New lipo charger designs

[David T]

Hi folk. I have been using L200-based chargers for some time. However, I wanted to be able to vary current with a standard pot. The datasheet has a circuit for this using an LM741. A few tweaks have improved the circuit and I've built five in a 'milking station' configuration. Current is variable smoothly up to 2A with adequate cooling.

Linear regulators like the L200 and LM317 require an input voltage that is greater than the output. To charge three lipos requires an input of 15v or more. Making a switch-mode power supply to operate from 12v can be a challenge. So I have been experimenting with a 'universal car laptop adapter'. Units that can deliver 5A at 15v are available on ebay for under £8. Mine operates at 87% efficiency which seems pretty good. I have it powering my milking station and it has been working well for a couple of weeks.

I believe this adaptor could also be modified to make efficient chargers for 4 or more lipos. For instance, the device should yield about 3A at 25.2v for 6 lipos with minimal mods. Watch this space!

I've posted a couple of photos here. The circuits with a more detailed description are on my site: http://www.flyelectric.ukgateway.net/lithium-charger.htm

David.

[JRuggiero]

David T,

Nice work. Having worked with the L200 for two homebrew chargers, I'm looking forward to your design, because I also don't care to adjust current with hefty low-resistance pots or switched resistors.

Jim R

[David T]

Thanks. It's working well and the pot adjustment is very nice.

With more use I am finding that the L200 does not like being 'instructed' (by the op amp) to go beyond its 2A max. I built a fixed resistor (R7) in for this adjustment but in practice each circuit seems to need a different value. I'm updating the design to use a trim pot instead. This issue only manifests itself at the max current setting.
David.

[JRuggiero]

David T,

Another thing I like about your circuit is that it uses the good ol' analog LM741. We don' need no steenkin' PIC!

Jim R

[David T]

The variable L200 design is neat. It only needs 15V in for 3 lipos even with a schottky diode reverse polarity protection on the output so is much more efficient than the dual LM317 approach which needs 3V more (ie: 6v more to boost and shed as heat).

I've been experimenting with the laptop PSU. I've bypassed the switch with a trim pot to set the voltage for direct lipo charging. It needs 25.4v with no load to deliver 25.2v under load. I need to see at what current it drifts up and scope the output. I'm trying to develop a 6 cell high current charger and it seems suitable for 4 to 6 lipos, maybe more. It puts out 82W as per the spec (~3A) at 89% efficiency which is pleasing.

I'm also experimenting with different current sense resistor values (marked 'R1' on their board). The standard R1 is about 0.030 ohm. With ~0.018 I'm getting 4A out at about 85% efficiency (100W). However the inductor heats up quickly even at 3A and obviously more at 4A. A higher R1 value reduces charge current. I've tried to create a switchable current with taps on R1 (to the IC) but the circuit is not suitable for this. It's looking like it only suits fixed charge currents. Does anyone have bright ideas for switching/changing such a low value resistor without spending a fortune? R1 carries high currents (about 5A when the output is 4A).
David.

[GlowFly]

Does anyone have bright ideas for switching/changing such a low value resistor without spending a fortune? R1 carries high currents (about 5A when the output is 4A).
Leave the low value resistor fixed, and make a tap of two higher value resistors to divide down/reduce the current feedback. Reducing the feedback will increase the output current, so make the fixed resistor value suitable for the lowest current. But make sure the fixed resistor power rating can cope with the consequent highest dissipation.

--
Steve

[AndyOne]

For comprehensive L200 data see...

bwir.de/downloads/L200_englisch.pdf

Andy.

[David T]

Leave the low value resistor fixed, and make a tap of two higher value resistors to divide down/reduce the current feedback.
--
Steve

That sounds a good idea. I need to think about how I do this but I can see the potential.
Thanks, David.

[killer_d76]

David,
pretty cool website!.. ;), i plan to build your Lithium balancer, looks pretty simple and i got lots of spare LM358N, though L272 is very rare here so i plan to use TDA2822 (http://www.utc-ic.com/spec/TDA2822.pdf) instead would this be suitable as an alternative IC?.. and forgive my ignorance but should i use the balancer before charging or right after charging the batteries?.. thanks!

[coro]

.. and forgive my ignorance but should i use the balancer before charging or right after charging the batteries?.. thanks!
Balancing before charge, it means equalizing voltages in discharge state - that is wrong.
Balancing after charge, it is way too late for overcharged cells.

[David T]

The TDA2822 has the same pin connections as the L272 and a slightly better output capability. However it is rated at 12v, 15v peak. Three lipos are 12.6v which I guess is OK but it certainly would not be suitable for the designs that do more than three cells. Let us know if it works OK. I suggest you measure individual cell voltages via the balancing plug occassionaly while balancing for the first time to ensure they are getting closer to each other and not worse.

I believe the low offset voltage of the LM358 makes it harder to substitute.

You can use this balancer at any time, ie: before, during or after charging, and with a charger connected or not. However, it will do most good when used during charging as this should prevent over-charge. Nevertheless, there is still some benefit in balancing at other times.
David.

[killer_d76]

Balancing before charge, it means equalizing voltages in discharge state - that is wrong.
Balancing after charge, it is way too late for overcharged cells.
:confused: :confused: :confused:

[killer_d76]

David ,
thanks for the quick reply!.. i'll try to build one using the TDA2822 since i'll be using it on 2cell li-ion only!, i have built electron's head charger ( http://www.rcgroups.com/forums/showthread.php?t=130567) too, so i'll be needing a balancer then i found your design that is quite simple! ;) thanks David!

[coro]

:confused: :confused: :confused:
Just wanted to explain why i disagree either, and I am sure that charger have to avoid any cell overvoltage (by reducing charge current, instead of wasting energy by balance bleeding). After the current approaches zero, it means when some of the cells are really full, it is possible that some other cell(s) have lover voltage and they need few more charged mAh to be full - so only after this conditions it is reasonable to bleed higher cells by balancer or bypass them and charge the lower cells. For sure, this situation is clear sign of pack damage in previous discharge, so the warning have to be issued to user.
It is reasonable and clear, imo, isnt it?

[David T]

Coro, You mention 'wasting energy' and 'balance bleeding' so I think you are referring to the 'charge guard' type of circuit. These circuits bleed off excess charge when charging. They have to be able to handle the full charge current. The part that is 'bled' off is often wasted as heat.

The circuit on my web site works differently. It takes power from 'high' cells and transfers this to 'low' cells. Some of this is lost as heat, but not all. This process is not directly related to the charging process. It does this at a very low rate, typically only a few humdred mA. So I think it is of benefit to level the voltages at any time, empty or full, because it might take a long time to balance them (depending on how much they are out of balance).

I think you are also making the point that if lipos are becoming damaged, their voltages can drift at different rates, empty vs full. The circuit on my site will simply try to keep them in balance, whenever you balance them. If you watch 'my' balancer while a pack is being charged, if different lights will come on at different times during the charge this indicates that current is flowing to different cells at different rates and the cells are probably not perfectly matched. If this happens and the pack feels a bit 'flat' even when fully charged, it can be good to discharge each cell individually to a safe voltage (eg: 3v per cell) and compare how much was removed from each. This should indicate whether one cell is dying.

I think we agree that the best time to balance cells is when they reach 4.2v per cell during the charging process. The graphs on my web site show that this occurs at about 40% into the charge (ie: quite early). Because the balancer on my site takes a long time to balance, starting before or at the beginning of charge I think is most helpful.

On a more practical note, I think it is better to balance lipos occasional rather than never. So doing it every now and then (at any time as I suggested) I believe will be of benefit. I hope I have not misunderstood your points; please correct me if I have.
Regards, David.

[coro]

David,
All I wanted to say is that Lipo cell become damaged at higher than certain voltage - lets say and use 4.20V as a standard - and the allmost only necesary job for charger is keeping the voltage below. Not just sumary voltage over couple of cells, but the voltage of every single one. Look at my www for very simple circuit for fooling the charger.

It is not a problem if some of the cells remain at lower than maximal voltage during charge - they are charged slower than they could, but there is no reason to charge some of the cells faster than others and then wait.

Fastest balancing in meaning of equalising no-current voltage, is achieved by Suzanne like design, as simple keeping the voltages at equal level takes very long time until balance current is really zero and rest voltage remain equal.

Second point I wanted to pick is that any pack I did test (creature of very different cells in series as well), remain balanced if it was not overdischarged or damaged, so real balancing is very ocassional, not everyday practice (in case that my charger will not overvoltage some of the cells in a pack).

Balancing is used as a supplement to avoid overvoltage during charge now, but its current capability is laughable even it is 500mA, so I think that the time for better idea is here.
The name "NEW lipo charger designs" wakes me temporary, ..
http://mujweb.cz/www/coro/LimiterEN.htm http://mujweb.cz/www/coro/

[David T]

I believe I agree with your points. I'll take a look at your site in a few days I hope.
Thanks, David.

[killer_d76]

thanks guys anyway.. ;) i'm working on the balancer now.. but only to use it on when-needed-only circumstances, hope to hear more comments from you guys! ;)

[theburro]

Hi David,

I'm having a hard time getting my hands on a L272M, I'm wondering if you know of any other substitutes (other than the tda2822 which I also can't get). Looking forward to building the balancer. Lowell

Also I'm a little confused by a couple of the resistors. ON the schematic page it calls for a 27R 2W, where as on the BOM page it calls for a 47R 3W. Are these interchangeable or which one should I be using for R4 & R6??? :confused:

[David T]

I have not researched alternative components for the L272. I would imagine that any op amp that allows a supply voltage and output current greater than you need (say >500mA) would work. To avoid changing the veroboard or PCB layout, this would need to be an 8 pin DIL package that contains two op amps with the same pin configuration. I'm happy to express a view on an alternative if you find one.

An alternative is to follow the links on my site to the orginal circuits. Richard Szym's site shows designs with transistors in place of the L272. He also suggests an OPA544 op amp but the pin config is different.

Sorry for the resistor confusion; I do discuss it briefly in the text. The 27 ohm resistor allows higher balancing currents but can get hot enough to melt the heat shrink covering that I use. The ICs can also get a bit hot. So I changed the design to 47 ohms to reduce balancing current and therefore heat. I actually buy 4W resistors now as the ones I can get are still compact and reasonably priced (the higher the wattage the better they are at dissipating heat).

[theburro]

Hey David,
I finished the balancer a few days ago and it works great. I had to go over the board a few times to make sure I had everything done. This was my first project, and it works great. Easy to make. Thanks for the help. Lowell