No power in battery switching curcuit

[whodunit]

Everytime the left battery is connected to the fridges, a random number of fridges don't get any power for a while, while the right one has no such issues. There's no wire damage anywhere and so I can't figure out why that happens. Could anyone please help me understand? has it always been like this?

[SharraShimada]

have you tried replacing the normal wires with conductive ones?

Lately transformers behave a bit strange if the power consumed is above wire max. And when the left battery comes in place, there is normal wiring in the game. When you switch to the right one, its only conductive wire from the power shutoff. Maybe thats your problem.

[whodunit]

14 minutes ago, SharraShimada said:

have you tried replacing the normal wires with conductive ones?

Lately transformers behave a bit strange if the power consumed is above wire max. And when the left battery comes in place, there is normal wiring in the game. When you switch to the right one, its only conductive wire from the power shutoff. Maybe thats your problem.

I just tried that but the same. Just so weird that it only happens with the left one. Is it because i'm using the left one to turn the shutoffs on/off? idk what else to think

[SharraShimada]

It was just a hunch. 

[nakomaru]

You must synchronize your automation circuit.

Add an OR gate to introduce a 1/10 second delay to your positive line to synchronize with the 1/10th delay of your NOT line.

And have your battery switch at around 20 percent.

And your current power backbone can only supply 1kW. Of course your 1560W of consumers will brown out.

[psusi]

5 hours ago, nakomaru said:

You must synchronize your automation circuit.

Add an OR gate to introduce a 1/10 second delay to your positive line to synchronize with the 1/10th delay of your NOT line.

No.  I don't know if a NOT gate doesn't add any delay, or if it just gets rounded down, but I've never had a battery flipper have any trouble using only one NOT gate.

5 hours ago, nakomaru said:

And your current power backbone can only supply 1kW. Of course your 1560W of consumers will brown out.

Yes, the transformers should not be there and because only 1 is hooked up, will cause rolling brown outs, but the whole circuit should brown out at the same time when the batter dies, not only half of it.  That's really fubar.

[nakomaru]

4 hours ago, psusi said:

the whole circuit should brown out at the same time when the batter dies, not only half of it.  That's really fubar.

If there's 800J in the battery, what do you think should happen? Should it supply 0 machines requesting 120J, or it should supply 6 of them? Well, we know the answer Klei went with.

[whodunit]

9 hours ago, nakomaru said:

If there's 800J in the battery, what do you think should happen? Should it supply 0 machines requesting 120J, or it should supply 6 of them? Well, we know the answer Klei went with.

That's a fair point. But in my original post, 1st picture, the battery connected has 15k J and yet it's only supplying only 6 fridges. That same battery, only a couple seconds later and now being at 12k J is happily supplying power to all. And this behavior is entirely absent for the second one. It had ~4k J and had no brownouts.

It's this inconsistency that has me puzzled. I corrected it by using a big transformer and now there are no brownouts. But now I'm wondering when the shutoffs disconnect the 1st battery from the entire circuit, why are there brownouts at all. Battery has no limits when discharging and the 2k wire can handle the load, shutoffs ensure that the circuit "sees" no 1k wire so what goes wrong that gets corrected in a couple seconds?

[nakomaru]

5 hours ago, whodunit said:

But now I'm wondering when the shutoffs disconnect the 1st battery from the entire circuit, why are there brownouts at all.

Because your automation doesn't merely disconnect one battery from your circuit, it disconnects both. See above for the solution.

And on the other flip, which doesn't have that problem, is because it's connecting both batteries and your main line directly to your consumers.

[psusi]

19 hours ago, nakomaru said:

If there's 800J in the battery, what do you think should happen? Should it supply 0 machines requesting 120J, or it should supply 6 of them? Well, we know the answer Klei went with.

Sure, but that would only happen for the last second before the battery dies, and in his pictures, the battery is nowhere close to dead.

5 hours ago, nakomaru said:

Because your automation doesn't merely disconnect one battery from your circuit, it disconnects both. See above for the solution.

And on the other flip, which doesn't have that problem, is because it's connecting both batteries and your main line directly to your consumers.

No it doesn't.  The pictures clearly show one way the left battery is connected to the loads, and not the feeder, and in the other the right battery is.

10 hours ago, whodunit said:

It's this inconsistency that has me puzzled. I corrected it by using a big transformer and now there are no brownouts.

I assume you are trying to do this:

The idea is that you shouldn't have that transformer there at all.  The transformers should be on the generator side and the trunk line just runs between all generators and all flipper substations without any transformers to isolate the substation from the trunk line ( that's what the shutoffs are for ).

[Gamers Handbook]

@whodunit You have 13 fridges for a total of 1560W, but are only supplying them with 1000W from a small transformer.  Even with switching batteries you can't keep all the fridges running at the same time, because the switching batteries aren't supplied with enough power for the fridges running 24/7.  Basically you can't keep both batteries charged and that's why you're seeing the power go out.  If you change your setup so that the batteries stay charged (like by hooking that second transformer in and supplying 2kw to the batteries), everything should work how you want it to.

@nakomaru  They patched the timing months ago and an OR gate is no longer needed for switching/flipper battery setups.  One NOT gate is perfectly fine for a switching/flipper battery.

@psusi Hey I know that design!  My guess is he's trying to do a different older method that was developed before I posted my transformer flipper though, something like @Mullematsch shows on page 1.

[psusi]

2 minutes ago, Gamers Handbook said:

@whodunit You have 13 fridges for a total of 1560W, but are only supplying them with 1000W from a small transformer.  Even with switching batteries you can't keep all the fridges running at the same time, because the switching batteries aren't supplied with enough power for the fridges running 24/7.  Basically you can't keep both batteries charged and that's why you're seeing the power go out.  If you change your setup so that the batteries stay charged (like by hooking that second transformer in and supplying 2kw to the batteries), everything should work how you want it to.

Again, that would not run only half of the loads; it would run them all right up until the battery died, then all of the loads would die ( within 1 second of each other ).  So they should be all going on and all going off at more or less the same time, rather than half of them being on all the time.

4 minutes ago, Gamers Handbook said:

@nakomaru  They patched the timing months ago and an OR gate is no longer needed for switching/flipper battery setups.  One NOT gate is perfectly fine for a switching/flipper battery.

Funny.  I recently returned to the game after taking almost a year off and before that I had used a flipper without needing an OR gate and it seemed to work fine.

5 minutes ago, Gamers Handbook said:

@psusi Hey I know that design!  My guess is he's trying to do a different older method that was developed before I posted my transformer flipper though, something like @Mullematsch shows on page 1.

If he's just trying to use a flipper instead of a transformer flipper, then still there is no point in having the transformer there, unless you also have other loads in between the transformer and the flipper.  And that would only be so that you can have peak surge > 2kw while averaging < 2kw on everything after the transformer, and you don't want to bother just running another line all the way back to another transformer on the heavy watt backbone.

[Gamers Handbook]

I believe they used the transformer so they could use a smart battery and have their generators shut off when power wasn't needed.  Without a smart battery, the generators would run endlessly.  Without a transformer the smart battery would always remain full and therefore be useless.

@whodunit  A picture of your automation overlay may be helpful as well as the settings on your smart battery

[psusi]

9 minutes ago, Gamers Handbook said:

I believe they used the transformer so they could use a smart battery and have their generators shut off when power wasn't needed.  Without a smart battery, the generators would run endlessly.  Without a transformer the smart battery would always remain full and therefore be useless.

Sure, but then the battery flipper gains you nothing; you may as well just take them out and run all of the loads directly from the transformer.

[Gamers Handbook]

28 minutes ago, psusi said:

Sure, but then the battery flipper gains you nothing; you may as well just take them out and run all of the loads directly from the transformer.

I was operating under the assumption this is one small segment that will be expanded into a full power grid.  In that case, people use it as a way to get around heavy watt wire.

[psusi]

Just now, Gamers Handbook said:

I was operating under the assumption this is one small segment that will be expanded into a full power grid.  In that case, people use it as a way to get around heavy watt wire.

I think you are saying the same thing I said here:

48 minutes ago, psusi said:

unless you also have other loads in between the transformer and the flipper.  And that would only be so that you can have peak surge > 2kw while averaging < 2kw on everything after the transformer, and you don't want to bother just running another line all the way back to another transformer on the heavy watt backbone.

 

[whodunit]

10 hours ago, nakomaru said:

Because your automation doesn't merely disconnect one battery from your circuit, it disconnects both. See above for the solution.

And on the other flip, which doesn't have that problem, is because it's connecting both batteries and your main line directly to your consumers.

I tried with the OR gate and synchronize it but it's the same situation. Even if I'm synchronizing it wrong, the brown-out should only be for 1/10 of a second but it lasts much longer.

 

4 hours ago, psusi said:

I assume you are trying to do this:

That post was what made me try this but I was just figuring out the basic system first which @martosss has in this post

before moving on to the more complex one.

 

@Gamers Handbook Battery is at 80-60. It also tried it with 40-20 and 90-85 but same issues. Sorry about the long automation wires, just this way for me to get a clearer picture.

[psusi]

Wait... does either battery ever hit zero?  I have noticed that if a battery is set to zero, then when it gets there, the loads decide they have no power, and continue to show they have no power for a few seconds after the switches connect the full battery to them.  Since your system is not getting enough power I would think that sometimes one battery does hit empty, some of the fridges go into "out of power", and when the other battery is finally fully charged and switches, then sometimes it takes the fridges a while to notice they have power again.

 

[nakomaru]

2 hours ago, whodunit said:

the brown-out should only be for 1/10 of a second but it lasts much longer

Some machines can't cycle off and on instantly, like the aquatuner. I'm guessing the fridge is one of them.

If you are still not satisfied with the explanation please post a save.

[whodunit]

6 hours ago, psusi said:

 

Yeah the 2nd battery does hits zero. Your explanation fits perfectly but i don't know if it's officially how some building like fridges behave.

 

5 hours ago, nakomaru said:

Some machines can't cycle off and on instantly, like the aquatuner. I'm guessing the fridge is one of them.

If you are still not satisfied with the explanation please post a save.

Here's the save. I think the machines being unable to cycle on/off instantly fits and makes sense. It's just it sounds a little hard to believe and I hadn't thought it was possible.

powertest.sav

[psusi]

11 hours ago, whodunit said:

Yeah the 2nd battery does hits zero. Your explanation fits perfectly but i don't know if it's officially how some building like fridges behave

I don't think the building matters; I think any machine when it wants power and sees the battery is dead will throw up the "no power" flag and only checks to see if power has become available once every several seconds, so when the shutoff flips it takes them a while to notice.

[nakomaru]

Here is the delay introduced by unsynchronized NOT gates. You may observe that for one frame there are two red signals, and another with two green signals.

However, GH is right: the behavior has indeed been changed for a NOT gate interacting directly with primary sources such as a smart battery.

Perfect synchronization in this case. Interesting fix for imperfect engineering.

Anyway, the problem is indeed merely the delayed responsiveness of certain buildings after losing power. It does depends on the particular building, and it also depends on whether a dead battery is involved.

Dead Battery:

Spoiler

Disconnected charged battery:

Spoiler

[Gamers Handbook]

That's the point I was [in-eloquently] trying to make earlier, it's because of the drained battery.

@nakomaru NOT gates can still desync every other one when they're chained.  Whatever they changed in that update, it made it so one NOT gate works with switching batteries, but timing can still be an issue if you are trying to do more complex stuff.  I assume for your small battery example you were sending the automation pulse with another gate?

https://imgur.com/a/ayiE3rB

 

[nakomaru]

1 hour ago, Gamers Handbook said:

NOT gates can still desync every other one when they're chained.  Whatever they changed in that update, it made it so one NOT gate works with switching batteries, but timing can still be an issue if you are trying to do more complex stuff.  I assume for your small battery example you were sending the automation pulse with another gate?

Yes, another gate. This was the point of the demonstration of differing behavior / confirmation of your claim, and what I meant by the unique behavior of "primary sources" (as opposed to gates) such as batteries and in a separate case, a clock signal.

Note: previous builds which utilize the asynchronous nature of gates still work even with primary sources (i.e. it is still asynchronous, but there are sub-frames in automation that are not delivered to the user.)

[whodunit]

@nakomaru @Gamers Handbook @psusi Thank you for all the answers and helping me out! It's a bit comical that fridges don't realize instantly that they have power now but it's easily avoidable and explains everything.