Parallel "Groups" or Parallel Strings or Neither ?

eMark

100 kW
Joined
Nov 2, 2019
Messages
1,164
Location
Minne-apple, USA
The current thinking by at least one member seems to be that the Parallel cells in a 18650 pack should always be referred to as a "group" and never (God Forbid) as a "string" otherwise those in the KNOW will see it as "Ridiculous poppycock" if ever someone were to refer to parallel cells in a rectangular pack as a "string". This lock-step thinking is in no small measure (IMO) due to the preponderance of larger triangle packs (e.g. 14S8P, 13S17P or 12S6P). However, not all packs are wired the same. Let's use as an example a rectangle pack (14S5P, 12S5P or 10S5P) that can be wired at least a couple different ways.

One way would be to use all five Series Strings of 14, 12 or 10 cells with each of the five "strings" connected with busbars of sufficient size/magnitude. These would be tied together with parallel busbars between each of the cells in the five (14, 12 or 10) Series Strings. But to be correct these single (parallel) connections are perpendicular (not parallel :wink: ) to the Series "Strings", and as lone rangers they aren't in a group whose purpose is to maintain/equalize continuity (e.g. as a pack ages) in case a Series "string" cell should become somewhat disfunctional. Another way would be to just use two Series Strings with each one separated by single (so-called parallel) busbar connections. So whether a pack (rectangle or triangle) is 14S5P, 12S5P or 10S5P do we know if the pack has 5, 3, 2 or just 1 Series (center) "String" without seeing the layout sketch ?

Another way would be to wire a rectangular pack (14S5P, 12S5P, 10S5P) in 5-cell Parallel "Groups" with only the center "String" being a Series "String" of the appropriate number and size of busbars having sufficient size/magnitude to handle the ebiker's need for raw performance. In some of these DIY youtube triangular pack builds it appears there is only one meandering Series "String". Is that possible (only ONE)? Could that be one reason why another member eluded to his (non-longevity) belief that most ebikers using triangular pack layouts having Parallel "Groups" of 4, 6, 8 (and up) may be fortunate to get 200 cycles out of their expensive triangular pack with large Parallel "Groups" and only one meandering Series "String".

So here's 4 questions to consider:
(1) Has anyone ever buillt (or heard of someone who built) a triangular 12S_P 48V pack having as many as 6, 4 or even 3 Series "Strings" and if so would you post the layout sketch showing either the 6, 4 or 3 Series meandering "Strings" ?
(2) Has anyone ever built (or heard of someone who built) a triangular 14S8P 52V pack having two Series "Strings" with Parallel "Groups" of 4 cells instead of 8 cell "Groups" and if so would you post a layout sketch showing the two Series meandering "Strings" ?
(3) Can you visualize a rectangular pack layout where both the Series and Parallel cells are in a "String" OR when the Parallel cells are neither in a "Group" or a "String" ?
(4) Is there more than one meandering Series "String" in this 13S17P customized triangular? pack ... https://endless-sphere.com/forums/download/file.php?id=262218 ... and if so could you please explain/elaborate ?
 
No "personal attack" is intended here, please don't infer that, but I have no idea what your intended meaning is for most of your post, specifically what I quoted below.

I sincerely believe and hope what follows might help you understand this topic better.

Feel free to suggest other terms if you like for parallel groups, but strings do always signify serially connected sets.

Yes you **can** wire a given set of cells in different layouts, but the overall best way is also the simplest.

Parallel grouping **first** to get to the desired Ah capacity, then those groups strung together in series, in order to get to the required voltage.

Note the groups in effect become single cells at that Ah capacity; if single say 18Ah cells are available in a physical format that fits, that would be better (easier, better performing, more reliable and likely cheaper) than using 6P groups to get there.

A 6S pack of LI cells will always be ~22V nominal, and charged at 24-25V

Getting to that voltage is the only reason to choose a 6S layout. But since 22V is not used for ebikes much, we see other layouts, using LI chemistries to get to more usual "nominal voltages":

13S for "48V" packs, charged at 52-54V
14S "52V" nominal, apparently em3ev calls it 50V
20S for "72V"

It is not a good idea to connect strings in parallel within a single pack.

Modules at the same voltage (same S-count strings) **can** be paralleled; this is sometimes done to flexibly gain greater range.

More than 3 strings in parallel can lead to imbalance issues, uneven wear between the sub-pack modules.

I'm assuming the talk about overall shape - rectangle vs triangle, physically parallel and so on - making a difference to the electrical xPyS layout, is tongue in cheek?

If not, then please know that the one is not related to the other.

______
eMark said:
as lone rangers they aren't in a group whose purpose is to maintain/equalize continuity (e.g. as a pack ages) in case a Series "string" cell should become somewhat dysfunctional

Another way would be to just use two Series Strings with each one separated by single (so-called parallel) busbar connections. So whether a pack (rectangle or triangle) is 14S5P, 12S5P or 10S5P do we know if the pack has 5, 3, 2 or just 1 Series (center) "String" without seeing the layout sketch ?

Another way would be to wire a rectangular pack (14S5P, 12S5P, 10S5P) in 5-cell Parallel "Groups" with only the center "String" being a Series "String" of the appropriate number and size of busbars having sufficient size/magnitude to handle the ebiker's need for raw performance. In some of these DIY youtube triangular pack builds it appears there is only one meandering Series "String". Is that possible (only ONE)? Could that be one reason why another member eluded to his (non-longevity) belief that most ebikers using triangular pack layouts having Parallel "Groups" of 4, 6, 8 (and up) may be fortunate to get 200 cycles out of their expensive triangular pack with large Parallel "Groups" and only one meandering Series "String".

So here's 4 questions to consider:
(1) Has anyone ever buillt (or heard of someone who built) a triangular 12S_P 48V pack having as many as 6, 4 or even 3 Series "Strings" and if so would you post the layout sketch showing either the 6, 4 or 3 Series meandering "Strings" ?
(2) Has anyone ever built (or heard of someone who built) a triangular 14S8P 52V pack having two Series "Strings" with Parallel "Groups" of 4 cells instead of 8 cell "Groups" and if so would you post a layout sketch showing the two Series meandering "Strings" ?
(3) Can you visualize a rectangular pack layout where both the Series and Parallel cells are in a "String" OR when the Parallel cells are neither in a "Group" or a "String" ?
(4) Is there more than one meandering Series "String" in this 13S17P customized triangular? pack ... https://endless-sphere.com/forums/download/file.php?id=262218 ... and if so could you please explain/elaborate ?




 
If by "string" you mean a series of cells (none in parallel) that equal the pack voltage, and by "group" you mean a parallel set of cells (none in series) that equal the cell voltage, then the following are consequences of each.

If you mean something else, explicitly define that and we can explain the consequences of each.

It does not matter what the pack shape is, or physical shape of cells are used, it's the same results.


With multiple strings of series cells, paralleled only at the main + and - ends, keeping them balanced is potentially much harder than the other way. Each string would have to be balanced separately, either by having a BMS (or balancer unit) for every string, or using an RC charger balancing lead set for every string to charge them separately.

Similarly each string would need it's own LVC warning unit as well, so you don't overdischarge any cell while riding.

Alternately you can use the RC charger balancing lead set on every string, and connect those in parallel, and monitor those with one unit--but this means they are no longer separate series strings (because you have to leave these always connected), and you might as well have built the pack the common way, as below.

Now, there is an advantage to wiring as no parallel cells: one bad cell can't take down al the ones in parallel with it. HOwever: that one bad cell will, if not monitored, instead destroy the series string it is part of, because every other cell in it will be overcharged by the voltage of the pack divided by the number of series cells, minus the one that's internally shorted. The other possible failure mode of an internally shorted cell, related to the above but while riding instead of charging, is that the series string it's in drops a cell's worth, and then the strings paralleled to that drain into the string with a problem. How far they drain, is dependent on the number of cells in series vs the one that shorted. If the pack was full when it happened, it'll also be overcharging the other cells in series with the one that shorted.

Similarly, a stuck-on balancing shunt in a BMS or balancer cant' take out an entire parallel group of cells...but it has similar potential consequences to the above. (just that the cell it's drained is probably not internally shorted, so it will take charge, and it's likely taht all the cells are individually monitored in this situation, so overcharging the other cells in any other part of the pack shouldn't happen).

There are ways ot prevent cell strings from discharging into other strings (power diodes, etc), but that's added voltage drop, resistance and heat in the pack, as well as additional failure points in every string.


**************************************************

So, if the pakc 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.

A disadvantage is that a cell dies shorted in a parallel group, and drains all the other cells in that group. But this won't affect any other cell in the pack during discharge, at least. During charge, the entire pack's other cells would be overcharged if not monitored. Monitoring being much easier, that's less likely.

Similarly, a stuck-on balancing shunt on a parallel group will kill the whole group, but again, overcharging any other cell in the pack shouldn't happen because everythign is monitored.


*****************************************


A pack can be built either way, and it will operate the same either way under ideal conditions--but the failure modes are different, and the cost of monitoring for failures is much higher on the series string version, vs the parallel group version.
 
amberwolf said:
If by "string" you mean a series of cells (none in parallel) that equal the pack voltage, and by "group" you mean a parallel set of cells (none in series)
My understanding / usage is a bit broader, and of course you know all this, what follows is to help clarify for OP.

Yes a string **of cells** is connected in series, as are a string of "groups", which are paralleled cells at 1S.

But you can also connect a series of modules into a string, each of which has their own internal xPyS layout, e.g. ex Tesla modules.

But then that's an example outside of our usually smaller and simpler context here.

More common are "split packs", where a high voltage requiring a long string is broken down for ease in finding reasonably priced chargers and balancing gear.

So, a 24S system composed of a string of 3x 8S modules, each of which may also have paralleled cell groups internally, and may also themselves be grouped in parallel, when occasional long range riding requires high Ah capacity, but we don't want to carry all that extra weight around all the time

going with multiple (same voltage, xS string) modules connected in parallel as needed.

Pro's and cons existing for each of these different layout options, and simpler usually being better,

one monolithic pack, parallel first into groups at the lowest level, one big string to get to the target voltage.

Hope this helps.



 
I have been guilty of using "string" and also "group" when referring to cells, whether they are in parallel or series. I'm still on the lookout for a definitive reference, after which I will follow the proper electrical engineering term.

This might be a good time to mention something. It has been widely embraced to parallel cells first, and then connect those paralleled cells in series to finalize the pack size and shape.

Clearly that works, however, when hoping to draw high amps from a pack...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:
With multiple strings of series cells, paralleled only at the main + and - ends, keeping them balanced is potentially much harder than the other way. Each string would have to be balanced separately, either by having a BMS (or balancer unit) for every string, or using an RC charger balancing lead set for every string to charge them separately.
Assumed by now this was/is common knowledge to most that have been paying attention since this thread ... http://endless-sphere.com/forums/download/file.php?id=98636. In the bottom right photo (Perfect current share for High power) the impression is given that two Series Strings (with paralleling) are better than one and others might say three Series Strings would be even better in that diagram (rectangular pack) which is an obvious improvement over the others.

Anyway, i truly appreciate you taking the time to post your very informative reply. It will take me time to digest it all. So, according to your explanation it makes good DIY sense that the reason even a substantial 13S17P pack ... https://endless-sphere.com/forums/download/file.php?id=262218 ... has only one meandering 13S String is for ease of balance charging (with a smart BMS) the 17P Groups. Is that correct or did i miss another reason(s) with respect to a triangular pack with several Parallel Groups and only one meandering Series String?

So getting back to the above 4 questions let me put/ask it another way. Which of these two rectangular pack options would make the most electrical sense (performance and/or cycle life longevity) whether it be: 14S5P, 13S5P, 12S5P or 10S5P?
(1) One center Series String (whether it be with 14, 13, 12 or 10 cells) with the same number (14, 13, 12 or 10) of 5-cell Parallel Groups feeding into each cell of the center Series String.
OR
(2) One center Series String (whether it be with 14, 13, 12 or 10 cells) with a short string of only 2 (interconnected) Parallel cells on each side interconnected to and feeding into each of the cells in the center Series String.

Correct me if i misinterrupted your reply, but got the impression in your above triangular DIY pack rationale that having more than one Series String down the center of a 14, 12 or 10 cell rectangular pack isn't good electrical sense but it is in a triangular pack? Yet, in videos about soldering the BMS balance leads (whether 14, 13, 12, 10 or whatever) the impression is that there is more than one Series String, but the leads only have to be soldered onto one of the Series strings because of the Parallel connections. So, what am i missing as there must be another reason why there is only one Series String meandering among the Parallel Groups in all triangular DIY pack builds.

EDIT Note: I just now read Ron's reply after posting this and wouldn't you know it that now i'm getting confused :thumb: which is to be expected :confused: My fault not john, amber or Ron as i truly appreciate the posts even if at first i don't completely understand ... it takes time to digest.
 
spinningmagnets said:
I'm still on the lookout for a definitive reference, after which I will follow the proper electrical engineering term.

This might be a good time to mention something. It has been widely embraced to parallel cells first, and then connect those paralleled cells in series to finalize the pack size and shape.

Clearly that works, however, when hoping to draw high amps from a pack...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...
Reading between the lines of Ron's post raises more questions than answers to any of my 6 questions, unless i'm blind 8) . So reading between the lines what was Ron's intent? Was it his way of saying there is no "right" answer to my questions if most everyone has their own opinion (even electrical engineers)? Ron's following comment is certainly interesting to say the least ... "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."

Who in their right mind is going to argue or debate Ron's belief when even Elon Musk might not know whether it's meant as an insider joke or from a man more brilliant in his own right than Musk himself.

By the way for clarification (FWIW) on one of my last questions it wasn't a typo or mistake ...

(1) One center Series String (whether it be with 14, 13, 12 or 10 cells) with the same number (14, 13, 12 or 10) of 5-cell Parallel Groups feeding into and contributing one of its Group cells to the Series String of ten cells.

One cell in each of the ten 5-cell Parallel Groups (10S5P) makes up/contributes a cell to the one central 10S String. Have a photo showing this, but unable to upload it. In some ways one would think some similarity to the 17P cell Groups in this 13S17P pack ... https://endless-sphere.com/forums/download/file.php?id=262218 . At first (for whatever reason i know not why) i assumed one cell from each of the 17P Groups was one of the cells in the one lone meandering Series String of 13 cells. So, can someone please explain how you as the DIY builder decides which of the four 17P Groups in that 13S17P pack doesn't contribute a cell to the one meandering Series String of 13 cells?
 
Strings to me means multiple cells or batteries, so it applies best to series. Parallel "groups" can essentially be treated as if together they are a single cell with the summed amp-hour capacity. While you can group series strings together in parallel, if they aren't paralleled together at the cell level it is sub-optimal, because you'd have to check every individual cell for balance.

I don't know if there's technical nomenclature to say cells in parallel can't be called a parallel string or cells in series can't be called a series group, but I do think it would be contrary to common convention and potentially lead to confusion or worse devastating connection mishaps.
 
eMark said:
So, according to your explanation it makes good DIY sense that the reason even a substantial 13S17P pack ... https://endless-sphere.com/forums/download/file.php?id=262218 ... has only one meandering 13S String is for ease of balance charging (with a smart BMS) the 17P Groups. Is that correct or did i miss another reason(s) with respect to a triangular pack with several Parallel Groups and only one meandering Series String?
There's no difference regardless of the shape of the pack for the principle.

Cell monitoring and maintenance, regardless of method, is much easier and less expensive if all of the cells are wired into parallel groups that effectively make them one large cell.

Doesnt' matter if it's a "smart" BMS, a dumb one, separate balancer boards and LVC / HVC boards, or RC chargers, etc.

Functionally, the pack will operate the same whether you make parallel groups and then series them, or make series strings and connect them only at the main + and - of each string, *as long as all cells stay balanced at all states of charge*. Under those specific conditions, pack-level HVC and LVC (at the bulk charger and at the controller, respectively) are sufficient to prevent cell damage.

But realistically, common 18650 packs wont' stay balanced for very long; they get used harder than they should, discharged deeper, etc., and they aren't made from matched cells to start with, so the cells start out with different characteristics which only grow more different with time.

Packs made from matched cells, that aren't used near their limits, are more likely to stay balanced, and are less likley to matter which way tehy're wired up, without balancing connections on each one, etc.

Packs made from larger EV grade cells, that aren't used near their limits, are probably even more likely to stay balanced.

But everything changes with time, so eventually even the best most well-matched and least-abused pack will eventually come unbalanced, and that's when it will start mattering how the pack was built and what kind of balancing / monitoring access you have to the cells. ;)


So getting back to the above 4 questions let me put/ask it another way. Which of these two rectangular pack options would make the most electrical sense (performance and/or cycle life longevity) whether it be: 14S5P, 13S5P, 12S5P or 10S5P?
(1) One center Series String (whether it be with 14, 13, 12 or 10 cells) with the same number (14, 13, 12 or 10) of 5-cell Parallel Groups feeding into each cell of the center Series String.
OR
(2) One center Series String (whether it be with 14, 13, 12 or 10 cells) with a short string of only 2 (interconnected) Parallel cells on each side interconnected to and feeding into each of the cells in the center Series String.
I don't really understand what you are asking about. You should make a drawing and attach it to your post, with exactly what you mean.

Also, again, the pack shape makes zero difference to how the pack configurations I posted about work.

IF you are talking about how to interconnect the parallel groups of a pack serially to make pack from them, then there are a number of pack build threads that discuss that, and Spinningmagnets' post in this thread has good info, and there are at least one or two good threads just talking about different ways to interconnect cells in a pack when spotwelding. Sorry I don't have links, I can't remember the names of the threads. :/



Correct me if i misinterrupted your reply, but got the impression in your above triangular DIY pack rationale
Again, as I stated in my first post in this thread, pack shape is irrelevant to what I was talking about.

that having more than one Series String down the center of a 14, 12 or 10 cell rectangular pack isn't good electrical sense but it is in a triangular pack? Yet, in videos about soldering the BMS balance leads (whether 14, 13, 12, 10 or whatever) the impression is that there is more than one Series String, but the leads only have to be soldered onto one of the Series strings because of the Parallel connections. So, what am i missing as there must be another reason why there is only one Series String meandering among the Parallel Groups in all triangular DIY pack builds.
I don't know for sure, but it sounds like you are not talking about series strings and parallel groups, but simply about the specific pattern of spotwelded interconnect strips. It sounds like all of the packs you are describing are all built from parallel groups that are then connected in series to make the pack.

You'll need to draw specific diagrams to illustrate exactly what you are describing in each specific case, for us to be sure.
 
Yes, if you want us to understand your questions you need to not only stop using standard words in non-standard ways, but also stop injecting new words that have no standard meanings.

Also, I suspect your "reading between the lines" is getting ahead of yourself, first focus on understanding and using 101-level terminology correctly.

Locational words like "center/ central" here

> One center Series String (whether it be with 14, 13, 12 or 10 cells) with the same number (14, 13, 12 or 10) of 5-cell Parallel Groups feeding into and contributing one of its Group cells to the Series String of ten cells.

> One cell in each of the ten 5-cell Parallel Groups (10S5P) makes up/contributes a cell to the one central 10S String.

makes no sense, afaict none of it does.

> Have a photo showing this, but unable to upload it.

Already advised

Also, "meandering" has no clarity here :

> assumed one cell from each of the 17P Groups was one of the cells in the one lone meandering Series String of 13 cells.

In a normal pack with parallel connections, each "cell" is actually a group. Each cell in that group is electrically "invisible", has no identity or function on its own.

IOW **all** the cells in that group function as a single cell as far as any device is concerned, BMS or balancer or balancing charger

> So, can someone please explain how you as the DIY builder decides which of the four 17P Groups in that 13S17P pack doesn't contribute a cell to the one meandering Series String of 13 cells?

Again, makes no sense.

By definition, 13S17P means each of the 13 groups has 17 cells.

There is no "four" of anything.

No group ever "contributes a cell", that phrase has no meaning.

 
OP please read this bit carefully:

> I don't know if there's technical nomenclature to say cells in parallel can't be called a parallel string or cells in series can't be called a series group, but I do think it would be contrary to common convention and potentially lead to confusion or worse devastating connection mishaps

But please ignore what follows, at least for now.

______
Although replying to John, these questions are also addressed to @amberwolf, @spinningmagnets, @Doctorbass @liveforphysics @dogman dan @methods if he's around, even @flippy in constructive mode :cool:, really anyone who knows more than me on the topic.

Of course please link to past posts/threads where / if this concept has already been discussed at any level of detail, especially to diagrams or drawings.

John in CR said:
While you can group series strings together in parallel, if they aren't paralleled together at the cell level it is sub-optimal, because you'd have to check every individual cell for balance.
I've come across this idea a few times now. Bear with me while I set a hypothetical context, and posit a "standardized" meaning for another term,

besides "group" for paralleled units and "string" for series'd ones.

Now "modules", for standalone-functional packs, that can **also** be used as sub-pack units, combined either/both in series for higher voltages, and/or paralleled to increase overall Ah capacity.

* usually portable, physically self-contained
* at a specific useful/standard voltage
* can be used / charged separately
* and/or wired themselves, in series or parallel
to comprise a main bank, maybe even rotated in and out

As an example, an off-grid home (or vehicle, trailer or boat) uses a very large capacity backbone main "mothership" bank at nominal 96V, while auxiliary ebikes or a kayak/dinghy, or portable powerpaks

use different combinations of modules to comprise packs at 12/24/48 volt.

______
As a more specific example,

using LI chemistry, call it 3.6V, each cell @6Ah,

so 5P7S modules to yield 30Ah @ ~"24V", charged at 29V

Now, let's series four of these modules say for a cargo trike, can be represented as:

4S(5P7S)

for 30Ah @ ~101Vnom, charged at 116V

________
OK, now (finally, thanks for following so far) my questions, triggered by your mention of

xPyS sub-pack modules "paralleled together at the cell level"

knowing that that level here, as the lowest one, is the 5P groups, rather than individual cells.

Let's label each of those seven groups A to F, so if measuring voltage at each serial connection:

A = 3.6V, B = 7.2V, …F = 21.6, and

at the module level (main power leads) = 25.2V

If I understand this idea correctly, it is

{? possible | a good idea ?}

to connect these corresponding groups, in **each** of the 7S modules to each other in parallel so, **all** the A's to each other, all the B's, all the C's etc

**at the same time** as the modules are connected in series via their main power leads (?? if true, !!! mind blown !!!)

Also, there could be **a single** 7S BMS monitoring **all four** modules while in use to provide against [over-voltage while charging and under-voltage from discharging] at the cell / group lowest level.

(Please assume / accept that all balancing functions, temperature and (bank-level) voltage and over-current protections, are provided by additional separate specialised BMS-type devices.)

That same 7S BMS can be used to perform this cell/group-level [HVC/LVC protection] with **any** number of modules, no matter how many are wired in series

or also if the resulting strings are further paralled

so that the impact of uneven resistances causing imbalanced current rates between modules is eliminated or at least ameliorated, as all 20 of the A cells will naturally (without any active circuitry involved) tend to balance v/v each other over time.

True?

(?? if true, !!! mind blown !!!)

What say y'all?

Any and all feedback welcome, from anyone who knows more than me on the topic.

For others, let's hold off on further questions / comments for now, especially OP eMark, take all this as input-only for now, no insult intended though OK?

I suspect this idea (if valid) deserves its own thread.




 
OK, let's back up a moment. It was getting to be a long day yesterday when i posted this last question at the bottom of my previous post. A question i later realized was stupid during a relatively sleepless night-- so apologize for this non-sensical question as i get-it, get-it, get-it so no need to explain my temporary brain amnesia (senior moment :wink: ) ...

eMark said:
So, can someone please explain how you as the DIY builder decides which of the four 17P Groups in that 13S17P pack doesn't contribute a cell to the one meandering Series String of 13 cells? ( https://endless-sphere.com/forums/download/file.php?id=262218 )

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). 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. 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).

John, i apologize for not replying to your posts which i do get a definite sense that you are very knowledgeable about DIY pack building (do's and don'ts). Just have a difficult time concentrating on your double spaced individual sentences without any grouping into paragraphs for ease of readability and understandability of the context. It just takes me longer to read again and again to digest the implication as there is definitely 'food/meat for thought' to digest. Just like a cow i have to re-chew (re-read) your cuds a few times before being able to digest them.
_____________________________

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 down/between the center/middle of the parallel groups in doctorbass' diagram for the sake of getting the best safety performance from a 'smart' BMS (right---wrong?) 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 ?

Also, want to get back to spinningmagnets (Ron's) intriquing comment. Even though i don't understand his rationale i assume it's important to explore further in this Series/Parallel thread as Ron chooses his words carefully with relevant purpose ...
spinningmagnets said:
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...
 
I have had only 1 cup of coffee.

Proof of that is seen here, by date-stamped picture of this empty (but pilfered) coffee cup

View attachment 2

In trying to be of assistance to my fellow enthusiasts I find myself at a loss. What comes to mind is that MORE PICS is always a good solution.

Two strings in Parallel.png

1) Space out Text (as with legal docs)
2) Break up text with pictures

Doing so allows those who are still in early stages of MorningTardation to understand.



It is ok if your pictures are hideous and ugly and unprofessional.
It is better to have a bad picture than no picture.

...

I will return to try and help once I can understand. My decoder goggles are back at the Magic Shop getting worked on.

-methods
 
Don't mean to be disrespectful, but it would be helpful if methods would not make light of this Series/Parallel discussion as if he has all the answers. Perhaps, methods can explain why most all triangular packs have just one Series String; whereas most rectangular packs have two or more paralleled Series Strings (maybe by his third cup of joe), but don't hold your breath :wink: .

Perhaps method can also expound upon the rationale and further enlighten us on spinningmagnets post ...
spinningmagnets said:
This might be a good time to mention something. 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...
 
john61ct said:
I've come across this idea a few times now. Bear with me while I set a hypothetical context, and posit a "standardized" meaning for another term,

besides "group" for paralleled units and "string" for series'd ones.

Now "modules", for standalone-functional packs, that can **also** be used as sub-pack units, combined either/both in series for higher voltages, and/or paralleled to increase overall Ah capacity.

* usually portable, physically self-contained
* at a specific useful/standard voltage
* can be used / charged separately
* and/or wired themselves, in series or parallel
to comprise a main bank, maybe even rotated in and out

As an example, an off-grid home (or vehicle, trailer or boat) uses a very large capacity backbone main "mothership" bank at nominal 96V, while auxiliary ebikes or a kayak/dinghy, or portable powerpaks

use different combinations of modules to comprise packs at 12/24/48 volt.

Ahhhh, now I understand what you meant in that other thread!

I've seen this done in a smaller extent with power tool batteries. I forget which brand, but you'd see their large tools used multiple of the "standard" batteries in series. I thought it was fairly neat.

This concept is interesting, but IMO requires the BMS chips to have some kind of detection/communication between them, at the very least least for parallel groups. (to implement safe parallel connections without a diode)

The limiting factor for series connections would be the blocking rating of the FETs. For example consider the following scenario: You have 6 of these 24V modular mini-batteries in series (144V), each with its own FETs to shut off the output...

Things are going great until one battery hits the LVC. It shuts off its output FETs to prevent overdischarge, like normal. The FETs are rated to block 100V but now that they're off, 120V (144-24V) is seen across the FETs. POOF, the FETs let out their magic smoke, fail as a short circuit, crippling the BMS's LVC/HVC protection, making way for a very bad scenario.

You might ask: "Well why not use relays?"

The relays would have to be highly specialized HVDC relays (bulky, $$$$), "normal" relays are not big fans of switching large DC voltages either. One of the many reasons our grid uses AC is because AC tends to automatically quench arcs at the zero crossing. DC just keeps going and going. This is why a relay might be rated at 250V for AC, then 30V for DC.

https://sound-au.com/articles/relays.htm#s2

Alternatively, the BMS could put all its trust in the load (and the charger, if charging in series) to shut itself off upon request, with a RESET pin or something.

With a modular system, there is a lot of risk in this. What if for some reason, there is no continuity between the battery's RESET pin and the charger/load? Connector might not seat properly, might corrode, wires break, etc.
 
MAXIMUM_AMPS said:
This concept is interesting, but IMO requires the BMS chips to have some kind of detection/communication between them, at the very least least for parallel groups. (to implement safe parallel connections without a diode)
I know BMS is your schtick, but that's a bit premature here, any chance you could address the fundamental questions I actually asked?

Is the conceptual layout itself sound?

If I understand this idea correctly, it is

{? possible | a good idea ?}

to connect these corresponding groups, in **each** of the 7S modules to each other in parallel so, **all** the A's to each other, all the B's, all the C's etc

**at the same time** as the modules are connected in series via their main power leads (?? if true, !!! mind blown !!!)

Also, there could be **a single** 7S BMS monitoring **all four** modules while in use to provide against [over-voltage while charging and under-voltage from discharging] at the cell / group lowest level.

(Please assume / accept that all balancing functions, temperature and (bank-level) voltage and over-current protections, are provided by additional separate specialised BMS-type devices.)

That same 7S BMS can be used to perform this cell/group-level [HVC/LVC protection] with **any** number of modules, no matter how many are wired in series

or also if the resulting strings are further paralled

so that the impact of uneven resistances causing imbalanced current rates between modules is eliminated or at least ameliorated, as all 20 of the A cells will naturally (without any active circuitry involved) tend to balance v/v each other over time.

True?
 
methods said:
file.php

Two strings in Parallel Cell Level.png[/attachment]

Close but yours is a 2P(1P4S) example of the concept.

Whereas my four modules are connected in series

4S(5P7S)
 
eMark said:
Don't mean to be disrespectful
Comes across that way, very aggro, unnecessarily so and unlikely to improve the odds of people taking time to help you with your confusion.

And anyway, I believe methods was considering my separate question, not your issues.

 
eMark said:
explain why most all triangular packs have just one Series String; whereas most rectangular packs have two or more paralleled Series Strings
Again, shape has nothing to do with it.

The right way for a single pack is one big string of single cells, or paralleled groups if cell capacity is too small.

Some may choose to serial connect single cells, then parallel the resulting strings, but it is **definitely** not true for "most" rectangular packs.

spinningmagnets' hypothesis is interesting, but certainly not proven afaik, and in fact I've never seen a pack done that way.

I would if I were you, just set that advanced idea aside for now and as methods advised, see if you can make some layout drawings to confirm the basic principles are gelling.
 
john61ct said:
I know BMS is your schtick, but that's a bit premature here, any chance you could address the fundamental questions I actually asked?

Is the conceptual layout itself sound?

I thought I did, sorry if I was unclear.

Yes the concept of modular mini-batteries is sound at a small level, and has already been done with power tools. But in no way is it possible without a BMS for each module, that would be extremely dangerous. Sorry if I bored you with the small details of how it would work.

ALSO: I found an example: https://www.makitatools.com/products/details/XBU02PT1
 
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). YES, john61ct i'm all onboard with your belief that Series "String" and Parallel "Groups" is the correct teminology in light of amberwolf's post, your posts and spinningmagnets belief ...

spinningmagnets said:
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...
Perhaps "due diligence" shouldn't be viewed as "fussy" while fusing may be "fussy", but that's for another thread :wink:

The Bottomline: It's best if there's only one center/middle Series String in a rectangular pack being fed by the Parallel Groups on each side. One cell in each Parallel Group contributes to and becomes one of the cells in the Series String. Isn't this basically what spinningmagnets is getting at?

In other words having only one Series String (fed by Parallel Groups) in a retangular pack is the preferred cell layout for getting the best use of a BMS and the best cycle life longevity and performance. IF this is indeed true (not just for the sake of making the best use of the BMS) then doesn't the bottom right diagram by Doctorbass need to be revised ... http://endless-sphere.com/forums/download/file.php?id=98636
 
my attempt at a diagram

each little block is a 5P group, comprising

7S strings (vertical), joined *4 to create a 28S string, power leads at pack level ~101V

Each string's groups are A-G

which are parallel-joined via balance-lead pairs to the other three corresponding groups (ommitted negative lead for clarity)

All of these connections being live concurrently.

Does that work?IMG_20191206_133259.jpg
 
You apparently posted before reading my latest post. Would appreciate your take on my post ... while i study your post as it relates to amberwolf's and spinningmagnets belief. That is having just one center/middle Series String in a rectangular pack being fed on each side by Parallel Groups with one cell in each P Group becoming one of the cells in the Series String.
 
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.
 
Mark, your phrasing is still causing me to think you aren't getting it. At this point give up on figuring out what spinningmagnets wrote, it's leading you down some garden path.

eMark said:
just one Series String being fed by Parallel Groups

Drop the redundant (thus confusing) adjectives, a group is paralleled internally, a string is a series, both by definition.

The groups (acting as single cells) get put into a single string, and done.

Do not reference anything to do with physical shape or location, all that is irrelevant.

no triangles vs rectangles

nothing is "central" or "middle".

There is no function of "One cell in each Parallel Group contributes" anything, all the cells in one group together are a single unit

nothing "feeds" that single string

those group-units **comprise** the string, the string **is** the pack, nothing is "outside" the string.

Now, the Doctorbass drawing http://endless-sphere.com/forums/download/file.php?id=98636

All of those are **the same exact** logical xSyP layout, at the level being discussed here so far.

Just different physical wire-length schemes, relating to best practices keeping things "in balance" (nothing to do with BMS).

Similar conceptually to what's discussed by Gibbo at this famous page http://www.smartgauge.co.uk/batt_con.html

but those are in fact string units being paralleled, as is common with lead using 12V batteries, and not a best practice their either, so don't let yourself get derailed discussing it here, it will just confuse things more.


 
Back
Top