Is the power shutoff bug still a thing with battery switchers/flippers?

[BLACKBERREST3]

My own battery switching guide for fixing the power shutoff bug is too confusing for me to understand 😭, maybe I can fix it some other way with all the new sensors we have like the wattage sensor. I don't even know if that bug was fixed at all tbh.

even ai is telling me that this bug still exists. They did fix a bug when constructing a new shutoff switch on an existing automation circuit it seems. not that I ever noticed that happening before from what I recall. "[Game Update] - 719533 - [Oxygen Not Included]"

let me know of any recent builds that address this issue. it seems like it should be an easy fix with the wattage sensor now.

[Zarquan]

Never trust Google's AI overview

[BLACKBERREST3]

31 minutes ago, Zarquan said:

Never trust Google's AI overview

Either way whether it's fixed or not, I've built a solution already. I'll post it in a sec. If it's fixed, I'll never need to spend the extra refined metal.

been a while since I worked on one of these.

I think this would work.

 

If I remember correctly, it was either pair of shutoffs that could get stuck in the off position even with an active green signal. I don't think it could ever fail in the on position. So we just need to send a green pulse when power draw is not detected. The only caviat to this is that we do not have a watt supplied sensor, only watts consumed. This means there is no real way to tell if you actually have incoming power generation on the wire going into the battery switcher. This doesn't matter too much since any power draw will disable the pulse timer AND even if the pulse does connect your consumers to your generators it will only do so when it is NOT consuming.

Logic Explained

Buffer = How long in a power outage (or just 0 watts consumed) until the pulse starts

Filter = How long every X seconds it should check before repeating the pulse

Double OR Gates = Input/Output Isolation so it doesn't create a feedback loop into the battery switcher

The 4 power shutoffs and NOT gate are part of the normal bat switcher setup

 

[BLACKBERREST3]

Condensed it.

[MinhPham]

I can confirm that the bug still exists ......

 

Also the current power distribution breaks the battery level synchonize across the network, they still works relatively but not as accurate as before

[KittenIsAGeek]

Huh. I've never had this particular problem.  Sometimes when the automation is built before the shutoff, it will be built "active" even when the automation is actually "inactive," but the first time the automation changes the problem goes away.

 

6 hours ago, MinhPham said:

Also the current power distribution breaks the battery level synchonize across the network, they still works relatively but not as accurate as before

On a grid with multiple generators and multiple batteries, I always have one battery set to run a generator to 100%.  It doesn't completely solve the problem, but the rest of the batteries usually end up pretty close to 100% by the time it shuts off.  

 

It would also be nice if the wattage sensor read the power going into charging batteries.  Of course, that would break a bunch of things such as fast-charging by using a pair of large transformers to send power down a regular wire...  but I digress.

[BLACKBERREST3]

You know what sucks is that I just built this, but I now i need to redesign this into my new Bat Switcher build. I'll make another post with my attempt at a power grid, but here is small snippet of what I've been working on.

For me, this seems to work better than just the one not gate.

 

18 hours ago, MinhPham said:

I can confirm that the bug still exists ......

 

Also the current power distribution breaks the battery level synchonize across the network, they still works relatively but not as accurate as before

In the build I'm working on, I think I've solved how to synchronize battery levels remotely without them having to be close together. Generator Priority is such a hassle when combining it with battery switchers and not only that, transformers have build priority that makes them charge not all at the same time, it's awful. I'm not sure if i ever did figure that out 7 years ago XD. I am bringing back the 50Kw relay though.

 

Just gotta combine the two rows of batteries with some automation. so close to a working grid 😮‍💨

In the one picture with the transformer -> battery -> transformer, I found that it's not completely accurate because build priority/flipping/magic idk. It gets it really close like within 1-4kj most of the time.

[Curry MAMA]

I thought we were talking about switched capacitors and charge pumps, but it seems I was wrong... (I walk away dejectedly)

(Here, we are using a green light that lasts only 0.5 seconds within a 60-second interval, and a red light that lasts only 0.5 seconds within a 60-second interval.)

[BLACKBERREST3]

5 hours ago, Curry MAMA said:

I thought we were talking about switched capacitors and charge pumps, but it seems I was wrong... (I walk away dejectedly)

(Here, we are using a green light that lasts only 0.5 seconds within a 60-second interval, and a red light that lasts only 0.5 seconds within a 60-second interval.)

I haven't even touched the bit writers/readers yet, is that a ribbon writer and an automation ribbon and what else is on that part? The problem we are talking about here is that on reload of a save (or rarely while playing) you get the shutoff bug where it can fail open. A lot of smarter people than I have used all sorts of powered devices to trip when the shutoff fails as that was the only means of detecting it. Past me helped test a build that used a zero power gas/liquid filter that when unpowered would trip. This would still work today, I was just hoping to make something smaller and cheaper in metals. any clock that pulses red to green AND green to red will fix it. Gamerhandbook made nice psa video about a while ago. It's nice to just have something that detects it instead of pulsing every x seconds. The build I did above is way too expensive in refined metal to use, but it's nice that it relies on automation only and not physics like doors opening below a battery and gasses/liquids merging in a filter. I was really hoping to use the wattage sensor for this. I might be able to make it cheaper yet, it's just that I had to isolate the outputs of the batteries with 2 OR gates. Maybe the ribbon readers/writers are key to isolating circuits. I haven't played with them enough to know yet.

edit: I think your build gets stuck when the battery needs a charge and it fails.

@Curry MAMA see this pic.

This is why I have 2 OR gates it my build to isolate the circuits and pulse each leg independently of the flipping logic.

Edited Monday at 04:17 PM by BLACKBERREST3

[BLACKBERREST3]

when I looked up switched capacitors and charge pumps, I found this

This looks pretty genius. I might use this later.

[BLACKBERREST3]

This also looks pretty cool, just gonna link it to look back here later.

So if I understand this correctly, I have to use a 1.92/2 ratio to prevent overload damage on the wires, 1.92s on and 2s off to prevent overload damage. The second rule is that if I try to use any power shutoff setup alongside a pulse power system, it will remember the oldest time as the system needs to be powered for it to start forgetting the overload accumulation. there's a lot to keep in mind, but I think I can make it work.

Edited Monday at 05:39 PM by BLACKBERREST3

[BLACKBERREST3]

This is an absolute game changer! This simplifies almost everything I was working on. and it's infinitely scalable! where was this all my life 😭

 

This is a direct upgrade to the 50kw relay per battery switcher i was using. I set the timer to .1s on .2 off for speed.

OH MY GOD! This is it! it was this simple the whole time! I'm gonna scream.

[MinhPham]

The problem of pulsed power grid is ... the disabled transformer loose their charge when disabled, as fast as a jumbo battery, you can't afford to spam them everywhere

The idea is :

•  Have a centralized pulsed transformers somewhere, run a power production line and consumer line all over the map.
• Connect the produciton line to the consumer line through the pulsed transformes.
• Connect your consumers to the consumer line through a power transformer. It acted as a battery as well so you  don't even need to have batteries all over the map.

Also the power transformer doesn't loose their charge if not disabled, minimum power losses

 

If you need more power just connect multiple transformer and use heavy watt wire on the endpoint

 

Edited Monday at 07:35 PM by MinhPham

[BLACKBERREST3]

What do you have as the time for the pulse? I think mine is a little too fast rn.

 

Edited Monday at 11:39 PM by BLACKBERREST3

[MinhPham]

I used the suggest timer on the article, 1.8s/2s

[BLACKBERREST3]

 

 

10 minutes ago, MinhPham said:

I used the suggest timer on the article, 1.8s/2s

I must have missed that, I thought he said the ratio was 1.9/2 but maybe that was just absolute worst case scenario

Edited Monday at 11:39 PM by BLACKBERREST3

[MinhPham]

It was me who missed the number, it was 1.9/2s on most of his article.

Because of that timer each large power transformer on the consumer side can deliver 1.8kW of power consistently, no battery needed. If the timer is 1.9/2s then it should be able to deliver 1.9kW

So far it does it work, only problem is i have no idea how much real power consumption is

Edited yesterday at 12:06 AM by MinhPham

[BLACKBERREST3]

I've got the fix for this bug figured out. Turns out past me already made the majority of it, I just had to relearn it. I've implemented that same principle as found in this article into my new bat switcher build that has better i/o. Here is the post I am working on.

https://forums.kleientertainment.com/forums/topic/171258-power-grid-battery-switcher-20-power-shutoff-bug-fix-50kw-capacitor-remote-battery-level-indicator-decentralized-pulsed-power-logistics/

 

Edited 23 hours ago by BLACKBERREST3