Anyone else not enjoying the new update? :(

[Planr]

Doing police-work on servers is completely difficult, since the logs don't help a lot. I can tell that by my own experience. It's impractical. Even the top servers do fail on taking action against planet miners because it's hard to keep one staff online at all time.

shame there isn't citizen's arrest in video games :p

 

[AtraUnam]

Edit 1: Apparently I was wrong about the cause of the power consumption increase (see Lancake's Post)

Edit 2: Upon testing a pre-update reactor I found it worked exactly how it used too suggesting that power consumption for power-supply beams hasn't changed at all.

 

[Lancake]

Good news everyone, power supply beams consuming twice as much power is indeed a bug and not a balance change. You may recall that beams were recently changed to draw power per tick instead of per second? Well previously power-supply beams consumed 240 power per second, after the change they consumed 240 power every tick, and since they tick every 0.5 seconds this effectively doubled their power consumption. All the devs need to do is change the power consumption per tick from 240 to 120 so I can practically guarantee it';l be fixed next release.

Power consumption for those were already per tick before and it still is now.
It provides 240 and it consumes 300. There's no bug with the power supply system now but it seems previous release and the ones before that could have suffered from a particular issue (although I don't know if that applied for logic fired reactors)

I quote from http://phab.starma.de/T384:

Power supply beam:

• the power consumed is incorrect when you're hitting a target (it's fine if you don't). It's in total the normal base consumption + the power supplied.

this and other issues related to that task were fixed.

 

[Panpiper]

Power consumption for those were already per tick before and it still is now.
It provides 240 and it consumes 300. There's no bug with the power supply system now but it seems previous release and the ones before that could have suffered from a particular issue (although I don't know if that applied for logic fired reactors).

I did a pretty rigorous test the other day to measure the output of power supply beams. I did not calculate exactly how much they were drawing, but I had 2890 power supply modules drawing 'roughly' 900,000 power and they transferred ~90 power per second only. It took 2890 of them precisely 766 seconds to fill up a 200 million battery on a test ship that had no power generation of it's own. The consumption figure of 300 power sounds right. The output? Not even close.

 

[Planr]

I did a pretty rigorous test the other day to measure the output of power supply beams. I did not calculate exactly how much they were drawing, but I had 2890 power supply modules drawing 'roughly' 900,000 power and they transferred ~90 power per second only. It took 2890 of them precisely 766 seconds to fill up a 200 million battery on a test ship that had no power generation of it's own. The consumption figure of 300 power sounds right. The output? Not even close.

Yeah this is true, I tested a similar thing myself with my own battery design and got the same result. I can share the blueprint with you if you'd like lancake.

 

[Tunk]

For the most part I've moved away from back injection.
Its just not robust, it is also fiddly, laggy and horribly inefficient compared to other power system designs.

 

[Lancake]

I did a pretty rigorous test the other day to measure the output of power supply beams. I did not calculate exactly how much they were drawing, but I had 2890 power supply modules drawing 'roughly' 900,000 power and they transferred ~90 power per second only. It took 2890 of them precisely 766 seconds to fill up a 200 million battery on a test ship that had no power generation of it's own. The consumption figure of 300 power sounds right. The output? Not even close.

Yeah this is true, I tested a similar thing myself with my own battery design and got the same result. I can share the blueprint with you if you'd like lancake.

Did both of you use logic to fire these power supply beams? I assume Planr at least did since that's about a battery/reactor design. There is/was a bug where logic beam duration was always fixed to 2.5 seconds when fired with logic. The attempt to fix it worked for normal beams but not for it support/slave ones, I believe a second commit was done to fix that but there were also a few more issues that were left unattended: http://phab.starma.de/T46

I would like to see a blueprint of that planr, PM me it if you don't want it to be publically shared.

I checked your numbers panpiper and they seem off. You say these transferred 90 power per second only, so how did you they fill up 200 000 000 power in 766 seconds which is 261 000 per second, or 130 000 per tick?
Your 2890 power supply should supply 138 720 per second and consume 173 400 per second.

 

[Panpiper]

I checked your numbers panpiper and they seem off. You say these transferred 90 power per second only, so how did you they fill up 200 000 000 power in 766 seconds which is 261 000 per second, or 130 000 per tick?
Your 2890 power supply should supply 138 720 per second and consume 173 400 per second.

I was using logic to control the power beam firings, a very simple clock mechanism. The beams were indeed firing for 2.5 seconds.

The calculation is 200 million power divided by 766 seconds divided by 2890 power supply modules equals 90 power per module transferred per second.

 

[Lancake]

I was using logic to control the power beam firings, a very simple clock mechanism.

The calculation is 200 million power divided by 766 seconds divided by 2890 power supply modules equals 90 power per module transferred per second.

I see, so both of you are using logic now. Is there a difference if you fire the power supply beam manually?

 

[Panpiper]

I see, so both of you are using logic now. Is there a difference if you fire the power supply beam manually?

I would have to rebuild a test facility to test that. It is not something I can do instantly. Maybe later today.

I cannot imagine much of a use for power supply beams that are 'not' controlled by logic.

I also suspect there may be an issue with shield supply beams as well, as the amount of shield they supply for the weight of the reactor is truly abysmal.

 

[Lancake]

I would have to rebuild a test facility to test that. It is not something I can do instantly. Maybe later today.

I cannot imagine much of a use for power supply beams that are 'not' controlled by logic.

I also suspect there may be an issue with shield supply beams as well, as the amount of shield they supply for the weight of the reactor is truly abysmal.

Hm, that would be odd. It has the supply value of a power supply, it uses 10 times less power though since you're basically transferring 100% of your shield to your target, power use is negligible.
I'm a little bit confused now since apparently support tools are broken yet I can't find anything broken on them, besides some sketchy sync issues.

 

[Panpiper]

I'll try to get more precise numbers on shield supply beams as well. The very rough numbers I can give you now is that a dedicated shield supply reactor massing 6700 blocks was capable of supplying around 2500 shield points per second. The reactor was generating far more power than it needed as at the time, I assumed it was simply generating shield out of power and that it would be as power hungry as a power transfer reactor. After much fiddling that reactor wound up with (again very roughly) 35000 shields, a similar regeneration rate and 'way' too much power.

I did not try to optimize a shield supply reactor after seeing how little it seemed to generate. It seemed FAR more efficient to add a regular power reactor and add 6700 extra shield regenerators (the mass of the test bed I had built) to the ship normally.

 

[lilheartless]

I'll try to get more precise numbers on shield supply beams as well. The very rough numbers I can give you now is that a dedicated shield supply reactor massing 6700 blocks was capable of supplying around 2500 shield points per second. The reactor was generating far more power than it needed as at the time, I assumed it was simply generating shield out of power and that it would be as power hungry as a power transfer reactor. After much fiddling that reactor wound up with (again very roughly) 35000 shields, a similar regeneration rate and 'way' too much power.

I did not try to optimize a shield supply reactor after seeing how little it seemed to generate. It seemed FAR more efficient to add a regular power reactor and add 6700 extra shield regenerators (the mass of the test bed I had built) to the ship normally.

I realize this post is a little old now but I was looking through some old stuff and came across this topic which was kinda just left hanging.

I decided to do some of my own testing this evening and came to the conclusion that this has either been fixed since the time of the original post or it was never actually broken.

I was getting aprx 80% returns on power generated from a docked power supply turret using logic one thing to keep in mind is that the figures given in game for the power supply beam both power transfer and power consumption are per tick and it "ticks" twice a second do if you want an extra 900k power regen per second you need aprx 1005000 power regen per second on the turret and 1675 power transfer modules x2 because there is a 2.5 second cooldown and activation time so you need two sets of power transfer modules hooked up to a single turret to have sustained transfer rates as well as not wasting the turrets power regen you could do a single cycle turret that relied on a larger capacitor rather than sustained regen but it wouldn't be a space efficient.