A new take on Battery Switching / Regular Wire Power Grids

[psusi]

So I finally built a second transformer flipper exactly like the first and for some reason, the darn things won't stay synched.  I even forced some generators on to top up all batteries in both to get then synched, and a minute or two later they are out of sync again and the new one won't run its generators, leaving all the work to the old one.

 

[psusi]

Also I have no idea why it apparently fixed itself just in time for me to take those screen shots, but it is back now: the transformer on the right of both setups won't fully charge for some reason.. it only goes up to 800 J instead of 4000.

 

Weird... I decided to try giving them another equalization charge and this time it seems to be working... for now.

[Bloxxed]

@Gamers HandbookCan I have a link to the save file?

[Gamers Handbook]

@psusi  Yes, the power shutoff bug is very annoying when it hits.  Since it only activates when you save while the automation is toggling, allowing wider gaps in your smart batteries so they aren't cycling as often helps a little bit.  For automated fixes, I proposed this solution:

There was also a couple other solutions proposed in another thread but I can't find the thread.  (tagging @crypticorb here too)  I know in that thread I proposed using a liquid shutoff and sensor to monitor when the power cuts out and then flip the signal to reset the power shutoffs, and someone else proposed using a powered smart container to do the same thing.  They both worked, the smart container was a bit closer to the desired instant reset, and of course each is much more responsive than clocks once a day.  So there's options which can help alleviate or undo the bug, but ofcourse nothing is as good as if Klei actually just fixes the bug (no news there as far as I know).  It may be worth implementing one of the faster solutions if you're constantly finding yourself re-syncing your batteries or if they control something extremely power sensitive.  Personally, it's rare enough for me to get hit with it that I don't use any protections from it and just slide around the battery percentage bars and resync the batteries if it happens.

As for syncing, they will never be in perfect sync.  They'll be close enough though.  You can also just use relative battery percentage values between different transformer flippers though.  Like you pointed out, they're still charging and discharging at the same rates.  So you just set your generator "on" percentages based off of what feels like a good balance of your fuels.  I prefer to sync them so I know specifically when things are going on and off, but you don't have to sync if you don't want to.

@crypticorb I haven't submitted any additional bug reports.  This is my original submission: 

 

Oh ya, let's see how you slimmed the layout up.  That specific layout was chosen simply because it was the easiest to understand when trying to show it to someone.  I fully expect people to rearrange things to fit into their builds.

@ForTheQueen and @psusi  The bug is triggered when the signal flips from red to green during a save, and then that save is loaded.  If merely a delay was needed, then I wouldn't have been able to reproduce the bug with NOT gates.  @crypticorb is right.  See my test here: 

 

@ForTheQueen  I had experimented with timed transformer flippers (instead of demand/charge based like it is now), but it breaks the syncing between the two batteries in the transformer flipper too much.  It's just not consistent.

@psusi Each battery charges equally, so you're putting twice as much power into the side with 2 batteries.  I imagine this has something to do with your issues trying to get both sides to sync.

@Bloxxed I haven't ever made save files, so I'm not sure how to do that.  Either way, unfortunately the save file I made the video in got corrupted like a day after I made the video.  Since then my next 2 sandboxes have gotten corrupted as well as my original base I ran the design in.  I'm not sure what's up with the corrupting games; only my 2 most recent bases haven't gotten corrupted.  However, it's been working great in all my bases and sandbox maps.  If you let me know what you're specifically having trouble with, maybe I could post some pictures that would help?  I also made a long-form livestream video (2 parts since my internet died in the middle) showing the process of building it, perhaps that would be helpful?: 

 

 

[psusi]

58 minutes ago, Gamers Handbook said:

Each battery charges equally, so you're putting twice as much power into the side with 2 batteries. 

Technically it isn't twice as much since there are a number of batteries in the grid that are all dividing up the charge, but yes, it is more on that side.  Why should that matter?  Since the initial trouble they have mostly been staying in sync.  At least I haven't worried about it.  I think I had the shutoff bug strike once the other day and they got slightly out of sync and I haven't bothered giving them an equalization charge again.

[KittenIsAGeek]

5 hours ago, psusi said:

Technically it isn't twice as much since there are a number of batteries in the grid that are all dividing up the charge, but yes, it is more on that side.  Why should that matter?  Since the initial trouble they have mostly been staying in sync.  At least I haven't worried about it.  I think I had the shutoff bug strike once the other day and they got slightly out of sync and I haven't bothered giving them an equalization charge again.

Think of batteries like cups and electricity flow like a running faucet.   Now assume that the faucet is set up so that all cups fill equally.  When you're charging one side, "water" is pouring into one cup.  When you're charging the other, "water" is flowing into two cups.  You're flipping state based on the charge of the 'single' battery. 

Lets add another assumption: When you're discharging, you're drawing power at a constant rate.  Since we're also assuming that the batteries are charged at a constant rate, we get into this situation:  "Water" is poured out of one cup at rate X. It gets down to 50% (or whatever) and flips to the other battery to charge.  Now "Water" is pouring into that one cup at rate Y, meanwhile water is pouring out of the two cups at rate X.  The single battery reaches its "green" state and flips the state back to itself.  This lets the two batteries start filling at rate Y while the one is losing water at rate X.  Assuming that the rates remain constant ant that Y > X, your batteries will probably stay synchronized. (edit) Even with more batteries on one side, provided the draw/charge is constant, they'll stay synchronized.

Now, lets assume that your discharge current is not constant.  This is the case for most bases.  You have things powering on and off that change your load.  Since "X" is now changing while "Y" remains constant, you can end up with uneven discharge periods.  Since you're only checking charge on one side, this means that it is possible for the other side to slowly lose charge if power draw is greater while the single battery is charging.  This will become apparent when the temporary power draw maximum exceeds your power production capabilities. 

Lets say your average base consumes 720 watts/s of power and your generators produce 1kw/s.  However, your temporary draw can potentially hit 1.2 kw/s.  Since your running average is 720 watts/s, your generators can keep up provided there's enough batteries to hold the charge for those times when you are using 1.2kw/s of power.  However, if you're using a battery flipping system, and the charge level is checked on only one side, then it is possible to "short charge" one side of the system.  Your batteries will slowly fall out of sync.

 

This is a similar issue to lithium battery packs and one reason why its recommended to use a balance charger system on them.  Unlike other battery chemistries, electricity doesn't flow if the cell is charged.  Under ideal circumstances, power is drawn equally from each cell when used and delivered equally to each cell when charged.  However, in the real world, there's pesky things like wire resistance and the internal resistance of the cells themselves, as well as minor variations with individual cell capacities that results in a very slight inequality of charge/discharge.  Over time it adds up, and without a balance charger your battery pack will appear to slowly lose capacity.  

[psusi]

22 hours ago, KittenIsAGeek said:

Think of batteries like cups and electricity flow like a running faucet.   Now assume that the faucet is set up so that all cups fill equally.  When you're charging one side, "water" is pouring into one cup.  When you're charging the other, "water" is flowing into two cups.  You're flipping state based on the charge of the 'single' battery. 

Yes, but my point was that these three batteries aren't the only ones in the system.  If there are 8 total batteries then each gets 1/8th of the net charge.  Then you flip two batteries off and one on and now each is getting 1/7th of the net charge.  1/7th isn't double 1/8th.

22 hours ago, KittenIsAGeek said:

Lets add another assumption: When you're discharging, you're drawing power at a constant rate.  Since we're also assuming that the batteries are charged at a constant rate, we get into this situation: 

But they aren't charged at a constant rate; sub stations are absorbing various amounts of charge over time, leaving less for the transformer flipper.  Generators turn on and off changing the charge rate.

22 hours ago, KittenIsAGeek said:

This lets the two batteries start filling at rate Y while the one is losing water at rate X.  Assuming that the rates remain constant ant that Y > X, your batteries will probably stay synchronized. (edit) Even with more batteries on one side, provided the draw/charge is constant, they'll stay synchronized.

You seem to be talking about synchronizing the two sets of batteries in the one transformer flipper.  The left and right sides won't be and don't need to be synchronized.  You just need the left side of the two different transformer flippers to stay synchronized.  Since they both accept 1/n of the net charge when in charge mode, and they both output 2kw in discharge mode, they will.  Unless the timing of the substation loads causes n to tend to have a different value at different times ( which it will ), and those times happen to line up such that n tends to be higher at the same time when one transformer flipper is charging the left batteries and lower at the times when the other one is ( possible, but unlikely on average ).  But that can also happen with 4 battery flippers.  Unless.. I suppose if you fully charge all of them to get them synchronized, then they would all start out with the same side in charge mode and the other side in discharge mode, and always flip at the same time, then they would always stay in sync.  Hrm..

22 hours ago, KittenIsAGeek said:

However, in the real world, there's pesky things like wire resistance and the internal resistance of the cells themselves, as well as minor variations with individual cell capacities that results in a very slight inequality of charge/discharge.  Over time it adds up, and without a balance charger your battery pack will appear to slowly lose capacity.  

The battery pack we produce for our product at work uses 4 lipos in series without a balancer and doesn't seem to suffer from this.  We do have to be careful when we assemble the pack to make sure that all of the cells start out with close to the same capacity though.  Otherwise, one weak one drags the whole pack down.  Usually when we have old low capacity packs, it appears to be due to one cell having much reduced capacity.  It still charges up to full voltage, but runs down to the low voltage cutoff well before the other cells.  It doesn't seem like a balancer would help that.

 

[KittenIsAGeek]

2 hours ago, psusi said:

The battery pack we produce for our product at work uses 4 lipos in series without a balancer and doesn't seem to suffer from this.  We do have to be careful when we assemble the pack to make sure that all of the cells start out with close to the same capacity though.  Otherwise, one weak one drags the whole pack down.  Usually when we have old low capacity packs, it appears to be due to one cell having much reduced capacity.  It still charges up to full voltage, but runs down to the low voltage cutoff well before the other cells.  It doesn't seem like a balancer would help that.

 

First, I was responding to the post about synchronization.  I agree with all your rebuttals.  I was explaining why the specific system you showed was getting out of sync.

Now, about the lipo packs you use: You're correct.  A properly balanced pack can work great without a balance circuit.  However, you contradict yourself in your reply when you say "Doesn't seem to suffer from this" (getting out of balance) and then later in the paragraph state "When we have low capacity packs, it appears to be due to one cell having much reduced capacity."

Allow me to elaborate a bit, as I work with LiFePO4 cells a lot (a particular type of lipo chemistry).  For about 85% of an individual cells capacity, the actual voltage on the cell changes very little.  At either end (charged/empty) the voltage varies a great deal.  Under ideal circumstances, all cells have identical capacities and internal resistances, so these greater voltage deltas are reached simultaneously across all cells.  In the real world, this is rarely the case.  Properly tested and sorted cells can minimize the differences, but currently its very difficult to get all your cells to have identical internal resistance and capacity.

So, over the bulk charge of the process (the 85%) you're delivering a constant current to your pack.  The very slight differences in internal resistances mean that each individual cell is receiving very slightly different amounts of charge.  Couple this with a very slight difference in total capacity, and you will get one cell that will become an outlier.  Each charge/discharge cycle will push it further out of sync with the other cells.

Both excessive charge and excessive discharge are damaging to all lithium chemistries.  So in the case where one cell is reaching its full charge earlier than the other cells, it is possible that for a short time it is getting over charged because the total pack voltage is not yet up to the "full" point.  Or in the case where it reaches its discharged state earlier than the other cells, its possible to over-discharge it because the total pack voltage is not yet down to the "discharged" point.  Both can damage the cell and reduce its capacity.  Without balancing circuitry, the cell will remain an outlier and continue to be damaged by either the charge or discharge cycle -- or both.  By time the pack as a whole needs to be replaced, the problem cell shows a much reduced capacity because of this damage.

Yes, you can build lipo packs that don't use a balance charger and if you're careful in your assembly, they'll work just fine.  However, just because they work fine does not mean the problem has gone away.  Its still there -- minimized, certainly -- but the problem of cell balance still exists.  Many companies use packs built without a balancing circuit simply because its cheaper to replace a bad cell (or an entire pack) when it fails.  The extra cost of the circuitry, the cost of individual cells, and the potential increase in points of failure due to more complicated circuitry must always be considered during the design process.  One method isn't necessarily better than another when you take your considerations for design into account.   For one particular example, I have a 4-cell 100 AH battery running my computer off a solar panel.  Individual 100 AH cells are not cheap (particularly at the time I acquired them in a recycling project).  During the day when the sun provides more power than my computer uses, it charges.  During the night, when solar is unavailable, it discharges.  The balancing circuitry keeps the cells fairly consistent with each other.  The battery is 15 years old and when I tested it a couple weeks ago, it still has a capacity of 98.3 AH.  For my particular case, the balancing circuit was much less expensive than replacing a cell and has more than doubled the lifespan of the cells.

[psusi]

13 hours ago, KittenIsAGeek said:

Now, about the lipo packs you use: You're correct.  A properly balanced pack can work great without a balance circuit.  However, you contradict yourself in your reply when you say "Doesn't seem to suffer from this" (getting out of balance) and then later in the paragraph state "When we have low capacity packs, it appears to be due to one cell having much reduced capacity."

The idea of a balancer is that the cells have different state of charge, and the balancer fixes that.  The bad packs I've looked at don't have one cell that simply has less charge, but actually has less capacity.

13 hours ago, KittenIsAGeek said:

Both excessive charge and excessive discharge are damaging to all lithium chemistries.  So in the case where one cell is reaching its full charge earlier than the other cells, it is possible that for a short time it is getting over charged because the total pack voltage is not yet up to the "full" point.  Or in the case where it reaches its discharged state earlier than the other cells, its possible to over-discharge it because the total pack voltage is not yet down to the "discharged" point. 

The safety circuit required by law prevents this from happening.  Once any cell reaches its min or max voltage, the pack either stops accepting charge, or stops providing power.

 

13 hours ago, KittenIsAGeek said:

For one particular example, I have a 4-cell 100 AH battery running my computer off a solar panel.  Individual 100 AH cells are not cheap (particularly at the time I acquired them in a recycling project).  During the day when the sun provides more power than my computer uses, it charges.  During the night, when solar is unavailable, it discharges.  The balancing circuitry keeps the cells fairly consistent with each other.  The battery is 15 years old and when I tested it a couple weeks ago, it still has a capacity of 98.3 AH.  For my particular case, the balancing circuit was much less expensive than replacing a cell and has more than doubled the lifespan of the cells.

Wow, that is a BIG battery.  We only warranty our packs for 1 year but most still work fine after 5.   Good to know that a balancer can keep them working for 15 years... makes me feel better about my Tesla

I'm not sure if my Prius has a balancer but it is using NiMH.  I had to replace that pack last year at 10 years old when it failed.