After reviewing most of XTAR’s 1.5V Li-Ion AA offerings, I’ve had a few people ask me whether any of the cheaper alternatives are any good, with the most frequent request being to test the imuto TF0CD0998KDC 8 x AA battery kit from Amazon. Usually, I wouldn’t pay much attention to such a thing, but owing to a promotional discount, I decided to plonk down some of my own money to find out, buying two sets so I could take one apart. I hope you enjoy …

Unboxing

The unit arrived in a thin colour print cardboard box, The set is called the “AA Station” with an AA Charger Box and eight 3600mWh 1.5V Lithium-ion cells.

The back of the box indicates it to be model TF-CD-009LDC which takes a 5V 3A input and provides eight 350mA charging channels, leaving 200mA for the circuitry itself. Apparently, a lower specification adapter can be used too. The unit carries a number of approvals, but there is no Australian RCM. It is Made in China.

The box has a gold, white and black colour scheme. The top has the brand, while the sides have a mixture of features and regulatory information. The barcode label is an Amazon internal label.

Opening the lid, we can see the use of a milky white plastic bag to protect the AA station.

Everything in the box is shown above, including a customer care card, an instruction leaflet, a USB-C to USB-C charging cable (required for the 3A current) and the AA station itself. The batteries come pre-packed inside the case.

The case has a plastic body with a matte finish, black in colour, with the imuto logo on the front. The hing can be seen at the back, which is moderately sturdy.

The USB-C power enters from the left side top, while other surfaces are generally featureless. The case is a rounded rectangular prism in shape.

Swinging up the lid, we can see some round spring-loaded ball bearing style contacts at the top for all positive connections and a common negative. The lid is held closed with the use of a pair of strong magnets.

A similar arrangement is used for the negative contacts.

The cells themselves have the same colour scheme, but are mostly white. The cells claim to be Model AA1100, 1INR15/51, with an energy of 3600mWh and capacity of 2400mAh (although in a very small font, that is easy to miss). It claims 1.5V output, but if this was the case, then the converter has to be 100% efficient or one of these figures isn’t what you might think it is. We’ll find out about that later.

The cells have a rather small nipple diameter for the positive terminal, flanked by a white plastic insulator cover. The negative has a mostly bright finish, subtly raised.

The details of the charger are printed on its underside.

The included black USB-C to USB-C charging cable, although you’ll have to provide your own power adapter.

Who Makes These?

If you’ve been in the market for 1.5V AA Li-Ion batteries, you might have noticed something very strange – namely that this design of battery (with its narrow nipple) and eight slot charger is seemingly everywhere under almost every brand under the sun (albeit sometimes with lower capacity cells, or in a four-cell charger variant or with different colours).

While I purchased it under the imuto brand, while writing this post and browsing through the normal online marketplaces, I found it under the brands of Vennerli, Pullomi, Sanpima, Dracutum, Artman, Homesuit, OXWINOU/QBLPower/GeiliEnergy, Pujimax, Nohon, Sidande, Chicnod, Sinceholy, Newell, Batzone and Delgeo. The list is hardly exhaustive at all, but I had excluded some lookalikes that had ribbing, curved edges or a power inlet port at the base as those may be different designs from a different company. China does make a lot of “copycat” products.

But who is the true originator of these products? I had to do some digging but I believe it is Keeppower Technology Co. Ltd of Guangdong, China. Based on the specifications, this product is their i8 with their P1450I cells. Other variants including cells up to 4700mWh (P1450I3) are available, as is their 4-slot i4. This can be found on their new products page. I’ve previously reviewed their 26650 cells and found them decent, so it is nice to know a trustworthy brand is behind the “wall” of brands on Amazon.

User Experience

Having been used to the XTAR batteries, I wasn’t expecting the best experience moving to these lower-cost alternatives but I was somewhat pleasantly surprised. It would seem most of my devices worked fine with the cells (but then again, most of my devices work fine with Ni-MH as well). One notable downside seemed to be the “narrow” nipple at the top of the cell, meaning that devices that stack cells in their battery compartment were more likely to have the cells spring out from the compartment when bumped or during loading. Compared to the XTAR, the lack of LED indicator was not really noticeable in most cases (as often batteries are in fully-enclosed opaque compartments) and there was no sudden unexpected power-downs but these cells did seem to be a bit more current limited and dropped off in voltage with use (a characteristic often known as “linear”). The actual available capacity, however, was much more similar to an eneloop in my experience … the XTARs ran noticeably longer in all cases.

The charger, however, was a disappointment. While it certainly works as a storage case and looks the part, I found it took quite a bit of time to charge – part of this is due to the cells themselves having a limited charging current capability, but the other was the charging case itself being a bit conservative about avoiding overloading the USB input. On many uses, the case either got stuck in blinking (i.e. indicating charging) on one or more channels which never seemed to complete or rapidly indicated solid (i.e. charging complete) on a channel that did not actually complete, requiring a restart to clear. This makes charging take longer than expected.

The design also is a bit annoying. The clearance around the batteries is generous, meaning that the batteries rattle inside the case. The magnet that closes the flap is close to the bottom row of batteries, meaning that it likes to grab onto cells that are being inserted or removed from the middle two slots in the bottom row. Because the flip-open lid has all of the circuitry in it, it’s not possible to add or remove cells without disturbing the charge of all other channels … noting that the case only handles 1.5V Li-Ion type cells making it a little less versatile than some dedicated chargers. The USB-C power input being on the top lid to the left was not ideal but given the weight of the cells, it didn’t matter too much as the cable weight wasn’t going to move the charger anywhere.

Still, the price advantage of these particular products cannot be denied, even if the capacity was not particularly special.

Performance Test

Testing of the cells were performed soon after the XTAR USB LR 3000mAh cell, using my suite of equipment and test script. This includes the Rohde & Schwarz NGM202 two-quadrant power supply, HMC8043 power supply, MXO4 oscilloscope, Tektronix PA1000 Power Analyser, Keithley 2450 SourceMeter, B&K Precision Model 8600 DC Electronic Load, modified car fridge and ZKE Tech battery rack.

Weight and C/5 Capacity/Energy

Unfortunately, as I had not read the cells carefully, I didn’t know what capacity they were, so some of the tests were done based on an assumed capacity of 2500mAh (multi-rate) and 2100mAh (cycle life). The actual advertised capacity was 2400mAh, which is not that different, but this means that I didn’t get an actual exact C/5 rate value.

Overall, four cells were tested of the 16, but all of them were weighed. It seems the weights are fairly close, generally speaking. The capacity was a bit disappointing – averaging 2136mAh which isn’t even quite eneloop territory. The energy averaged about 2969mWh, both values notably less than the claimed 2400mAh/3600mWh as that figure may just represent the bare cell without any conversion losses.

Multi-Rate Discharge

Multi-rate discharge tests show that the available capacity and energy is dependent on discharge rate. The best results seem to be around 100-250mA, with available capacity decreasing as current increases.

There is a disparity between the two cells which may represent individual unit variances and perhaps the settling of the battery in its first few cycles. However, this trend does reverse somewhat past 1.5A and this is because the output voltage actually dips.

The same information plotted based on energy, where the reversal seems to happen above 1.75A.

In all tests, the temperature was fairly tightly controlled, so the modified car fridge was doing its job.

Plotted based on time, the voltage characteristic appears to be linear until about 1.25V, then linear but steeper, down to nearly 1.1V where it cuts out.

This trend is reflected for most of the discharge rates, however, rates above 1.25A seem to show unexpected voltage perturbations which may be due to internal heating effects, suggesting that such high currents should be avoided. The voltages are also noticeably reduced at the higher currents.

This is made especially clear when it is plotted with reference to capacity.

The same data is plotted here as a function of energy.

Cycle Life Endurance

To see how the cells would survive a “lifetime” of usage, I decided to run 50 back-to-back cycles at 0.42A (estimated capacity of 2100mAh at C/5 rate) on two cells.

The raw data is tabulated above, and as usual, the first cycle seems to be a little anomalous likely due to being charged from an already charged state at the beginning of the test.

Based on the trend, both cells were losing about 0.197 to 0.290mAh per cycle based on a linear fit. If we take the claimed 2400mAh, then 1920mAh would be its end-point. Based on an initial capacity of 2144mAh/2153mAh, this would result in 1136/803 cycles respectively if the same linear trend were to persist. This is a fairly high number which suggests more than sufficient durability for most users.

When plotted as a function of energy, the curves appear to be less steep.

The round-trip efficiency of the cells is about 61% efficient, but with a difference between the two cells, likely representing component differences (e.g. Rds, inductor DCR, etc).

Temperatures were held stable throughout the testing as confirmed by the chart of average temperatures for each cycle.

Output I-V Curve

It appears that the output voltage is mostly stable with a slight hump, but Cell $6 shows a bit of an odd bump at about 1.25A while Cell #3 shows a similar discontinuity at about 1.5A where the voltage starts to curve upwards again, peaking at just shy of 2.5A and then decreasing as the converter is overloaded but does not shut down. I would recommend the use of <1.25A to ensure that the cells can perform in a stable manner.

Charging

The behaviour of the AA station is both perplexing as it is inconsistent. When connected to my HMC8043 with my own adapter set up to emulate a USB DCP, it dutifully caps it consumption to 1.5A which is “safe” but results in a long charge time of 5h 36m for all eight cells. During charging, it seems the case interrupts the charging current periodically. Even with one cell, it would seem that the current is periodically interrupted.

Testing a single cell on its own shows that charging takes just shy of three hours, with a peak current of 354mA and termination of 63mA set by the cell’s internal BMS. Quiescent was 0.24mA with 986.8mAh delivered at 5V into a linear charger. If the cell has a 3.6V nominal voltage, this would correspond to 3552mWh; 3.7V nominal voltage, 3651mWh, close to the label claim of 3600mWh.

To fully achieve the maximum charging speed, I tried using the case with the Novoo RG68 monitored using the Tektronix PA1000 Power Analyzer. While the initial charging seemed to complete by around three hours – the time for a single cell, unusual behaviour was noted in that some banks simply did not stop flashing while others finished prematurely. Restarting the charge, it’s clear that some cells were taking quite a bit of current for another three hours or so. Once it seemed to stop taking power, some bays were still blinking, so I restarted the charger again whereby some further charging was achieved.

Ultimately, I couldn’t rely on the charger to bring all the cells to a fully charged state as both the user feedback indication and termination seemed to be flaky.

AA Station Case Quiescent

The charging case was advertised as an ideal storage solution for batteries between uses, so I was interested to see how much would be lost from the battery if stored in the case.

Satisfyingly, my Keithley 2450 SourceMeter indicated just 21.242nA draw, meaning it would take around three years to consume 2000mAh and empty a cell.

Ripple and Noise

This was measured with the 10:1 probe with 20MHz bandwidth filter.

Unloaded, the voltage is a sawtooth indicating that the converter has a low-power mode. The ripple is about 20mV peak-to-peak, frequency is about 140Hz.

At a low draw of 25mA, the waveform becomes a bit more sinusoidal in a way, with around 20mV peak-to-peak ripple at 255kHz.

Increasing the load to 100mA, it seems that the waveform has irregular dips rather than being a consistent repetitive waveform. Perhaps it’s in the cross-over between modes, nevertheless the ripple is well controlled at 25mV peak-to-peak.

At 200mA load, it seems this irregularity started to result in larger ripple spikes, reaching 123mV peak-to-peak on average. This is perhaps not ideal for some sensitive instruments, but seems rather normal for such 1.5V Li-Ion cells.

At 500mA, the ripple waveform returns to a clean sinusoidal shape, likely as the converter is no longer pulse-skipping. The ripple averages 155mV peak-to-peak which isn’t as low as it could be, with a frequency of around 253kHz.

At 1A loading, the ripple increases to 359mV peak-to-peak. Considering batteries (normally) don’t produce any ripple whatsoever, this level of ripple is rather excessive and might leave some sensitive instruments a little unhappy.

At 2.3A (the final current I could achieve before the waveform disintegrated into nearly randomness again), the ripple reached 503mV peak-to-peak – this is about 33% of the output voltage. So I wouldn’t consider these cells capable of delivering high currents.

Transient Performance

For some fun, I decided to see what would happen if the cell was exposed to some load steps – the output is well regulated and no severe excursions are seen, but instead we can see the ripple being low at 0.1A load and increasing to large levels at 0.9A load.

Teardown

I can’t help my own inquisitive nature but to at least attempt some sort of teardown to see what is inside of it.

Cell

I wanted to know more about the switching converter used in the cell, but …

… when unwrapping the outer label, I discovered the unit has a solid metal outer shell. That’s a nice construction – no chance of the circuitry shifting around or getting damaged.

From the top, I see that there is a yellow PCB that seems to somehow be crimped in, perhaps soldered to the shell for one of its contact. The silkscreen has a logo (EL?) that I don’t recognise, with the text A0. This has me second-guessing whether Keeppower is really the true OEM, or is there someone behind that too?

AA Station

The charging case’s guts are mainly in the flip open lid, which has a translucent plastic outer. The insert is clipped in, but despite trying to pry around the edges multiple times, I was not able to release it without resorting to destroying the plastic.

Once broken apart, a black PCB is seen from the top with the channel LED indicators.

Each of the LEDs is a bi-colour SMD LED package. There are header pins for power and I2C. This board is fastened to the carrier with five screws – care is needed when disassembling here …

… as the positive terminals are all seemingly made of a ball bearing and spring combination – disassemble this wrong and you’ll have to look all over the room to find them!

The charger PCB is dated 19th June 2024 with a name of TF-CD-009LDC-V3.1. It has a HASL finish which has pads contacting the springs – perhaps this is not an ideal situation (usually ENIG/ENEPIG might be used to better resist tarnishing and oxidation) but perhaps one constrained by cost. There seems to be banks of resistances and MOSFET switches probably to run the initial trickle for cell type detection, or to limit charging current (or cell internal dissipation due to linear charging).

There are some ICs that are anonymous – perhaps an opamp?

The main brains of the charger is most likely a custom-programmed microcontroller, marked F806A43214.

Conclusion

The Imuto 1.5V Li-Ion AA (2400mAh/3600mWh) Battery + Charging Station Set is a rather inexpensive entry point into 1.5V rechargeable Li-Ion batteries. While it does seem to work and cycle life testing suggests it has adequate endurance, it seems there are a few drawbacks worth noting.

The first is capacity – the 2400mAh/3600mWh refers to the “raw” capacity equivalent based on the Li-Ion cell itself, without consideration of converter efficiency losses. Actual output energy and capacity is dependent on load, but it seems that close to the standard C/5 rate, the actual resulting capacity is just 2136mAh/2969mWh which makes it a bit less than some decent Ni-MH alternatives.

The output characteristic is also not a consistent 1.5V which is something which some users like to demand. Instead, it seems to be a two-gradient linear, first down to 1.25V then steeper down to 1.1V to try and avoid unexpected shutdowns. This does result in some equipment showing varying performance as the cell depletes. The output also seems to have quite a bit of ripple, above 10% peak-to-peak especially when loads increase above 500mA, with the cells seeming to have average voltage wander above 1.25A loading but did not cut out entirely even when loaded up to 3A.

While the AA station charger looks good, it holds the batteries loosely resulting in some rattle while its magnets seem to have a habit of grabbing onto cells as they are inserted or removed from the middle two slots in the bottom row. Its performance was overall, inconsistent, with some channels seemingly never terminating charge at random, or other channels terminating charge early requiring multiple manual restarts to eventually bring all cells to a fully charged state.

I suppose, to some extent, you do get what you pay for. Being a lower cost product means you’re not going to break any capacity records, nor will you have the greatest output power capacity or power quality. But for some non-demanding applications, these cells will be adequate and the cost savings may be worthwhile. This allows a wider range of consumers and appliances to reap the benefits of rechargeability, higher voltage and no electrolyte leakage. But if you really need the best (and you’ll probably know it if you do), then there’s always XTAR …