3
\$\begingroup\$

This is my first post on EE, so I'll try to be as simple but detailed as I can.

I have a 12S10P pack that I want to segregate into a 6S4P + 6S4P, 6S3P + 6S3P, 6S3P + 6S3P for "cautionary" LiIon balance charging. (My charger only goes up to 6S anyway..) I still want to draw power from all 6 at the same time in a 12S10P config. I broke out MS Paint and unleashed my art skills to demonstrate the masterpiece. In this picture, imagine the top right cells are being balanced charged. There are no diodes in place. I don't know if balance charging will interfere with other cells in the pack (the ones in blue.) If there is unwanted adjacent charging, how do I stop that from happening? Ask questions if you need clarification. Steamy pile

\$\endgroup\$
7
  • \$\begingroup\$ you can't balance parallel cells, only the 1st cells to reach full charge get bypassed, the others may be undercharged \$\endgroup\$
    – D.A.S.
    Commented Oct 26, 2016 at 5:06
  • \$\begingroup\$ As Tony said - if you have cells in hard parallel you will not be able to adjust balance on any individual cell. ie your 6s4p pack needs to look like 4 x 6s1p at balancing time. How much this matters probably depends on usage and cell quality and how close to limits you charge/discharge them. || Thoughts only. May be invalid. : If you "bottom balance by eg holding a joined 4p row at say 3V, you may be able to get reasonable matching. Also, if you hold at 3.7V or less for a while you get minimal charge rate, minimal energy in a cell and may persuade individual cells to be closer matched. \$\endgroup\$
    – Russell McMahon
    Commented Oct 26, 2016 at 11:36
  • \$\begingroup\$ @RussellMcMahon Oh yes, I know I cannot balance charge individual cells. I just want to know if charging one of the 6SPx rectangles will affect other 6SPx in the pack, since they are all connected in some way. \$\endgroup\$
    – heilstalin
    Commented Oct 26, 2016 at 14:57
  • \$\begingroup\$ @TonyStewart.EEsince'75 Balance charging probably isn't the word, and I know I cannot charge individual cells-But I still need to charge the 6SPx rectangles. I broke up the pack like so because Its probably slightly easier to detect a fault somewhere. \$\endgroup\$
    – heilstalin
    Commented Oct 26, 2016 at 14:58
  • \$\begingroup\$ If all cells are hard connected in an 6SNP arrangement and you do not disconnect strings while charging one string then if you are applying voltage across a 6S string to charge it and doing "something" (not specified by you) to achieve balance then you MUST be charging all strings. It is not clear why you would think this is not the case. Rather than drawing large arrays of cells in general form without details of what you wish to do or wish to try to achieve, you would probably be better off to draw a simplified diagram with just a few cells ... \$\endgroup\$
    – Russell McMahon
    Commented Oct 26, 2016 at 22:46

3 Answers 3

2
\$\begingroup\$

You can't balance the cells with this configuration (6SPx). The Problem if one cell is charged more than the others you'll not be able to figure it out. You have 12S10P configuration, for example if you use a balancing circuit with a configuration of 6S4P you'll balance each individual cell of four branches of your battery with different voltage which can leads to over voltage or under voltage of the cell and damage your battery pack. You can use your balancing circuit with 6S to only balance 6 cells in serial in your case. So you need about 20 balancing circuits to balance your individual cells safely.

\$\endgroup\$
3
  • \$\begingroup\$ I understand that I cannot charge individual cells, and that some may be undercharged. I broke up the pack for (what I assume to be) better fault detection and slightly better cell matching. They're all spot on 3.742v from the factory. I havent wired the pack yet, but if you have any suggestions it would be much appreciated. \$\endgroup\$
    – heilstalin
    Commented Oct 26, 2016 at 15:07
  • \$\begingroup\$ In your (new) picture it's look like you wire the cells with 4P12S and 3P12S configurations is different from 6S4P + 6S4P, 6S3P + 6S3P, 6S3P + 6S3P. In this case you can balance your branches with 6 BMS (one BMS for each branch xP6S) \$\endgroup\$
    – R Djorane
    Commented Oct 28, 2016 at 12:38
  • \$\begingroup\$ Yeah, I guess I made it sound more complicated than what it was. \$\endgroup\$
    – heilstalin
    Commented Oct 28, 2016 at 17:17
1
\$\begingroup\$

I have a 12S10P pack ... I still want to draw power from all 6 at the same time in a 12S10P config.

Assume there are no diodes in place.

I don't know if balance charging will interfere with other cells in the pack. If there is unwanted adjacent charging, how do I stop that from happening?

If all cells are hard connected in an 6SNP arrangement and you do not disconnect strings while charging one string then if you are applying voltage across a 6S string to charge it and doing "something" (not specified by you) to achieve balance then you MUST be charging all strings.

It is not clear why you would think this is not the case.
Rather than drawing large arrays of cells in general form without details of what you wish to do or wish to try to achieve, you would probably be better off to draw a simplified diagram with just a few cells.
I have shown a hard connected 3S2P array below. Using something like thatto explain what you plan and using enough detail to make your proposed method clear will help people understand what you are proposing.

Use this to (try to) explain how you think you could possibly "balance charge" one 3S string without affecting the other one.

There is a very good circuit diagram drawing tool incorporated in this system - click on the cct diagram icon at the top of the question window or use Ctrl-M to answer it. It is easy to use and quite intuitive (albeit a little "clunky"). I used wires to 'build' the battery pack as their battery symbol show 2 or 3 cells with no intermediate connections.

A major problem is that if you DO somehow balance charge a string it is then different in charge state from other strings and imbalanced outputs or inter-string current will flow.

What you SHOULD do is explain

  • what you wish to achieve overall
  • what application
  • what cells
  • what currents
  • what charger
  • what balancer

and then ask how to achieve what you wish.
If you do not yet have a balancer etc then asking how rather than telling us what you wish to do when you do not know how to do it makes more sense.

This is a rough 3S2P circuit diagram.
How could you balance the left string without the voltages affecting the right string?

schematic

simulate this circuit – Schematic created using CircuitLab

\$\endgroup\$
3
  • \$\begingroup\$ This is my pack. Can I charge the area in the red without affecting/charging the rest of the pack (in blue)? what application - Powering a 1kw motor what cells - Samsung 26F what currents - ~20A what charger - 6S, 6A Balancer+Charger \$\endgroup\$
    – heilstalin
    Commented Oct 27, 2016 at 0:49
  • \$\begingroup\$ @heilstalin Nice diagram. Which circuit lab was that drawn with ? - via embedded version here or elsewhere? | ALMOST there. Knowing HOW the balancer works and could be made to work is crucial to the answer. I'll assume the balancer can work magic as required. Then -> Answer: Only with EXTREME care and cunning. -> IF you measure the voltage across the pack before balancing AND if the main pack voltage does not change AND if you do not change the net voltage across the subset that you are "balancing", THEN you may be able to do this ... \$\endgroup\$
    – Russell McMahon
    Commented Oct 27, 2016 at 4:23
  • \$\begingroup\$ ... without affecting the others. But only if there is NO net current between the subset and the remainder of the pack at any time . IF you could monitor this it would make it easier (small sense resistor per 6S. eg 1 milliOhm at 20A = 20 mV at full current = much less at balance currents. And/but, as the balancer is not affecting the rest of the pack then opening the string (1 FET) during balancing and THEN voltage matching the string to the rest of the pack and then closing the switch may work as well and be easier. But the string is now not strictly the same as the rest of the pack. \$\endgroup\$
    – Russell McMahon
    Commented Oct 27, 2016 at 4:24
1
\$\begingroup\$

schematic

simulate this circuit – Schematic created using CircuitLab New batteries are balanced <0.1% but aging and C/40 usage can quickly age this imbalance to 1% and at 10% the battery is no good in the array. This can be measured by internal ESR and meter or with LED indicators in the array to help balance.

  • Smart balancers use a half bridge PWM with series L between each string cell and can get complex unless designed on a PCB or purchased.
  • dumb balancers use passive loading like TVS Zener with PTC protection rated power of 10% during CV mode charge current.
  • I suggested cheap Red Power LED's as both passive balancer and indicator if carefully chosen.
  • at first this may seem like overkill until batteries age and you want to extend life by balancing then removing mismatched failing cells.
  • IF you install thin short wires between cells, they can act as Fuses just like the ones in Tesla 2 batteries. Then you can reduce but not eliminate the risk of mismatched cells shorting out. Tesla batteries all have microfuses on each anode and cathode of 4000++ cells

  • then with <0.5% matching you can string many hundredsin parallel

** - compute max ESR of string and increase 25% with matched short ESR fuse wire divided by total N cells using AWG tables for ESR/mm **

\$\endgroup\$
5
  • \$\begingroup\$ I'm already going to make ~500 solder joints, I honestly don't see myself making any more than that due to time constraints. I just need to know: Will balance charging a 6SPx block (not individual cells) in my supplied picture, while everything is connected together, charge cells in the other 6SPx blocks? \$\endgroup\$
    – heilstalin
    Commented Oct 26, 2016 at 19:12
  • \$\begingroup\$ I can't say for sure, but what I recommend is use the wire resistance to match the ESR of the battery. Same with LED's in parallel I know ESR=1/Pd and in same batch I can get Vf @If within <1mV yet specs are +/-200mV so I discovered best way to put 3W LEDs in parallel is Normalize the ESR with a fixed wire resistance approximately equal to the LED ESR (1/3 Ohm in my case xx mOhm in your case as a safety against Thermal Runaway using a low Z Voltage Source. But if your charger is Current limited you need to ensure 1 string does not hog all current and lowest ESR will do this. \$\endgroup\$
    – D.A.S.
    Commented Oct 26, 2016 at 19:28
  • \$\begingroup\$ There is a risk of Thermal Runaway when Battery Voltage drops from lower Voltage chemical sensitivity (I believe) due to self heating thus drawing more current from other parallel cells.... this positive feedback effect can lead to self-destruction.. Using much less than max rated Charge speed reduces self-heating and this risk. In diodes. Vf drops with rising temps, I am not sure about LiPo's if it is PTC or NTC \$\endgroup\$
    – D.A.S.
    Commented Oct 26, 2016 at 19:32
  • \$\begingroup\$ Wouldn't it be easier to just put a bunch of thermistors throughout the whole pack wired to an arduino or 2 and just flick off a relay if the temp gets to a threshold? I'm gonna be drawing about 1200w max sustained from the pack and thats half of the theoretical maximum @ 25°c. Its simple, cheap protection. \$\endgroup\$
    – heilstalin
    Commented Oct 26, 2016 at 22:35
  • \$\begingroup\$ It is better to design current limiting than rely on OTP. My schematic is best for this, as a starting point. OK? Take a better look at balancer designs. rather than 24 thermistors and string relays \$\endgroup\$
    – D.A.S.
    Commented Oct 26, 2016 at 23:47

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.