0

For a project at uni, I'm trying to learn to work with 18650s and TP4056s. One of the tasks is to be able to USB-charge three 18650s (Panasonic NCR, 3350 mAh each) in parallel, whilst also being able to draw current from the batteries (not when charging; we already have a switching circuit to separate charging and discharging automatically, but this is besides my question).

After days of research, it seems to me that over various forums, people are split on the use of multiple batteries and possibly multiple TP4056 in this regard. I've seen people recommending for and against the use of each of the following three scenarios:

  • One TP4056 with a 3P parallel pack;
  • Three TP4056s in parallel with each having one battery;
  • Three TP4056s in a series/parallel configuration with the first having a 3P pack.

This is an example of people approving the first, this is an example of someone approving the second, and this is what I mean by the third configuration.

All these seem conflicting to me, and so I'm wondering which is considered bogus, and which is considered good practice. Personally, I'm most inclined to go with the first approach, since I don't see why a TP4056 would treat three (balanced) batteries differently from a single battery with triple the single-battery capacity. A diagram of this proposal - not including switching mechanisms as mentioned since those are not fundamental to the question at hand, and therefore assuming the TP's OUT will never be outputting anything when there is current flowing into the IN - is found all the way at the bottom.

I was also told it'd be bad to connect the inputs of three TP4056s in parallel to one USB breakout board, and I don't know what's up with that (I'd be fine in using diodes, if current flowing in and out of the modules would somehow be the problem). This would of course not be an issue with my proposed way of doing things, but it would be for the other two methods.

Furthermore - notify me if this requires me to open a separate question -, this SE thread's accepted answer advises that one connect a protective DW01A in between TP4056s and batteries, but as far as I know, that exact circuit is already present on the TP4056 itself. I have read about there being TP4056s on the market with and without protective circuitry, and I'm thinking this might be what that SE thread is referencing, but I might be wrong about this. Any extra thoughts on this last point appreciated.

Edit: Added diagram. I forgot to add that the charging time is of little importance.

Mew
  • 141
  • 7
  • 1
    What does the charger manufacturer recommend? – Leon Heller Jun 27 '19 at 17:34
  • Please post a schematic of what you are proposing. – winny Jun 27 '19 at 17:41
  • @LeonHeller: The [TPower spec sheet](https://dlnmh9ip6v2uc.cloudfront.net/datasheets/Prototyping/TP4056.pdf) barely mentions "single cell", but my examples show this is totally not the only configuration people go for with these modules (the answer by Sunnyskyguy also attests to this). – Mew Jun 27 '19 at 21:26
  • @winny: Done - apologies for not doing so in the first place. – Mew Jun 27 '19 at 21:26
  • That’s not a schematic but rather a wiring diagram. Non the less, it should work. – winny Jun 28 '19 at 10:51

1 Answers1

-1

There is no conflicts with any examples.

  • parallel operation is to boost currents by sharing
  • series to charge 1 card to 1 battery from a higher voltage.
  • arrays to combine both above

  • You can use as many cards in parallel as long as do not exceed the source current limit.

  • Each card has a Rprog to set say for 0.5A or 1A and USB charger hubs often support 2.4A per port such as 6 ganged ports with 60W max which are low cost.(eg Blackweb)

  • However charging and discharging simultaneously requires independent Ibat and Ichg sensing so battery current can be cutoff while charging the load directly.

  • this is to prevent cooking the battery at 4.2V CHRG drives the load draws more the 5%CC used for battery cutoff.

  • The TP4056 uses a high side FET switch to regulate battery current

  • the DW01 balancer disconnects charge discharge FETs on the low side

  • so you need to consider how to regulate both output ports ( Bat. and load) independently and have flow control from both voltage sources. (Charger and Bat.)

So draw a block diagram / wiring schematic and define all interfaces specs for voltage and current and see what needs to be done with parallel load and battery charge to prevent the above failure when current sharing.

I.load + I.bat < I.charge.max.
If > > I.charge.max, reduce I.bat then if still > Icharge, disconnect load

Consider;

  • priority is given to external load unless faulted or overload on hard start.
  • charge service to battery is 2nd priority.
  • Gang as many batteries in parallel as long as you know charger capacity
  • 5V Molex has more current capacity than USB ports
  • loads are responsible for protecting / regulating themselves
  • supplies are responsible for protecting themselves
  • chargers are assumed dumb otherwise and just responsible for regulating voltage @ some I max.
  • ICL’s (NTC) at Digikey may be used as a soft start
  • PTC’s may be used a fuses
  • limit the time spent with 4.2V on Battery to reduce aging is another option for cutoff with some timer in minutes
Tony Stewart EE75
  • 1
  • 3
  • 54
  • 182
  • Why was this downvoted? – DerStrom8 Jun 28 '19 at 10:25
  • @DerStrom8 thanks for your vote. The answer to your question is moderators and frequent contributors here can’t be bothered to moderate this frequent activity or say they are unable to do anything ( no comment) – Tony Stewart EE75 Jun 28 '19 at 10:53
  • I don't really understand what you mean when you say "parallel operation", "cards", "arrays" and the like. Do you mean having TPs in parallel, or packing the batteries in parallel, for example? Furthermore, as stated, the situation has a strict separation between charging and discharging, so the interspersed advice about this makes the answer more confusing to read; I can't really find an answer to my original questions here. – Mew Jun 28 '19 at 12:12
  • MEW I don't know how it could be clearer. Your 3 links are all valid and not conflicting, The differences are due to the series and parallel strings of batteries. WHich part do you need more clarification? – Tony Stewart EE75 Jun 28 '19 at 12:14
  • Well, for one, what do you mean by "operation", "card" and "array"? They don't mean anything to me, and could be about batteries or TP modules. Thus, I am still wondering, for example, why I might benefit from having one TP per battery and having the TPs in parallel, as opposed to having one TP for a parallel pack of three batteries - different kind of parallel, as it seems to me. – Mew Jun 28 '19 at 13:50
  • “Parallel operation” of 3 cards to boost 1 battery current, a “card” like a playing card is the PCB with the IC on it., “Arrays” means P x S matrix of par. X ser. Like 3S3P is 3 in series with 3 of those strings in parallel for 3 times the voltage and 3 times the current with 3 batteries and 9 cards. It for LEDs 3S3P means LEDs – Tony Stewart EE75 Jun 28 '19 at 13:56
  • Ah, that makes the post clearer. So, as a final question for clarification - since I was discouraged from hooking multiple TPs to a single input in the past, I really want to make sure I've got this right: I can make whatever array of TPs I want, without them creating damaging currents from one to the other, aiming to achieve the effects you've described? – Mew Jun 29 '19 at 20:25
  • In short yes, but PTC or some inbuilt fusing ought to be considered carefully (no pun intended) but also consider a voltage balancer to extend life of batteries. So the weakest cell does not cutoff the string. – Tony Stewart EE75 Jun 29 '19 at 20:33
  • you may also want to consider a tiny 2” fan for keeping cool – Tony Stewart EE75 Jun 29 '19 at 20:50
  • -1 low BW user.. – Tony Stewart EE75 Jun 29 '19 at 23:56