller stations. But for large stations I think double-headed stations are actually much bigger. Large stations often can't make effective use of the double-headed ability to back out the way you came in simply because there is always a queue of trains waiting to get in. So you always need a separate exit anyway which negates the advantage of double-headed. So then you've got trains that are 60% longer (LLCCCCLL vs LCCCC) making your stations 60% larger with no advantage. Then you still have to deal with the fact that longer trains also make congestion higher and traffic junctions slower.

[Frightning]

The reason for double-header trains is that the stations are much smaller (and easier to build), and it turns out that smaller builds make a huge difference when you're scaling up really high.

If you're only talking about outposts then sure, but I think that the concerns of central unloading stations are more important, and for those I don't agree with this logic. For small and medium stations, yes, double-headed makes for smaller stations. But for large stations I think double-headed stations are actually much bigger. Large stations often can't make effective use of the double-headed ability to back out the way you came in simply because there is always a queue of trains waiting to get in. So you always need a separate exit anyway which negates the advantage of double-headed. So then you've got trains that are 60% longer (LLCCCCLL vs LCCCC) making your stations 60% larger with no advantage. Then you still have to deal with the fact that longer trains also make congestion higher and traffic junctions slower.

...and their slower too. 1-way offers the highest throughput potential.

[iceman_1212]

It looks like train stop speed has gone up from 259.2km/hr. I had a 2-4-2 that was going 272km/hr. Not sure which of the many new factors is responsible but it was running on solid fuel and I have the first three braking force researches.