Parallel "Groups" or Parallel Strings or Neither ?

amberwolf said:
So, if the pack is instead built using parallel groups of cells wired in series (the common way to build packs), balancing is easy, only needing one BMS, balancer, or RC charger. Same thing with monitoring LVC.
In other words the preferred build for a triangular pack is with one central Series String being fed by Parallel Groups ... https://endless-sphere.com/forums/download/file.php?id=262218 for making the best use of the smart BMS, for pack safety, performance and cycle life longevity.

So, as you and others have implied what's good for a triangular pack is good for a rectangular pack. Have you ever built a rectangular pack having just one center/middle Series String being fed by Parallel Groups with one cell from each Parallel Group contributing one of its cells to makeup the one Series String in a DIY rectangular pack build -- just as is the case with a DIY triangular pack build ?

Edit: Ok, as john corrected me earlier no cell in the P Group is counted as one of the S String cells to which that P Group of cells is connected to. Thought i better clarify before john jumps all over me. Now every time i post i have to be so careful to get it right or john will make a POINT of reprimanding me in front of the class.
 
John, please STOP your demeaning comments that border on disinformation if only to discredit and derail. It's apparent you've got an ax to grind. Start your own thread.
 
You are trying so hard to impress others with your head knowledge that it gets a little old; especially when you have to demean and discredit someone along the way. Lighten up and don't take a little humor once in awhile as a personal attack (e.g. Plz Xpln).
 
The problem with putting series strings together in parallel with only small gauge parallel connections at the cell level is, what happens when one cell in the pack is weak or is dead. ie high current will try to pass through small wires. Other than maybe some of those handmade Chinese packs that so commonly fail, no manufacturer in the world makes battery packs that way...for good reason.
 
eMark said:
You are trying so hard to impress others with your head knowledge that it gets a little old; especially when you have to demean and discredit someone along the way. Lighten up and don't take a little humor once in awhile as a personal attack (e.g. Plz Xpln).
Not at all.

My point is it is impossible to understand your questions

therefore impossible to help you understand how this stuff works

But I'm happy to give up, obviously communication's just not happening same bizarre phrases just getting repeated.

I'm honestly just trying to help.

Best of luck, maybe on other topics things will click.
 
eMark said:
https://endless-sphere.com/forums/download/file.php?id=262218
Any idea why three colours are used there, what they are supposed to represent?

 
John in CR said:
The problem with putting series strings together in parallel with only small gauge parallel connections at the cell level is, what happens when one cell in the pack is weak or is dead. ie high current will try to pass through small wires. Other than maybe some of those handmade Chinese packs that so commonly fail, no manufacturer in the world makes battery packs that way...for good reason.
I discard cells on the basis of regular benchmark capacity testing at the per-cell level, long before EoL looms near.

Also, fuses or maybe fusible links? since not in the main power path, a little extra resistance NBD?
 
eMark said:
Amberwolf's justification makes sense for just having one Series String in a triangular pack being fed by Parallel Groups for the sake of incorporating a "smart" BMS. However, for some reason that same elogic doesn't seem to apply to rectangular packs having 2 or more Series Strings paralled throughout the pack in which case the Series Strings are Paralleled by other cells (single/grouped) so as to be one happy functional family (if that makes e~sense).

Again: it makes zero difference what shape the packs are.

None.

Zero.

Seriesing parallel groups, or paralleling series strings, is electrically EXACTLY THE SAME in a round pack, a star shaped pack, a sphere, dodecahedron, rectangle, triangle, or tesseract. ;)

I think you may not understand how these packs are wired.

Pretty much every common pack is wired with cells paralleled into groups. Then those groups are seriesed to make the pack. Just like every pack in the diagrams.

Sometimes the spotwelded strips are laid out in a way that makes the series connections first, and then the parallel ones, but electrically it is the same.

If there are parallel connections at anything *other* than the main + and - ends of the series strings (such that cells are paralleled into 1S groups with series connections to each other), then the pack is not built as paralleled series strings. It is built as seriesed parallel groups.


And so too is the addition of a BMS as standard DIY build protocol even when there are two or more paralleled Series Strings. With the BMS balance leads only needing to be connected to the outer edge/closer of the 2 or more paralleled Series Strings in a rectangular pack.

No.

If you have multiple series strings that are paralleled, then there will have to be multiple sets of balance leads, one for each series string, because they are completely independent of each other at the cell level.

Each set of balance leads must then be connected to it's own BMS or balancer, or it's own RC charger, etc.

If you connect the balance leads together (parallel), then it's no longer multiple series strings that are paralleled, it's now electrically parallel groups in series.


SORRY for any confusion as i still don't understand why most all triangular packs only have one so-called Series String; whereas most all rectangualar packs have two or more so-called Series Strings ... http://endless-sphere.com/forums/download/file.php?id=98636 ... (bottom right diagram).
They don't.

That is not two series strings. That is showing three groups of 4p (four cells in parallel) in serie, exactly like every other version on that diagram.

They are all electrically the same, as far as that goes.

AFAICS, the only thing that diagram shows is different ways to setup the spotwelded series connections between the parallel groups, to try to get the best current flow out of each cell into the next parallel group.

Realistically, the very best current flow is not to have individual strips between the cells, but to have a big plate that connects all of the positives of one parallel group to all of the negatives of the next parallel group in the series set of groups. Most people don't build their packs like taht because they are using cheap nickel strip instead of custom-made plates for each pack. Some commercial packs are built that way, however.


Anyway, the diagram has nothing to do with whether the packs are built as paralleled series strings, or seriesed parallel groups, because everything shown there is the latter.


eMark said:
If it's OK with you all (as john mentioned) let's return to Pack Bulding 101 with this diagram ... http://endless-sphere.com/forums/download/file.php?id=98636 ... according to amberwolf he would prefer that the bottom right diagram (Perfect current share for High power) were wired with just one series string
No.

All of the packs in that are wired exactly the same way, as far as what I was discussing.

They are all three seriesed groups of four parallel cells.


What you are talking about is something completely different.



Please explain why that diagram by doctorbass may not lend itself to ideal BMS safety protection when using two or more paralleled Series Strings instead of just one Series String (i.e. Triangle pack) ... https://endless-sphere.com/forums/download/file.php?id=262218 ?
None of those are paralleled series strings.

All of them are seriesed sets parallel groups.

So ALL of those are electrically exactly the same regarding how a BMS or other balancing/protection/etc is wired to them.


And none of this has *anything* to do with what *shape* the pack is.
 
eMark said:
So, spinningmagnets belief makes e~sense in light of amberwolf's belief that it's best to have just one Series String being fed by Parallel Groups (whether a DIY triangular or rectangular pack).
No.

That is not my belief at all.

I only explained to you the difference between the two ways of wiring a pack made of more than one parallel set of series cells.

What I believe is best is completely irrelevant, and is not discussed at all in this thread (not by me, anyway, and since I havent' said what I believe is best, no one else can be discussing it accurately since they don't know).
 
Note taht I am replying to each of the replies above that has misconceptions, so I can be sure the correct information is being passed along.

There are multiple concepts being discussed in this thread, and some of them are being confused with each other.


eMark said:
amberwolf said:
So, if the pack is instead built using parallel groups of cells wired in series (the common way to build packs), balancing is easy, only needing one BMS, balancer, or RC charger. Same thing with monitoring LVC.
In other words the preferred build for a triangular pack is with one central Series String being fed by Parallel Groups ... https://endless-sphere.com/forums/download/file.php?id=262218 for making the best use of the smart BMS, for pack safety, performance and cycle life longevity.
No.

There is no "central" string of any kind.

What you have, in any pack, that is built of seriesed sets of parallel cells (necessary if you wish to use a single BMS or other monitoring/protection system), is exactly that: seriesed sets of paralleled cells.

*How* you series them is completely irrelevant to this. (though it may be relevant to how current flows thru the system, if it's high enough)

You're applying one choice in pack building (whether to series parallel groups, or to parallel independent series strings) to a completely different part of pack building (how to series the parallel groups).
 
Wow @amberwolf, fantastic effort to punch through the fog there.

Enormous gratitude and thanks, not just for your work here, but across the forums in general.
 
john61ct said:
MAXIMUM_AMPS said:
I thought I did, sorry if I was unclear.
No need, and just want to focus on the foundation layer first, otherwise more convoluted than necessary.

Those Makita modules' "lowest cell level groups" are not paralleled together, and that is the question at hand.

Can the two "dimensions" of interconnectivity be active at the same time?

Will all the "A's" cross-balance each other as if a single 35P without any active assistance?

And will each "35P A group" usefully get monitored **as if a single cell** by my hypothetical "voltage only BMS"?

Ignoring any other BMS-like functionality for now.

Alright, I'll keep it simple, omitting protection/balancing circuits, just cells and wires.

While it is true that cells do automatically balance their parallel'd neighbors, if they are significantly out of balance, a huge dangerous current can flow between them. Which is why there are no battery systems on the market like that. Modular batteries are "isolated" on the cell level to make them safer.

Simply put, by connecting them at the cell level, you'd be shorting together 2 groups of cells with potentially very different charge levels. For example: 1 parallel group of cells could be fully charged, and the other parallel group could be at 10% charge. A huge current flows from the fully charged cells into the lower ones, and 2 minutes later, you hear a loud FSSSHHHHTTTT and we have a lithium fire.
 
amberwolf said:
What I believe is best is completely irrelevant, and is not discussed at all in this thread (not by me, anyway, and since I havent' said what I believe is best, no one else can be discussing it accurately since they don't know).
Thank you for your patience and clarification without finding it necessary to run me down/over. Actually, i would be most interested (as would others) in what S/P wiring you believe is best as it isn't irrelevant and was the reason why i started this thread. However, now realize each has its pros and cons. Such a S/P scrimmage might only turn into an endless debate of S/P wiring "theology" :wink: as if there is only one correct way.

My misunderstanding as i was visualizing what i now realize was wrong thinking as there possibly being another way of S/P wiring as better suited to the function of a BMS, but now realize (thanks to You) that was stupidity on my part (just one Series String) and why john was so dissatisfied/upset with my questions.
amberwolf said:
There is no "central" string of any kind.
That's what i originally understood (until recently) before starting this S/P thread. Started thinking that perhaps a S/P pack could have just ONE Series String fed by Parallel Groups and in so doing grossly misinterupted some of the ES post/replies along the way ...
spinningmagnets said:
It has been widely embraced to parallel cells first, and then connect those paralleled cells in series to finalize the pack size and shape. I now believe there is less line resistance if the cells are connected in series first. This may cause the pack-building to be more fussy and time-consuming, but it might be worthwhile for some builds...
amberwolf said:
No. That is not my belief at all. Sometimes the spotwelded strips are laid out in a way that makes the series connections first, and then the parallel ones, but electrically it is the same.
BINGO! Now, beginning to understand Ron's belief as well as your belief. Also coming around to why John was/is so dissatisified/upset with my questions. It's apparent from his posts that John is very intelligent, but like a grandmaster of chess he easily becomes frustrated with those wired with a lower S/P IQ.
amberwolf said:
Realistically, the very best current flow is not to have individual strips between the cells, but to have a big plate that connects all of the positives of one parallel group to all of the negatives of the next parallel group in the series set of groups. Most people don't build their packs like that because they are using cheap nickel strip instead of custom-made plates for each pack. Some commercial packs are built that way, however.
Early on and until recently i grasped that concept from reading Micah Toll's DIY Lithum Battery book, some of his youtubes and some of the many informative youtubes by that guy with the electrified volkswagon van. Also Shawn McCarty's no-BMS youtube and other youtubes showing spotwelding nickel/copper plates to mammoth S/P cell packs. It's all interesting, but grasping with the fundamentals; especially the use of to or not to use a BMS in lieu of using a RC balance charger.
amberwolf said:
If there are parallel connections at anything *other* than the main + and - ends of the series strings (such that cells are paralleled into 1S groups with series connections to each other), then the pack is not built as paralleled series strings. It is built as seriesed parallel groups.

Each set of balance leads must then be connected to it's own BMS or balancer, or it's own RC charger, etc.

AFAICS, the only thing that diagram shows is different ways to setup the spotwelded series connections between the parallel groups, to try to get the best current flow out of each cell into the next parallel group.
Yes, i get the wiring logic of those self-explanatory diagrams which make e~sense. That said in order to better understand/grasp S/P wiring how would you express the S/P wiring of that bottom right S/P diagram ... http://endless-sphere.com/forums/download/file.php?id=98636 ... where would you attach the balance leads of either a BMS or a RC Balance Charger to the cells in that bottom right diagram. One would hope that we all can agree on the S/P expression for that bottom right diagram and where to attach balance leads ??
 
john61ct said:
Wow @amberwolf, fantastic effort to punch through the fog there.
Enormous gratitude and thanks, not just for your work here, but across the forums in general.
AMEN! and AMEIN!

Just felt it only right/respectful to reply to amberwolf's most informative post with my apology for trying to justify what is apparently my wayward S/P thinking with only one Series String. Would also welcome john's answer as well as any others to my S/P question at the bottom of my above post ... being that amberwolf may not check back for a couple days or may not want to initiate what could become a debate. That said if we can't even agree on the proper/correct S/P (or P/S) expression for that bottom right diagram ... http://endless-sphere.com/forums/download/file.php?id=98636 ... then maybe my stupid 1 Series (middle/center) String wasn't so stupid. Or in other words if that S/P (P/S) diagram is 2S would someone please explain why it can't serve just as useful a purpose if wired as 1S with the proper size busbars and proper sized parallel connections ...
 
Here is a pic of a partially-assembled pack that only has the series bus connectors assembled FIRST (it took me years to find an example)...

BatteryBusStyle5.png


And here is a pic of a pack that did the series connections first, and then added the parallel connections over that...(this is what I am recommending)

BatteryBusStyle9.png


There is nothing "wrong" with parallel first and then series next. However, if you are forcing the high series current to pass through a 0.20mm thick bus thickness 14 times (in a 14S pack), then...you are adding a (14 X 0.20mm =) 2.9mm thick pure nickel resistor in-line with the series current for the pack.

All of this being said, I have no objection to referring to paralleled cells as a "group" and seriesed cells being a "string". I now feel I have learned something new today...thanks to you all...
 
Thanks for posting the photos as it's starting to sink in as to why you believe (i.e. resistance) the way you do :)
I'm always learning something everyday ... never too old at 75 to learn something worthwhile everyday :)

So in the top photo it's impossible to even guesstimate how many cells in that Series "String" which may be part of a Powerwall ?

In the bottom photo is it possible for you to guesstimate how many cells in "Series" and how many cells in each "Parallel Group" ?

Whether a 14S8P (https://endless-sphere.com/forums/download/file.php?id=259994) OR 13S17P (https://endless-sphere.com/forums/download/file.php?id=262218) the number of total cells is determined by multiplying the number of Series cells by the number of Paralleled cells. Would it be correct to express the S/P numerology of the simple pack in the lower right bottom of this diagram ... http://endless-sphere.com/forums/download/file.php?id=98636 ... as 3S4P ... (having a total of 12 cells) ?

If so then isn't each of the other three wired diagrams also 3S4P ? Do i get a passing grade or am i a failure :(
 
The top pic with red cells is 14S / 4P, the lower pic with green cells is 20S / 7P.

In the graphic posted by doctor bass, all of the packs are 3S / 4P. He was attempting to persuade builders that a certain method of pack arrangement had a better current-sharing flow, compared to other arrangements (meaning, all cells are as equally burdened as is possible)

Again, I want to state...the BMS does not care if you series first, and then overlay the parallel bus-strip on top, or...parallel first, and then overlay the series strip on top. Either way will balance the cells just fine.

However, if you add resistance to the series current, then there will be more "voltage drop" when the system is under load. All electric bike systems have "some" voltage drop under load. If you series the cells first, and then parallel, the voltage drop will be less, but...whether that result is worth the extra effort...it is something that only you can decide for your particular application.
 
spinningmagnets said:
BMS does not care if you series first, and then overlay the parallel bus-strip on top, or...parallel first, and then overlay the series strip on top
I've only ever used wires never strips,

so that crisscrossing is not something I'd ever come across.

So I prefer to visualize the xPyS layout with wiring, batteries with posts like the old-school big lead 2V cells.

As with the DoctorBass diagram eMark's been fixated on.

Those type of connections never have more than one pair of wires per post, one inbound / one out, from an electron flow POV.

Can you (or anyone?) please ELI5 or point me to resources that will help me grok the idea of those strips overlaid so electrons are concurrently flowing in multiple paths or????
 
I would like to add that I understand that it is still very common to assemble the paralleled cell "groups" first, and they will then all equalize and act as one large cell. There is nothing wrong with assembling a pack this way, it works.

I have taken efforts to ensure that I have emphasized that the "new method" I am suggesting is not a huge benefit. However, I believe the physics involved are solid.

If you intend to draw 20 amps from a common 4P pack, then you will only draw 5 amps as a temporary peak from each cell (the cruise phase on a ride will draw low amps, the acceleration phase is what draws peak amps). If you only intend to draw a temporary 5A as a peak load from 10A cells, then all of this is just an academic discussion. The "worst" configuration will still work just fine.

But...lets just say that you paid extra to get authentic VTC Sony cells that can put out 30A peaks, and your tiny 4P pack is going to be tasked with providing 120A...you might want to consider using a copper/nickel sandwich bus, and you might want to series before overlaying with the parallel buses. Then again, maybe I'm wrong. However, if you are building an expensive high-performance pack, I'd suggest doing some homework and research first.

Don't automatically trust anyone, even me.
 
MAXIMUM_AMPS said:
While it is true that cells do automatically balance their parallel'd neighbors, if they are significantly out of balance, a huge dangerous current can flow between them
**Of course** you only parallel groups at the same SoC/voltage, I'm not an idiot.

And yes, also obviously, having cells very closely balanced is sine qua non for all kinds of reasons, in discussions with me, please take such factoids as given.

I hate to assume anything, but from the deafening silence wrt the actual questions I asked, seems to imply "sure that setup would be fine". . .


 
john61ct, here is a pic of a nickel "ladder-style" bus that performs both the parallel and also the series functions.
BatteryBusStyle7.png


Here is a pic showing the series current-flow of 25A through a common 5P pack.
Battery186508.png


Here is a pic of someone who is using solid nickel plates to perform the series and parallel functions.
BatteryBusPlate2.png


Many pack-builders buy 0.15mm thick nickel ribbon to make both the series and parallel connections. Making the parallel connections first and overlaying them with the series connections is easier. Making the series connections first, and then overlaying them with the parallel connections has less resistance for slightly better performance.
image.jpg
 
Personally I would never use pure nickel strip for series connections, because nickel has 4X the resistance of copper. Instead I'd used nickel plated copper strips to retain the easy to tab weld properties of nickel as well as its resistance to corrosion.Pure nickel strip is fine for the parallel connections. Cheaper and more readily available nickel strip is fine for the low current parallel connections. Also, I do agree with SpinningMags regarding making your series connections first, though in a 14s pack there are 13 series tabs with 2 connections on each, so in his example making the parallel connections first would send the current through an extra 26 layers of nickel. Plus it is more likely to result in a poor weld to a cell of a series connection.

I disagree with one statement earlier by AW. Pack shape can matter to some extent, because increasing the area of the electrical loop made by the pathway too and from the controller positive and negative increases inductance. Inductance is a bad thing on the battery side of the controller, because it slows changes in electrical flow and with our controllers switching thousands of time per second, too much inductance can cause stress on the system. For example, you would never want to lay out your pack so that the positive battery lead goes along the top tube with the negative going along the down tube of your ebike. Always lay your pack and wiring out to minimize the area of the electrical loop.
 
eMark said:
Actually, i would be most interested (as would others) in what S/P wiring you believe is best as it isn't irrelevant and was the reason why i started this thread.

Other than as I noted before, of just using single large plates to make the serial and parallel connections at the same time, I don't have a preference, as I don't have any personal empriical testing as to which way to lay the interconnects (series first or parallel first). ;)

If I go purely by theoretical stuff, I'd say that what Spinningmagnets said makes sense, about the series resistance. One thing you want to do in any current-carrying circuit like this is lower it's resistance, because all resistance does is waste power as heat, instead of allowing you to use it for useful work.

That's one of the reasons it may be preferable to use single large plates for each set of parallel cells, that make all the series and parallel connections for that group at the same time. It also lessens the resistance in boht series adn parallel paths, because ths plate is wider (and potentially thicker) tahn common nickel strips.


It's all interesting, but grasping with the fundamentals; especially the use of to or not to use a BMS in lieu of using a RC balance charger.

That entirely depends on your preferences, and your dedication (or lack thereof) to pack maintenance. If you want to just ride, and not worry about anything other than plugging the charger in when you finish a ride, then a good BMS on a good pack will let you do that. If you like to tinker and test, then either no BMS or just alarms and balancers (or manual balancing), or various other methods, might be your thing. Or a BMS with bluetooth that lets you monitor everything on your phone...

If you use a dependable RC charger, you're covered for charging and balancing. But you don't have anything to protect against overdischarge while riding, unless you use some form of cell-level monitoring (like cell alarms).

If you use a BMS, there are at least several types, from cheap to pretty expensive. If the pack itself is cheap, you don't want to spend a huge amount on the BMS. But a better BMS does a better job of keeping the pack safe, and can make it last longer, so it's a compromise for what things are important to you.

The crappier the cells are to start with (or the less suited they are to the project they're used in), the more you *need* a BMS (or something) to help them not die quickly.

The better the cells are, and the more closely matched they are (from same batch, etc) and the less hard they are pushed, the less you need a BMS (at least until the cells age and begin to drift apart in characteristics).

For instance, on my SB Cruiser trike's pack made of EIG NMC cells, I don't use a BMS because they stay balanced, and I don't use them anywhere near their capabilities 99% of the time. Every so often (maybe once a year, or even less often) I"ll check the cell voltages, just in case. So far, only once (on an older pack of these cells, used a little harder than the trike's) on an older bike, has a cell been out of balance when checked, and it was a bad cell (internal resistance went up significantly).

On a cheap 18650 pack Iv'e got, if it didn't have a BMS it probably would already be toast; even with the BMS it's not very good, and the BMS it has is not very good either (sometimes kinda "locks up" and doesn't allow further charging and doesn't balance, until it's discharged sometimes a fair bit). The cells it's made of are not designed for the use (discharge rate/etc) the pack was advertised for, either--the pack heats up quite rapidly.


where would you attach the balance leads of either a BMS or a RC Balance Charger to the cells in that bottom right diagram.
It doesnt' matter where the balance leads are attached. They can be wherever physically works for a particular pack construction. The only necessity is ensuring each parallel group has a balance/sense wire to the appropriate BMS/etc pin.


One would hope that we all can agree on the S/P expression for that bottom right diagram

All of those packs have the same S/P:

3S 4P

You can call them 4P 3S if you want. Doesn't matter.

As I said before, every one of them is exactly the same, electrically.
 
The normal thing outside of this weird tiny-cylinder world y'all live in, is each cell is **either** in a group **or** in a string, never both at once.

Often there is no paralleling, just one string of (say 1000Ah) cells.

If a cell is in a group, then **that group** is the unit that gets series'd together to comprise the bank.

Of course with lead, 4/6/8/12V batteries are strings already, and those strings can be grouped and then series'd as well to get to say 48V, but it's much better (as I'm sure you know) to use just use 2V cells like Rolls/Surette, to build big lead banks.

spinningmagnets said:
john61ct, here is a pic of a nickel "ladder-style" bus that performs both the parallel and also the series functions.
Sorry as I said, I think need to visualize with wires only, I've looked at thousands of these plate / strip pictures and I just don't grok the both-together concepts that way.

So I'm hoping someone can draw a diagram that uses wires to replicate this weird layout y'all use with strip/plates.

____

> Here is a pic showing the series current-flow of 25A through a common 5P pack.
Battery186508.png


Well there's no way to see which cells are positive end up and which negative

______
Here is a pic of someone who is using solid nickel plates to perform the series and parallel functions.
BatteryBusPlate2.png


Those are 14 groups there right, 8 of which are "plated", 6 not yet welded, right? So with a thick enough connector, just one is required to go from positive to negative to form the string and complete the pack, right?

If so then that's bog standard to me, the weird bit (for me), is the idea of an individual cell being connected to other cells in more than one way at a time.

 
amberwolf said:
All of those packs have the same S/P:

3S 4P

You can call them 4P 3S if you want. Doesn't matter.
I know people usually put the S value first, but on other DC electrical forums (fora?) it is considered more accurate to put the "lowest level" first.

So 4P3S it is, to be nitpicky.

That is, as long as things are wired "normally", with groups connected in series via one big fat wire each, only.

With this possibility (which I don't yet understand) of both S and P happening at the same time? I guess it truly matters not. . .

 
Back
Top