Known power bug?

[Dupe_626]

Cycle 3900, the standard power setup, lots of generators connected to one heavy watt conducive wire and then, transformers. All the smart batteries controlling the generators have the same settings (20-95). For a split second, the power on the heavy watt conducive wire goes to zero and I don't understand why. The only somewhat wonky thing that I can think of is that I have some volcano taming turbines, but they are separated from the 'spine' by transformers and automated shutoffs.

Another thing: if I put 12 mercury lamps behind a transformer, when the power on the 'spine' goes to zero, the conducive wire connecting the lamps briefly goes red with a value of +10000 w (?!?) although it is, like I said, behind a transformer and the max power draw should be 720 w. Is this kind of behavior to be expected the more you build in the game or am I obviously doing something wrong?

[Knurek]

It's rather hard to say without taking a look at your save or some screenshots. Do you use any mods related to the power grid?

[Dupe_626]

Weird things are happening. The sweepers at the bottom were connected to the same CW. After disconnection, the CW on the sweepers seems to remember the connection and sometimes, just for a 'tick', draws 720 w, other times it draws thousands w. After deleting the disconnected CW, the problem persists. After deleting the transformer and the CW, the spine still fluctuates to 0 w. I installed some automated notifiers to show what's going on.

The game is original with all the DLCs and three mods installed but disabled. The map is a standard frost dlc map, not converted, not imported...

[Knurek]

Definitively a bug related to mercury lamp (this is the first time I'm using it). Good catch!

I've tried to recreate your setup on a smaller scale on my current save. I've put a large transformer between the spine and 6 mercury lamps plus a liquid pump. I've set the wattage sensor and notifier to pause the game if wattage exceeds 1kW. Potential load is 600W, but the sensor reads 1381,33W. The current on the spine also went down to 0W, just like in your case.

You should add this bug to the bug tracker, or I can do it if you'd like, just let me know.

[Dupe_626]

I ended up deleting the spine (maybe an exaggerated measure) and the lamps, the pump, the sweepers, the transformer, the circuit and then I rebuilt it all, just like it was and so far, everything seems to be back to normal, but... now that you mentioned the pump, I checked and realized that I forgot to rebuild the pump. So good catch to you! This could indeed be related to the pump. But yeah... they created an amazing game and programming is not an easy thing to do.

Anyway, thank you for replying and taking the time to test the problem. I'm not so familiar with this forum, so please add this to the 'bug tracker' if you can. Just let me know where I can follow it.

[Dupe_626]

Nevermind... Something is very wrong. In short, everything seems to be running fine until I decide to build some mercury lamps and power them. This time I tried it on a new separate grid (without the pump, only six lamps) and the fluctuation to zero still happened and more than that, it happened on both grids(!) at the same time, as if they're connected by an invisible wire. Then the game crashed. After resuming the game from a save file from before(!) the lamp experiment, the spine still showed the same behavior randomly fluctuating to zero w. After a restart of my system, nothing changed.

The save file is 31 mb so I cannot upload it here. My best guess is that I'm running low on memory and that's why weird things are starting to happen. 16 gb might not be quite enough for a big map and a bunch of planetoids. I ordered a 32 gb kit, it'll arrive next week and we'll see how it goes...

[gigamoi]

33 minutes ago, Dupe_626 said:

The save file is 31 mb so I cannot upload it here. My best guess is that I'm running low on memory and that's why weird things are starting to happen. 16 gb might not be quite enough for a big map and a bunch of planetoids. I ordered a 32 gb kit, it'll arrive next week and we'll see how it goes...

You may check if you're guessing correctly using your system's resource manager to see how full is your ram and if your OS has to put stuff on the disk to compensate some lack of RAM. That being said, even though 16Go should be more than enough to cover most light use cases nowadays, it's a bit tiny for anything above that and I would not be surprised if your decision to upgrade your RAM pays of in some way (even I think it is unlikely to fix this problem of yours).

[Knurek]

  Here's the bug report:

 

4 hours ago, Dupe_626 said:

My best guess is that I'm running low on memory

Your system memory is fine. Same behavior occurs on a tiny build on cycle 6. Check the video in the bug report.

[Dupe_626]

My memory usage for this map is above 75%. If I ctrl+esc the game and open five tabs in Edge, the game crashes. Anyway, good to know that wild farming of bonbon trees is the route to go for now.

[Dupe_626]

I did some more testing because I thought that the random fluctuation to zero, even after deleting the afflicted circuit, might be a completely unrelated issue, but it's not. What happens is that after deleting the red circuit, the bug will 'jump' to a random transformer and affect the corresponding conductive wire and that's how the random fluctuation on the spine happens. The only way I managed to fix it was by disconnecting all the transformers from the spine and then reconnecting. A more elegant solution might be a general power switch to shut off and restart all power across the base.

[Dupe_626]

Well... I spoke too soon. The fluctuation is still there. I tried the simplest setup on a map I use for testing and as you can see, the battery has juice, the lamp is shining, but there is zero load on the wire. So just to clarify one thing: is this fluctuation just how power is supposed to work in this game or am I the only one having this problem?

[gigamoi]

I don't think the low values are a bug per say. I didn't peek into the game code but I think power may work by transferring joules from one buffer (one per generator, battery and consumer) to an other very fast. The low values we see on wires would occur when consumers are running on their internal buffer but are not receiving power from producers' buffers yet (in which case, you just need a buffer gate set to a tiny time to get the reading you would expect from the sensor).

Large values could occur if some consumer with a buffer far larger than their consumption rate gets that buffer filed all at once by some supplier with a sufficiently large and full buffer. The consumer's rate limit would eventually work as a cap in in the long run but wouldn't affect the initial supply phase. There is a known delay between circuit overload and wires being damaged. This delay may be there to account for this powering-up phase.

[Knurek]

11 hours ago, Dupe_626 said:

is this fluctuation just how power is supposed to work in this game or am I the only one having this problem?

No, it's not how it's supposed to work. I did some more testing, and it looks like this bug doesn't always occur, but I'm still pretty sure the cause is the mercury lamp, as it seems that it only happens sometimes after building them. The issue persists even after destroying the lamps, saving and reloading the game. Maybe destroying the entire circuit helps, as you mentioned before, but I'm going to leave this bug for the devs to figure out.

As to if you're the only one having this - no, the game is clearly bugged. Maybe it's just that not a lot of people use this lamp in a circuit like yours, or they were just lucky and the bug didn't occur for them, or they didn't play long enough for the game to slow down enough for the wattage sensor (if they used one) to be triggered by that momentary 0W drop. I for example never touched the mercury lamp, until you started this thread, because I've got plenty of power on Ceres to use the sun lamps instead.

Here I've had this bug, deconstructed mercury lamps, saved and reloaded the game, and I still momentarily get 0W on the main line (the sensor doesn't even detect this change) with a separated wire getting wireless power(?):

[Dupe_626]

@gigamoi - Very interesting theory, but the fact that the issue is so inconsistent proves that it's probably a bug. On my old test map I used a stopwatch to see if the power drop is consistent and it is... kinda... if a random variation anywhere between 4 and 28 seconds can be considered consistent.

I started a fresh new map and I built the same simple setup with the sun lamp. No power drop! Then I deleted the sun lamp and I installed six mercury lamps. Still nothing! Everything works just fine!

So there is probably a set of conditions that trigger the chaotic behavior. Although it's not a game breaking bug, it is a bit of a bummer, to be honest, because I like to have everything working properly and power is such a crucial and cool part of this game. I mean a critter getting stuck somewhere - no big deal, but this...

@Knurek - Well... At least now I know I'm not completely crazy. Like you said, I'll let the devs figure it out, if they ever do. For now I'll try to take and enjoy the game for what it is...

_

Thanks!

[Knurek]

@Dupe_626 You could try replacing all automated power shutoffs in your save (and manual ones too if you have any) with automated transformers. It seems to fix the issue temporarily, although you might need to upgrade some wires to withstand transformer power draw.

[Knurek]

I've loaded up my old, pre-frosty (so no mercury lamps there ever) colony (9000+ cycle, 45 dupes, 8 FPS slideshow), I've put a power shutoff (with a notifier to pause the game on power cut) between ceiling lamps on a conductive wire connected to a large transformer, and I've put a wattage sensor on the heavy wire. Even with no mercury lamps I'm sometimes getting momentary 0W on the heavy wire the moment the shutoff cuts the power, but the wattage sensor fails to read it:

At the same moment a piece of wire at the shutoff reports -1W:

A nearby conductive wire reports 14710W:

And some other wires connected to transformers report 0W:

So the mercury lamp is not involved in this bug. It's all because of the power shutoffs.

[JRup]

Power shutoffs will mess with the grid. It's a known feature™.

If you need automation with the Wattage Sensor to be responsive & accurate you will have to account for this event. Filter and/or Buffer gates will help but the more power shutoffs used on the map the more it becomes a "random event generator" that can overcome even the best automation setup. I just use it in tandem with a hammer to add to the entertainment value it provides

This can help learn more about the wonder power shutoffs are:

 

[Dupe_626]

@JRup I don't care about the wattage sensor. I used it just to help me catch the bug and even that it couldn't do reliably. I had to chase it with the spacebar.

@Knurek The same exact thing happens on my FP map. I destroyed all my shutoffs and restarted the game, but that didn't do anything. The bug remains. And as I mentioned before, deleting the overloaded cw doesn't fix it, the bug will just move to another cw and the fluctuation will happen anyway.

[Knurek]

@Dupe_626 I'm out of ideas then, sorry.

@JRup It's hardly "known" if the only mention of this is in the middle of a wall of text in a totally unrelated thread, unless I missed something. Was this even reported as a bug before? I could not find it in the bug tracker. Filters and buffers are no solution if it can still get broken.

[Dupe_626]

@Knurek Thanks, man! I appreciate you trying to help. It doesn't affect the gameplay so much, but it's a nasty, ugly bug. It's like driving a fancy car with a broken engine. It works, but it doesn't feel right. I really hope someone from the dev team reads these forums and that it will be fixed soon. One thing I forgot to ask... Could there be a mod out there which we're not aware of, that alters/improves the way power works?

[nome]

5 hours ago, Dupe_626 said:

I really hope someone from the dev team reads these forums and that it will be fixed soon

They do, and it's being worked on.

[Dupe_626]

9 minutes ago, nome said:

They do, and it's being worked on.

Holy gassy moo! I didn't expect that! But I would imagine it's a question of pride for a programmer to fix such an insidious bug. Thank you very much and as someone with little experience in programming (python), I would be very curious to know if it's a tricky bug to fix. If that's not too much to ask, of course.

[nome]

41 minutes ago, Dupe_626 said:

Holy gassy moo! I didn't expect that! But I would imagine it's a question of pride for a programmer to fix such an insidious bug. Thank you very much and as someone with little experience in programming (python), I would be very curious to know if it's a tricky bug to fix. If that's not too much to ask, of course.

I'm not on the oxygen not included team so I can't speak to the bug myself, but it looks more than trivial at first glance and there are a lot of other issues competing with it for attention. As such there's no guarantees which release it'll get fixed in. It's on the radar though!

[myxal]

21 hours ago, Dupe_626 said:

I had to chase it with the spacebar.

FWIW, in debug mode you can use Alt + "=" to advance by one tick. Huge help in understanding/observing the mechanics tick-by-tick.

[Dupe_626]

@myxal Thanks! Very useful tip!