/       /       /    VariCore 32700 LiFePO4 Batteries and a Few Thoughts on Replacing Lead-Acid Batteries in UPS Units
Support the project on Patreon


VariCore 32700 LiFePO4 Batteries and a Few Thoughts on Replacing Lead-Acid Batteries in UPS Units

$30.24 (10шт)
Go to store
I had an idea for modifying various uninterruptible power supplies—more specifically, upgrading them to use more modern batteries. To try it out, I ordered a number of batteries, and today’s post will be not just a test but also a little investigation, if you can call it that.


Today’s review is split into two parts. The first is a conventional battery review, while the second is a small investigation directly related to using lithium batteries in buffer mode.


Part One

At first I wanted to order just a couple to try out. Then I thought, what would a couple actually tell me? So I ended up buying ten. There were some discounts on AliExpress when I placed the order, so they came to $30; the price is now $2 higher.

They arrived fairly quickly, but I was surprised to find them packed in a box with a LiitoKala logo. Still, everything was packed very neatly, with each battery in its own bag. The package weighed almost 1.5 kg.


The product title says the capacity is 6500 mAh, but the datasheet on the same page lists much more modest specs: at least 5500 mAh at a 1C discharge rate.
It also gives the internal resistance (no more than 8 mΩ) and charging parameters. However, these are all specified relative to capacity C, which is a little unclear, since capacity is normally specified at a 0.2C discharge rate, not 1C as stated here.


The datasheet continues, but some of the information is repeated.


So, I have ten cells with a pretty nice appearance. I don’t have a spot welder, so I ordered them with tabs already welded on. In fact, I couldn’t find them without tabs in the store.


The batteries themselves have a sticker for checking authenticity, and the manufacturer is listed as VariCore. All this caused a bit of confusion, so some of my graphs say LiitoKala and others VariCore. Don’t worry about it.
There are also markings like H5, H7, etc. on the negative terminal. I didn’t notice any correlation with the measured battery parameters.
The positive terminal—or, more precisely, what’s underneath it—has an unusual shape, with some kind of protrusions on the metal insert.


The dimensions are slightly larger than specified, but that’s not critical; we’ll just take it as it is.
The batteries weigh about 140 g, which is quite something compared with the familiar 18650 cells, especially when you pick one up.


At first, out of habit, I wanted to charge them in my usual charger. I even started thinking about how to fit them in, since it supports cells up to 26650 or 21700. But then I realized in time that these batteries have a different voltage, so I had to run the first “conditioning” cycle in the same charger I used for the other tests.
The batteries arrived almost completely discharged. On the first charge, about 5700 mAh went in.


To get ahead of myself, all the batteries were also discharged almost completely. The photo shows the first charge results for eight more batteries; I had to stop one of the charging cycles, so that battery isn’t shown.


And to get ahead of myself again, I’m not just going to show the capacity measurement from the first cycle, but from two cycles—the first and second, or zero and one, depending on how you count. I’ll also address a familiar claim I often come across: that lithium batteries need to be “conditioned” too. They don’t. Below are two graphs, for the first and second cycles. The first is blue and the second is red. Take a look and see if you can spot the difference—and which way it goes.



The second battery. I think the conclusion is obvious.
There is a notion that a battery is in some kind of “storage mode” and needs to be run through several cycles afterward. But that doesn’t apply to cylindrical batteries, and the difference itself is so small that it’s within the measurement error of the instruments. So I’ll say it again: lithium batteries don’t need “conditioning,” and they start gradually losing capacity after just a few cycles.


Now, about charging.
The datasheet specifies two charging options: at 1C (1 hour) and 6C (30 minutes). I tested them a little differently. First, I tried charging at 3 A, since I took 1C to mean 6 Ah—roughly halfway between the stated 5.5 and 6.5 Ah.
Charging at that current took 2 hours and 18 minutes, which isn’t a particularly impressive result—more like average.


The second attempt was at a charging current of 5 A. Unfortunately, that’s the maximum charging current my load can handle; the one that can charge at 35 A is still on its way. Yes, according to the datasheets, these batteries can be charged at around 30–35 A!

At this rate, charging takes 1 hour 18 minutes. All the following tests were done at a charging current of 4 A, simply to avoid overloading the charger.


The characteristic discharge curve typical of batteries like these is immediately apparent. Ordinary lithium batteries have a curve that declines fairly smoothly throughout the discharge, while the discharge curve of LiFePO4 batteries is more like that of nickel-based batteries: the voltage stays almost constant for most of the discharge, then drops sharply at the end.

Interestingly, capacity was over 5800 mAh at just 0.2C. In all the other modes, it barely changed at all, staying above 5700 mAh.


Heating was also relatively modest: even at 20 A, the temperature reached only about 67 degrees.


The second battery performed a little worse, though I wouldn't say by much.


Its temperature was also slightly lower.


To make the differences between the discharge curves easier to see, I put two types of batteries on one graph: a “classic” lithium-ion battery (red) and a LiFePO4 battery (blue). You can ignore the voltage difference; that's obvious enough. The difference in the discharge characteristics themselves is much more interesting, and easier to see this way. For comparison, I used a 5 Ah battery, as the closest match.
The LiFePO4 discharge curve is interesting because in this mode the energy delivered in Wh increases. The downside is that it's much harder to estimate the state of charge, so it's better to use a meter that tracks the capacity delivered. Those are much more complicated and expensive than a simple voltmeter, though.


To continue the test, as usual, I used a more powerful electronic load rated for up to 30 A. I've already described the test procedure; in short, we load the battery, measure its voltage under load, then use the difference to set the cutoff voltage.
The voltage also rose as the battery warmed up during the first half of the test, as shown in the two bottom photos.


At the discharge current, the capacity came to about 5750 mAh. The temperatures are shown in the photo on the right, and I can say I was very pleased with the results. The battery voltage did sag quite a bit at this current, though, down to 2.8 V, which isn't very good.


For the second battery test, I set the tester to logging mode so you can also see the discharge curve at 30 A.



Part two.
I'll share my conclusions at the end of the review. For now, I'll tell you about a small experiment.

A typical use case for batteries like these involves cycling: charge them fully, disconnect the charger, use them or put them aside, then charge them again when they're discharged.
Using them in an uninterruptible power supply is quite different, at least with less “advanced” models that can perform short discharge/charge cycles. The difference is that the batteries in a UPS are always under voltage—in other words, they operate in standby mode.

Why is this mode bad for lithium batteries? At high voltage, dendrites can form. In normal use, this will most likely just increase the battery's self-discharge, but if it's constantly connected to a charger, it could also heat up and develop an internal short, with the usual “special effects” to follow. Of course, I'm not a chemist, so this is just how the situation looks to me as an “ordinary user.” Unfortunately, the people who really understand all this chemistry aren't interested in doing this kind of thing, so this is the best I can offer for now.

Anyway, I “have a goal and believe in myself,” so I thought: if keeping batteries under voltage all the time is harmful, and the risk depends on the voltage level, why not try running them at lower voltages? Especially since:
1. Unlike lithium-ion batteries, LiFePO4 batteries drop sharply from 3.65 V to 3.35–3.45 V as soon as charging ends and the charging current is removed.
2. I have an idea to use 4S packs to replace UPS batteries. The standby voltage there is around 13.75–13.85 V, or 3.45 V per cell in a 4S pack.

1. First, I connected a discharged battery to a power supply and charged it until the charging current had dropped almost to zero. This simulated the transition to standby mode, when the charging current equals the battery's self-discharge current.
2. First, I tried a CV voltage of 3.55 V. The battery charged fairly quickly, taking only about a couple of hours at 4 A.
3. Then I discharged the battery and charged it to 3.45 V, or 13.80 V for the pack. This took longer, around 4–5 hours.
4. I got carried away and tried charging to 3.35 V, but since I was using a two-wire connection, the test started to look like it would go on forever. So after about 10 hours I raised the voltage to 13.8 V. The total charging time was around 15 hours.

The power supply display shows the capacity in mAh. Because the test took so long, the measurement error accumulated, so don't pay too much attention to these values.


I discharged the battery at 10 A, which is the most realistic current for this scenario.

Two interesting things emerged:
1. As the CV-phase voltage is lowered, charging time increases noticeably
2. The capacity measured during discharge is almost independent of the voltage at which the charger switches to the CV phase


I mentioned the two-wire battery connection above for a reason. The second test used a charger with a four-wire connection, which significantly reduced charging time. The charging current was 4 A, with charging stopping when the current dropped to 100 mA.
But the results were even better: in practice, the capacity I got was independent of the charge termination voltage.


Honestly, I thought that lowering the charge termination voltage would reduce not only the capacity but also the charging time. In reality, the opposite happened: capacity stayed the same, while charging time increased—and quite noticeably.
The lower the charge termination voltage, the sooner the charger switched to the CV phase, and the longer charging took. Below are examples of charging at 4 A with voltages of 3.65, 3.55, 3.45, and 3.40 V.


And, as always, I measured the internal resistance. The results were frankly disappointing, since some batteries were clearly above the specified 8 mΩ.
Below are the results before testing, after the first cycle, and after all the tests. In preparing this review, I ended up running more than 10 full cycles at different discharge currents.


Here are the results for the other eight batteries. The photos are grouped in pairs: first as received, then after a full charge. At first I thought the resistance was high because the batteries were completely discharged, but it turned out to be slightly higher after a full charge.
Now I'm wondering whether to open a dispute and ask for a partial refund for the batteries with excessive resistance. They did promise less than 8 mΩ, but in reality only half of them meet that spec, and several are at 10–12 mΩ.


In terms of size, LiFePO4 batteries are also more appealing: eight 32700 cells plus a protection circuit fit in the space of a standard 12 V 7 Ah battery. They are slightly wider, though, since the 12 V battery is also 64 mm (2×32), and lithium batteries need a casing. But there are ways to offset the cells slightly relative to one another.

With this new configuration, we get a real 11–11.5 Ah and can handle continuous load currents of up to 60 A. For conventional batteries, the maximum capacity in this size is 9 Ah. In fact, just today I was preparing six CSB batteries for a UPS. They aren't cheap either, at around 600 UAH (25 dollars) each.


And now for the conclusions.
First, about the batteries themselves.
Overall, I liked them. They deliver a decent real-world capacity of 5.7–5.8 Ah and don't heat up much, even at 20 A. However, their internal resistance varies a lot, and of course they don't deliver the 6.5 Ah claimed in the listing title. Also, at discharge currents above 15 A, the voltage sags noticeably, which reduces the energy delivered.
I'm not even sure whether to open a dispute. I don't like the wide variation in internal resistance, and half the batteries don't meet the stated specs. Also, a CSB 12 V 9 Ah battery has an internal resistance of 16 mΩ (I measured it today), while my 4S1P pack is at 32–44 mΩ, which is rather high. Ideally, I should build a 4S2P pack.

Now, about the experiment.
I'm more than happy with the results. These batteries can be used in standby mode at 3.4–3.45 V per cell, or 13.6–13.8 V for the pack, so there's no need to change the charging settings on UPS power supplies.
Naturally, my experimental conclusions apply only to these particular batteries. Batteries of other types—especially those with different chemistry—will behave quite differently.

Would I recommend these batteries? Definitely, with the caveats mentioned above about consistency and capacity. In other words, you need to know what you'll be using them for and whether those issues will matter. If you're not sure, ask me and I'll advise you.

I've already received a protection board for four cells and plan to write a follow-up where I'll try installing this pack in a regular UPS. For now, I regret not buying 20 cells, because 10 doesn't divide evenly by 4. Also, the second set of four batteries will go into a second version of the UPS—a low-voltage one for a completely different project, which you’ll hear about too, but much later.

By the way, there’s a sale on the same batteries, but with Litokala labels, for 28.56 per 10-pack. That’s even cheaper than what I paid, so I’m even thinking of picking up a pack of ten.

That’s about it from me. As usual, I’ll be glad to hear your questions and suggestions for tests.
$30.24 (10шт)
Go to store
This page was found when searching for:
varicore vc4850, обзор аккумуляторов varicore, в корпус от 7 амперной батареи ups сколько входит 32700 элементов?, varicore vc-4850 pdf, аккумулятор 32700 lifepo4 зарядное, отзывы о продукции фирмы varicore, акб varicore vc-12120 отзывы, аккумулятор lifepo4 12v 40ah varicore инструкция по эксплуатации на русском языке, как правильно собрать аккумулятор из 32700 58 вольт, какое зарядное устройство нужно для аккумулятора 3s2p lifepo4 32700, lifepo4 вместо свинца в упс, kirich дендриты, вскрыть акб lifepo ибп форум, схема подключения аккумуляторов 32700, батарея из 15 штук 32700, схема сборки аккумуляторных ячейки fe, 2, lifepo4 акб на 24в 25 амперчас заряжается только до 27вольт-что делать, пустая коробка для сборки аккумуляторов 32700, kirich lifepo4 ups, переделка упс lifepo4, сборка 32700 с винтами, замена аккумулятора в бесперебойнике на lifepo4, varicore 33140 как заряжать, 32700 аккумулятор

Questions and answers

What capacity did the VariCore 32700 LiFePO4 cells deliver?

Measured capacity exceeded 5,700 mAh in most modes and was over 5,800 mAh at a 0.2C discharge rate. The datasheet specifies at least 5,500 mAh at 1C.

Can VariCore 32700 LiFePO4 cells be used in a UPS standby setup?

The author’s tests suggest they can, at 3.4–3.45 V per cell, or 13.6–13.8 V for a 4S pack. This conclusion applies only to the cells tested.

What did the VariCore 32700 internal-resistance tests show?

The cells showed a considerable spread, and about half did not meet the stated resistance specification. The author measured 32–44 mΩ for a 4S1P pack and says a 4S2P pack would be preferable.


This model is included in the summary table: Battery comparison: measured capacity and resistance

The figures from this review are in the measurement database: Batteries

Calculate manuallyAll calculators

Discounted products


You may be interested in


Discounted products


Support the project

All the content on the site and the advice are free, but we’d be grateful for your support of the project and the channel!

Comments10

Comments are published in their original language.

  1. 30 January 2020 00:59#1
    Реально полезная тема, потому как часто нормальные свинцовые батареи стоят дорого, а дешевые для более-менее мощных УПСов не подходят, да и служат они мало.
    +4
  2. 8 March 2020 23:32#2
    Покупал эти аккумуляторы на алиэкспрессе для использования в ИБП для замены свинцовых.  Собрал 8s4p. использовал с балансирами на 1 ампер. Через полгода работы в буферном режиме умерли две ячейки в ноль пытался зарядить лабораторным блоком питания но они заряда больше не принимали. Выкинул их и пересобрал батарею на 8s3p, но через несколько дней умерли все. Потом выяснилось, что и ИБП перестал заряжать батарею.
    Теперь вот думаю - то ли купил подделку, то ли их буферный режим убил. ИБП в буферном режими выдавала примерно 3,43 вольта на банку. Смотрел сейчас спецификацию на леферы ЛИОТЕХ там указано, что для буферного режима напряжение должно быть (3,34÷3,35) В на аккумулятор.

    У меня в параллель лиферу титанатная сборка была подключена 12S так вот с  ними ничего не случилось они просто  разрядились, зарядил и померил емкость всё в порядке.  Ну с титанатом с их 2.4 номинала ничего случиться и не могло 12s, они  всегда в некотором недозаряде от 27,5 вольт свинцового буферного напряжения  на сборку.

    Как ваши VariCore поживают в буфере? Живые еще? 
    +2
    1. 12 March 2020 11:26#3
      Не, пока почти не использую, жду устройство, куда они будут ставиться.
      А так да, тоже почитал больше информации и узнал что лучше выставлять буферный резхим не 3.4, а 3.35 на ячейку. Там кстати говорилось что лучше сначала зарядить до 3.4, а только потом перевести в режим CV с 3.35. но думаю что вполне можно сразу заряжать до 3.35, но это сильно скажется на скорости заряда.
      0
  3. 2 December 2020 12:20#4
    Здравствуйте!
    Есть ИБП Cyberpower Value 2200. Решил поменять кислотные аккумуляторы на LiFePo4. Заказал. Пришли. 32700 7000 мАч. Тестирование показало, что это 6500 мАч, только перекрашенные в 7000 мАч. Мне в итоге нужны 2 аккумулятора по 12 В. В ИБП они подключены последовательно. С балансирами все понятно. Купил 4S, будет P3. Если правильно понял в каждой "S", в сумме, должна быть примерно одинаковая емкость. Последовательные "S" должны иметь одинаковое напряжение. Появился вопрос, а как балансировать 2 аккумулятора соединенные последовательно. В ИБП стоит простая перемычка между "-" и "+". Где-то видел ваше решение этой проблемы, но как еще раз не искал так и не нашел. Подскажите пожалуйста решение данной проблемы.
    0
    1. 10 December 2020 04:05#5
      Как раз разница емкости в S не критична, но вообще я как-то давал ссылку на программу, а точнее веб страницу, там можно задавать емкости ячеек и схему включения, она посчитает оптимальное расположение ячеек в схеме. Сейчас пытался найти ссылку, но что-то не находится, попробую завтра.

      +1
      1. 20 December 2020 14:31#6
        Здравствуйте!
        Вы обещали дать ссылку где можно посчитать оптимальное расположение ячеек в схеме.
        И оставался еще вопрос без ответа, что делать с перемычкой между аккумуляторами. Или на это можно не обращать внимания?
        0
        1. 20 December 2020 21:28#7
          Да, нашел, вот ссылка - https://www.repackr.com/#/pack-builder
          Думаю разберетесь, если нет, пишите.

          Насчет балансировки не совсем понял, там по сути батарея 24 вольт, значит ставим схему 8S, если для ИБП нужна средняя точка, берем её от сборки, если нет, ну и ладно.
          Балансировка возможна балансиром, обзор которых я недавно делал.
          +1
          1. 21 December 2020 09:14#8
            За ссылку спасибо, разбирусь с аккумуляторами и буду компоновать банки, вот тут она и пригодится.
            На счет BMS 8S, дело в том, что я уже купил 2 белые платы BMS на 4S 100А. Буду делать 4S3P. Места в ИБП CyberPower в обрез, т.е. в бункер влазит ровно 2 стандартных аккумулятора 12V 9A и больше свободного места нет. Как уместить LiFePo4 32700 уже придумал, буду ставить в два этажа. Сейчас установлены вертикально (длинной стороной вверх) и параллельно друг другу. В итоге появился вопрос: нужно ли балансировать два собранных аккумулятора или и так сойдет? Вроде бы ка балансиры будут сами делать нужное напряжение на выходе собранного аккумулятора. аккумулятора.
            0
          2. 23 December 2020 13:45#9
            Здравствуйте!
            Так и не нашел в интернете как правильно заряжать 32700 LiFePo4 7000мАч с нуля, т.е. с момента получения данного аккумулятора (datasheet не всегда прав). Опыта нет. Единственно, что получилось так это поднять емкость аккумулятора после не правильной зарядки, такое тоже бывает. Вследствие чего есть вопрос:
            1. Если аккумулятор полностью заряжен, по прошедствии 3 дней или дольше садится емкость аккумулятора примерно на 30 мАч и напряжение проседает примерно до 3.36 В. При тестировании повторного  разряда аккумулятора, нужно ли сначала дозаряжать до 3.65В и затем минут через 10 включать разряд или сразу включать разряд (EBC-A10H)? Как будет правильнее?
            Еще хотел поделиться денежкой, но могу скинуть только на карту Сбербанка или PayPal, если такая возможность есть скинте мне на e-mail куда перевести. Главное, чтобы с меня проценты за перевод не брали...
            0
            1. 25 December 2020 04:43#10
              Как заряжать, без разницы, обычно достаточно просто полностью зарядить до 3.6-3.65 вольта с отсечкой по падению тока до 100мА.
              Можно сначала разрядить до 2.5 вольтта и потом зарядить как я написал выше.

              То что напряжение падает через время это нормально для LiFePO4, оно падает почти сразу после снятия зарядного тока.
              Про повторное тестирование не совсем понял, если долго лежал, можно опять зарядить, выждать паузу и потом разрядить с измерением емкости.
              0

Discuss on the forumHave a question or some experience to share? Discuss the article with the author and readers: the forum is a convenient place to share schematics and photos.
Go to the forum

Only registered readers can comment. It takes a minute to sign up, and you can log in with the same account you use on the forum.

Log inRegister