Charging high voltage lipos a different approach to cell balancing

Charging high voltage lipos: a different approach to cell balancing.

Faced with 3 11.1 V RC lipos at 2200 mAh to make 33.3 V for 'buffer' for cooker, and solar power to charge it , I was wondering... how to safely do cell balancing.. I have seen many schemes,

But then had this idea, I once bought 20 pieces of MCP73831T single cell lipo / liion charger chips for < 4$ on ebay, only used a few so far, and those have proven themselves with daily charge, discharge. So 3.7 V per cell out.

9 cells (3 x 3).

Those chips need 6 V or a bit lower in. say 5 V to reduce dissipation. Say 6 turns on a transformer.

9 windings of 6 turns, and the primary some push pull 12 turns on solar panel. Costs: only the resistors and diodes... some caps, all in the [no junk] boxes. Charge current say 400 mA. 5 to 6 hours charge time... for 2200 mAh. Small transformer, 15 VA (33.3 * .5), can wind it in an hour.

Anybody overlook something?

Sure 400 mA (500 is chip limit) is not that much, but solar for that it is just right, would be < 1.5 A from the solar panel.

mm Yu kan fissiolize ze zirkuit no?

If not ye'r puzzies :-) ;-) (-;

Reply to
Jan Panteltje
Loading thread data ...

Simple discrete solution: make a shunt regulator around a TLV431, boosted with a power transistor. Set it for 4.20V (give or take). Put one across each cell.

The ~100uA bias means you'll want to have them switchable, maybe with some PMOS and resistors and zeners. Only needed during charging, after all. This also means you can charge with a constant current (as long as that current is less than the total current/dissipation capacity of the shunt chain).

Current mirror output from each shunt would be cool, so you can monitor as each cell tapers off during its constant voltage phase.

A more complicated version might use a boost converter per shunt, to deliver the excess power back to the cells above and below it, or directly back into the charger supply.

Tim

--
Seven Transistor Labs 
Electrical Engineering Consultation 
 Click to see the full signature
Reply to
Tim Williams

I don't understand what you're doing. A block diagram or schematic please?

Before proceding, you might want to check if you haven't re-invented the wheel. Google for BMS (Battery Management System) for LiFePO4 batteries. These are shunt regulators that you attach to LiFePO4 cells to keep them balanced. They're common on eBikes and EV systems and are powered by the individual cells. For example: Some are combined with battery protection circuitry. For example: The bad news is that you're going to dissipate some power in the shunt regulators (because that's what shunt regulators do). Details:

--
Jeff Liebermann     jeffl@cruzio.com 
150 Felker St #D    http://www.LearnByDestroying.com 
 Click to see the full signature
Reply to
Jeff Liebermann

On a sunny day (Sun, 20 Dec 2015 17:19:51 -0600) it happened "Tim Williams" wrote in :

Yes, that is one way, I still have a whole bunch of LM317, maybe I can use those, will fiddle with some circuit ideas. That indeed gives a bit more charge current, and less losses.

mm, yes, that is also an issue, the current draw when not charging. That MCP73831T has reverse current protection.

Yes.

:-) The ideal system takes the output voltage, boosts it, and feeds it back to the input so always free energy :-)

Reply to
Jan Panteltje

On a sunny day (Sun, 20 Dec 2015 23:16:32 -0800) it happened Jeff Liebermann wrote in :

Its very easy, sorry for the complicated story, it is simple 9 MCP chargers, one on each cell, each galvanic insulated with a RF transformer.

Yes, but I have 9 cells...

That finds among other things this circuit:

formatting link
I think that is also what Tim mentioned, but now wit ha MX chip.

Yes, need to search on ebay too, but fro the MCP I have all parts, for LM317 too.

Yes.

Thanks for the links.

Reply to
Jan Panteltje

On a sunny day (Sun, 20 Dec 2015 17:19:51 -0600) it happened "Tim Williams" wrote in :

PS, I could not resist, 50 for 1$84 free shipping:

formatting link

Now to find some PNPs.. Preferably TO220 or something so they can dissipate a bit.

Your idea has the advantage that Ican adapt it for lifepo4 or liion or lipo with just a resistor.

Reply to
Jan Panteltje

Ah, good ol' Wing Shing brand. They have some good prices on the ancient jellybean parts.

I suspect they're not really a manufacturer, but they get their brand put on stuff.

--sp

--
Best regards,  
Spehro Pefhany 
 Click to see the full signature
Reply to
Spehro Pefhany

Will this suffice?

5 to 13 cells at 5A max.

I couldn't find anything specifically for 9 cells, but maybe you can make a 10 cell balance system work: Careful. The title says LiFePo4 while the specs say LiCoO2.

Transformer isolated individual charging circuits seem like a good idea. Perhaps make an independent module consisting of the cell, the charger, the protection, etc. That way, when it explodes or catches fire, it's obvious which charging section is at fault. Hint... don't forget the fuses.

Best of luck on the project.

--
Jeff Liebermann     jeffl@cruzio.com 
150 Felker St #D    http://www.LearnByDestroying.com 
 Click to see the full signature
Reply to
Jeff Liebermann

On a sunny day (Mon, 21 Dec 2015 08:48:49 -0800) it happened Jeff Liebermann wrote in :

It says: Please be noted: you cannot program the BMS without the USB adapter which cost $199USD additionally. The USB adapter is free of charge if you buy 1000pcs smart BMS. ???

Yes, seems you can just connect the cathode to one or mroe 'B' taps lower.

Yes, 4.2 V looks like lipo.

Thanks, I ordered some TL431 and some 2N2955 PNP (TO220) to try it the way Tim described.

formatting link
formatting link
Still less than 5 $...

Sure many of you remember the 2N3055, the 2N2955 is the complement half. Could be fake Motorola, but as long as it works.

A big 24 V 15 Ah lifepo4 is under consideration.

Just did run some test with solar panel and up-converter to 24 V.

Reply to
Jan Panteltje

Den mandag den 21. december 2015 kl. 18.36.42 UTC+1 skrev Jan Panteltje:

formatting link

-Lasse

Reply to
Lasse Langwadt Christensen

On a sunny day (Mon, 21 Dec 2015 09:54:50 -0800 (PST)) it happened Lasse Langwadt Christensen wrote in :

Yes, bit puzzling, I want to use only the PNP...

39 Ohm is about 100 mA.. at 4.2 V, (.2V Vce drop). According to my neural net for a beta of 100 (typical listed) of the 2955, and the typical 100 mA for the TLwhatsit, makes 10 A? Well, but I am neural net. What bugs me now is I need 9 TO220 heatsinks. Or one big alu plate and 9 TO220 isolation sets. Big board, or make 9 small boards. or 6 or 7 (if I was to go for 24 V).

Charge current is limited by solar cell, say to 4 A at most after step up with this:

formatting link
That converter I just tested on the solar panel (not much solar today, and shortest day).

4 A at 4.2 V is about 16 W (worst case per transistor), so need to reduce charge current. Now I was thinking (neural net again) because I also have these:
formatting link
to [also] put that one in series, the current limit works great for lead acid charging, so should work here too.

Now ain't that fun :-)

And discharge protection, we need that too.

So, I was already contemplating using a PIC ADC with voltage dividers from each cell, adding a Microchip ENC28J60 for Ethernet, and sending status via UDP every few seconds (and control it from somewhere 'on planet earth or further once NASA opens deep space network to humans). Already wrote all the code for that chip for other projects (like LED lighting). And may as well add serial NMEA.

So, :-)

Reply to
Jan Panteltje

Eww-w-w-w! An, I can only imagine purposely designed, smoke factory!

  1. Divider is too small for battery use, I'd rather see like.. several 100k, or 1M.
  2. TLV431 is lower bias and just as cheap (but at least R6 is correct).
  3. TL431 can sink 100mA into the BD140, which is as vague a transistor as you can get, but it's going to be over 1A, maybe 2A even.
  4. D better be a pretty big LED.
  5. Q2 can't even draw 2A, let alone carry it in its base. Kaboom.
  6. 39 ohm implies this thing was only ever intended for ~microamperes at the TL431 (that is, in excess of its bias current), and for 100mA total sinking current. Which the TL431 can do alone (if maybe not thermally), and the BD140 can do handily.
  7. But because Q2 is allowed to saturate (the load resistor helps cheapen power dissipation, but actually not in this case), if the charging current exceeds 100mA, the whole thing goes cluster real fast.
  8. And incidentally, the first LM317 is set for 1.25A, so yeah.

There needs to be an additional series resistor between TL(V)431 "cathode" and Q1 base, to limit base current (and '431 dissipation) responsibly. An emitter resistor for Q1 might also be helpful.

A slightly larger Q1, maybe a TIP32, would be fine for present purposes. Delete all the D-Q2-39 ohm stuff.

Tim

--
Seven Transistor Labs 
Electrical Engineering Consultation 
 Click to see the full signature
Reply to
Tim Williams

it can, but will it, the feedback through TL431 should prevent that.

sure, there's no protection against connecting a cell that is overcharged, something would probably explode if that happens. So don't do that.

+1

I don't see that. R1 appears to be designed to destroy the trimmer.

looks more like 400mA to me, go much past that and the trimmer burns up

but I'd substitute in one of theose cheap 5A CV-CC buck converter boards for U1-U2

feedback should limit that.

--
  \_(?)_
Reply to
Jasen Betts

Also, the LEDs could usefully be the input side of an optiosolator, so that each optoisolator turns on when that cell is charged to full voltage. The output sides of the opto-isolators could be all in series, connected to some shutdown circuit, so that when all cells are charged it will turn off the input for a while. Still the scheme has enough problems that starting from scratch might be better.

Reply to
Chris Jones

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.