Battery paralleling circuit

If they are being monitored, then Arduino can notify when one needs to be pulled out to be recharged.
I know my battery recharger is a whole separate unit that the battery pack gets plugged in to.

If the batteries will be recharged in place, why not just run off the chargers source voltage to start? So I think pulling them is reasonable. I didn't get the impression they would be charged in circuit.

If you mean basically just use the charger as a power supply, the problem is that charging is a fraction the current of use.

For example in flight batteries, a factor of 100 charging current to discharging is not unusual.

But yes, it did sound like the batteries would not be charged while in the device. There was mention of a campsite. Around here, no place plug in, so.

a7

One battery wired to a circuit, when its thingy drop then add in another battery is the thing do wanted.

When the new battery is add into the circuit, the old battery will become a charging capacitor once the new battery is added into the circuit. An operation that will need to be compensated for. Continued operation will see that the old battery will drain the new battery till the new battery is charged. The new battery's charge is then diminished.

When the 3rd battery is switched into the system, the other 2 batteries begin their charge cycle.

If the batteries were Lead Acid or LiFePo4 batteries, cool.

The reply from @CrossRoads is "useful". I suggest this build for powering the measuring controller. The controller can select the most suitable, preferred, battery supplying the real load when the current load powering battery starts to go low, needing a recharge. Indicating the candidate for charging is natural.....
Different strategies can be used.

  1. Round Robin, use each battery in sequence, one after the other.
  2. Use the most desired battery. Could be a poor battery in order to squeeze out the last of its lifetime, or use "the best" battery in order to level their capacities.

CrossRoads:
Have you considered running the batteries in parallel, with each going thru a diode, so that only the battery with the highest voltage at any particular time is the one supplying current?
This schottky diode for example, rated for 25V, 10A, you would lose 0.46V across it at high currents, with 18V battery pack that could be okay.
STPS10L25D STMicroelectronics | Discrete Semiconductor Products | DigiKey

If you want to monitor all the batteries, use a voltage divider at the battery pack output before feeding into the diode to bring it down to a level that a 5V or 3.3V device can read.

Thank you very much, this is a great suggestion! I have actually looked at Schottky diodes as an option, there are a few very cool circuits out there that will let you set the pick up and drop out voltages, but I suppose I was just getting greedy with wanting more available current to power larger devices, similar to a battery generator.

A single Milwaukee battery is only capable of outputting 30A, the datasheet of the cells they use are 15A and they are in series parallel, 10 cells total. Now obviously this is also the best case scenario, as I would not want to continuously draw the maximum allowable current through the battery, but that is roughly a maximum of 540W or so? Luckily the compressor cooler is only about 100W, so that is suitable for that particular case, and maybe additionally some lights. Once again, I suppose I was getting greedy with wanting more available potential power output.

I really appreciate everyone's input, there were definitely some factors I had forgotten to consider, I was pleasantly surprised to wake up to so many suggestions and advice :slight_smile:

duder56:
Thank you very much, this is a great suggestion! I have actually looked at Schottky diodes as an option, there are a few very cool circuits out there that will let you set the pick up and drop out voltages, but I suppose I was just getting greedy with wanting more available current to power larger devices, similar to a battery generator.

A single Milwaukee battery is only capable of outputting 30A, the datasheet of the cells they use are 15A and they are in series parallel, 10 cells total. Now obviously this is also the best case scenario, as I would not want to continuously draw the maximum allowable current through the battery, but that is roughly a maximum of 540W or so? Luckily the compressor cooler is only about 100W, so that is suitable for that particular case, and maybe additionally some lights. Once again, I suppose I was getting greedy with wanting more available potential power output.

I really appreciate everyone's input, there were definitely some factors I had forgotten to consider, I was pleasantly surprised to wake up to so many suggestions and advice :slight_smile:

If You wish, and Your batteries can supply, 15 or 30 Amps go for power silicon diodes. You will loose another 0.5 - 1 volt for those loads.
For such high currents, and Your intention to switch in and out batteries, use cells that, as single, can supply that current. That's the only way to monitor each battery, or run one good and one poor in parallell permanently, as one single unit.

Parallelling batteries is not a new question. In my opinion, don't ever use it. You will never be able to measure which one of the 2 that is the bad one needing to be exchanged.

Specify the particular needs for current. Different needs for current for different loads can be handled but a collected load, widely swinging will be much more difficult to handle.
Limit examples and variations. They will cause quite some confusion and useless guesses.

duder56:
... just getting greedy with wanting more available current to power larger devices...

Then you can use power mosfets as "ideal diodes" for separate the various batteries ... a little bit more complex than a bare diode, but with better characteristics (you cannot use them for low voltages, usually 9-10V is the minimum, but using 18V there is no problems at all)

Actually they are available ready as "solar panels antibackflow diodes" circuits, but can be built easily ... a P-channel power mosfet and a matched-pair of PNP transistors is the simplest circuit, or can be made with specific chips like LTC4376 (all-in-one til 7A max), or LTC4357 and LTC4359 (need a power N-channel external mosfet with these ones, the current depend from it, so can also be 100A if you choose the right mosfet), and so on ... :wink:

Forgot to say, some of these chips also have "enable" pin, so you can use them also as switches :wink:

Hi,
Different voltages don't accurately indicate the state of charge.

You need to look at battery theory and a parameter called INTERNAL RESISTANCE.

Tom... :slight_smile:

TomGeorge:
Hi,
Different voltages don't accurately indicate the state of charge.

You need to look at battery theory and a parameter called INTERNAL RESISTANCE.

Tom... :slight_smile:

Yes of course internal resistance must be considered, this is why paralleling different batches of batteries is complicated, but it is the voltage you must look at first and foremost when attempting to parallel a battery. The capacity of the battery was just one of many considerations, but my primary concern was the state of voltage as to not get too much backflow current.

TomGeorge:
Different voltages don't accurately indicate the state of charge.

Modern phones obviously measure voltage and launch a percentage in the display. My old phone, no SIM card, is used for surfing via Wifi. When it shows 70% it will soon cut off completely, without warning. Just to plug in the power pack....
Old, used, lead batteries have their unexpected behaviours well. Using "garbage" batteries and hope for high amperes is just nonsense.

as a possible different direction, or perhaps complementary . . .

there are a handful of websites explaining how to rebuild/refurbish this kind of battery pack.

Regardless of brand or voltage, they are all built from the same type of sell, with several soldered together. With a voltmeter, you may be able to determine which battery isn't carrying its share of load, and recharge just that cell up to match voltage to the rest, and then get more normal/less reduced performance from the pack. You might also use one pack as a source of donor cells to replace bad cells in the others.

From my reading, it seems that there tend to be one or two cells that go wonky while the rest are still behaving, and can be rejuvenated or replaced.

This will also bring your packs of varying age closer in voltage to one another, alleviating some of the other problems mentioned above.

These are probably worth a read for what you're trying to do, just for the second-hand knowledge from people who apparently spend a lot of time working on this kind of battery pack.

I found this when researching "80v" (actually 72, but using peak charge to measure) chainsaws, as I already have an 80v 5Ah kobalt battery. All five brands are actually the same pack with different grooves in the battery case and tool receiver, and you can cut new channels or add ridges to make them work!

[oh, and if anyone needs to know, an 80v chainsaw is a match for a gas saw, and a bit easier to control. Not that I recently lost a big tree in a storm . . .]

@dochawk
For many years I designed the electronics of heavy fork lift trucks, mostly using 48 volt lead batteries. One competing company used 120 volt battery pack that had to be reduced to below 100 volt for export. And yes, those 1 ton batteries were checked for bad cells if they didn't perform.
My most fascinating (?) battery right now is a litium battery in an old phone used for surfing the Internet. It charges up to 100&. Disconnecting the charger, running Youtube videos, checking the battery status regularly it some times cuts off, shuts down completely at 80% displayed. Sometimes I can be lucky and read 50%.
That battery is done, is a wreck. It has no capacity left. What good would it do in any parallell battery system?
Batteries connected in parallell, as told already, gives no opportunity to monitor any single battery voltage.
This project is waisted energy.

I'm not suggesting that it's a good idea.

However, the whole thing is likely to act less like a Samsung phone (or at least a smaller model) if the batteries are "less dissimilar."

I'm really thinking more that a rehabilitated pack would provide more current longer, potentially eliminating the whole parallel game. It sounds like one good battery is enough, and the parallel business is to make up for an individual pack.

I've mentally toyed with the notion of individually swapping/charging cells in a pack, and switching them in and out, but it's an awful lot of complexity with most of the problems mentioned above by so many. just to isolate, you're talking about 2n-pole/DT high current switching just to switch between charge and discharge modes for he n cells, n charging circuits, n voltage sensor lines . . .

it's easier to just buy another battery and move along to am ore interesting project.

OK, and it'd really like to think I did what I could to stop the explosion . . .

Railroader:
Modern phones obviously measure voltage and launch a percentage in the display. My old phone, no SIM card, is used for surfing via Wifi. When it shows 70% it will soon cut off completely, without warning. Just to plug in the power pack....
Old, used, lead batteries have their unexpected behaviours well. Using "garbage" batteries and hope for high amperes is just nonsense.

The point is that I'm not using "garbage batteries" I'm not sure where you gathered that from? All of my batteries have been well maintained, and are no older than 3 years at the most.

I used an extraordinary example. If You have a battery good for, lets say, 50 - 75% of Your needs and try to back it up with less good batteries, You spend a lot of time and work for gaining what? You'll never manage to extract every ampere hour that way.

I could se a good way in Your idea. That would be to have batteries available, in parallell, and selected one by one to power a long time consumer. Writing to a display telling which batteries needing charging... That could be a good UPS.

This topic was automatically closed 120 days after the last reply. New replies are no longer allowed.