Review, Teardown: *Fake* Xiaomi (Mi) NDY-02-AD Silver 10400mAh Power Bank

It might seem that I’ve been singing the praises of Xiaomi/Mi power banks lately, and that’s because they’re very good products. However, the journey to obtain the product is often fraught with more difficulty than expected.

In my case, when I ordered my initial silver unit for my own review, it was lost in the post, delaying my progress. I decided to try and order a second batch of units from a different seller to try and speed up the process of getting one, while the first seller eventually agreed to ship me another unit.

The initially ordered unit was the one that went into the silver review, and that unit was a genuine. However, the order from the other seller turned out to be counterfeit. This is one of the big problems of trying to buy the Xiaomi – there are so many counterfeits! Alas, there are many sites which tell you how to spot one, but sellers rarely provide enough information to tell at a glance, and sometimes they will send you a product which differs from their pictures.

After a protracted several-week back and forth with the seller, which included some terms I was less than happy about, I managed to get a full refund for the products. Many people weren’t as lucky. This gave me certainty, as I could then proceed to test the products and find out if they were any good.

So in a bit of a twist from the regular, I’m going to do what few people have done before, and actually review a counterfeit product. This might come in useful for people who have purchased one, and need some support when it comes to identification and proof of counterfeit product.

The Items

It was pretty much a case of “I knew you were trouble when you walked in …” with these counterfeits. These were sold out of a trading company in Hong Kong which probably supplies many eBay sellers under different names.


The counterfeit units came in clear plastic cases with a white paper over-wrap. This over-wrap doesn’t fit nicely, and the edges are folded, which represents poor quality. There is even a pretend serial number, with the number 472700190770 which is identical on both units and should never be.


DSC_8232The rear of the wrap is plain and white, with no authentication sticker in site.

The unit arrives neatly packed in a clear plastic case which seems to be a nice touch, although depending on the handling, the case itself could have a few cracks in it. Visible is the cable, the unit and a manual – pretty much the same inclusions that the genuine power bank has.

But remember, the genuine version never comes in a plastic case!!!


DSC_8238Taking the contents out, the inclusions are similar, although the finish on the power bank is different. While both are aluminium, the finish on the counterfeit is a coarser more reflective look, with imperfections at the edges (you will see later).

The unit itself weighs in at 218gm, which is a little lighter than the 252gm/260gm result from the two genuine units.



Side by Side Comparison

Things are definitely clearer for all when the units are put side by side (or top-bottom).


The genuine unit is above, the fake unit is below. It can be seen that the fake unit has a slightly more translucent plastic, with a less-refined power button in terms of its inset and large LED indicator holes. It also has a different microUSB port, a similar looking USB port but omitting the mi logo – zoom in to see.


The genuine bank is on the left, the fake is on the right. The coarseness of the finish can be seen when placed side by side, although physically they share identical dimensions.


In my case, it seems the URL is printed a little higher, and it’s the older URL, although genuine banks do carry this URL as well.



The underside shows a clear difference, with the genuine above, and the counterfeit below. Note how the genuine has a smooth finish, whereas scored lines can be seen on the fake (just barely). The printing on the fake is much more contrasty, being a silver on white “label” whereas the genuine has a sharp grey paint print on a plastic cover that has no wavy texture. The letters in the Li-ion logo seem to spell Li-ien on the fake.

A look at the cables also shows the telltale fake sign – the genuine cable on the left has a black tongue and no Mi logo! The fake has a Mi logo, and a white tongue, which it shouldn’t!


Plugging it in, you will find that the fake cable also doesn’t perform data synchronization as it has no data wires wired through, but the genuine does.

Other differences are found in LED behaviour – with them remaining on during discharge instead of very occasionally blinking (as the genuine unit does). Recharging LED indication is identical, and so is the charge time at five and a half hours.


Taking the unit apart like a real Xiaomi isn’t going to get you anywhere. As it turns out, this unit is constructed differently. You will need to start at the tail end, by removing the label and unclipping the end cap which holds the guts in place.


Once you’ve done that, the whole unit can be pushed out of the metal shell. The imperfections in the manufacturing can be clearly seen here.


Unlike the genuine Xiaomi, there is rubber foam tape inside the shell after pushing the unit out.


What is it made out of? Well it’s certainly not the Samsung or LG cells which are advertised by Xiaomi – these are unbranded cells, which are unmarked and appear to be locally produced (of questionable quality and safety). The PCB is blue and marked 9113-M4 on this side. It seems the D+ and D- are shorted together, so this unit acts as a dedicated USB charging port.


There seems to be damage to the insulation from being “pushed into” the sharp edges of the metal shell during construction. There is also evidence of stray solder-balls. Should the insulation wear through, there is a potential safety risk in short circuiting with the metal shell should the positive somehow detach or touch the shell. As the cells are of questionable origin, they may or may not have any PTC protection internally. It wouldn’t surprise me if it didn’t …

There are no NTC thermistors to be seen either – this is part of the temperature sensing required for safety and higher speed charging and is another advertised feature of Xiaomi power banks.


The whole rear of the unit appears to be formed with a white plastic carrier, adding to the weight (and possibly compensating for cells which are light, a sign of low capacity). The underside clearly shows the indicator LEDs on the PCB and the holes in the body of the case. This design features no diffusing tape or dark plastic, so the indicator differs from an authentic Xiaomi in having certain viewing angles that are glary (direct view of LED) and also light leakages from each LED that light up the surrounding plastic. Definitely not of the same quality there.


The other side of the PCB shows just one unidentified controller, U1, which controls all operations. There is no Texas Instruments controller, as advertised by Xiaomi. The inductor is a regular, open, unshieleded type, with an undersized 1A SS14 Schottky diode. This is not proper design for a 2A converter. There are some ceramic capacitors on the output, but not many. The PCB is very recent, dated week 31 of 2014.


Speaking of which, the tabbing and battery connections leave a lot to be desired. The single tab strip doesn’t look like it was applied with much care, with poorly aligned spot-welds and the pack “falling out of shape”. The wire connection is made via very thin hook-up wire – definitely not the best choice here.


They didn’t fare much better on the top – arguably, it’s even worse, with poor quality wire soldering and solder splashing everywhere. The tab strip is too long as well, which increases the risk that the excess ends may eventually touch something they shouldn’t.

If you’ve read this far, there is absolutely no doubt that this is not an original Xiaomi product.

Performance Test

Keep in mind, this is not an authentic Xiaomi product and the results that follow are not representative of what you will get from a real Xiaomi product. They may even not be representative of the performance of your fake Xiaomi product, as there are several counterfeit designs out there.

Testing was performed using the same methodology that the other Xiaomi products were tested with. Of note is the nature of the resistive test load – when the voltage falls below the nominal voltage, the current falls as well. As a result, the load levels of 500mA, 1A and 2A (corresponding to 10 ohm, 5 ohm and 2.5 ohm) are only sustained by a true 5v source – the actual currents are actually a little less.

Lets take a look at the capacity results:

Load (mA) Run Capacity (mAh)
500 1 3857.50123
500 2 3873.520161
500 3 3828.532666
500 4 3804.587905
500 5 3787.785843
Mean 3830.385561
Range 85.73431843
StDev 35.6354511
Load (mA) Run Capacity (mAh)
1000 1 3714.424338
1000 2 3719.15866
1000 3 3706.623432
1000 4 3670.331475
1000 5 3683.2535
Mean 3698.758281
Range 48.82718495
StDev 21.04722632
Load (mA) Run Capacity (mAh)
2000 1 2960.503267
2000 2 2941.377836
2000 3 2924.84206
2000 4 2924.311745
2000 5 2902.682109
Mean 2930.743404
Range 57.8211585
StDev 21.57758676

At about 500mA, 1A and 2A loadings, the capacity was 3830mAh, 3699mAh and 2931mAh respectively. The results were very consistent, with all ranges below 86mAh. This is below the advertised 6800mAh minimum, and 10400mAh typical/nominal value. It’s no surprise – it’s a fake. Using the provided rating, the efficiency is 36.8%, 35.6% and 28.2% respectively, a far cry from the up to 93% claimed for real Xiaomi products.

As a result, it can be concluded that the batteries are significantly under capacity. It is likely the pack is 4800mAh in total, with 1200mAh per cell, with a conversion circuit that is about 80% efficient. Not the worst 18650 cells, but not good either.


During testing, however, it showed significant problems with voltage regulation. As mentioned in the teardown, I was skeptical based on the Schottky diode rating that the unit was able to sustain 2A load, and indeed, it couldn’t.

At the “2A loading”, the voltage started at below 4.5v and continued falling to about 3.9v towards the end. Because of the low voltage, the actual current delivered was closer to 1.4A rather than 2A due to the resistive load. As a result, the unit is incapable of even meeting the specifications printed on it in terms of current load and will not be able to charge any device at 2A.

At the ~1A loading, it only just managed to stay within the 4.75v lower limit for USB voltage, until the end of discharge. At 500mA, it was very stable at 5v, rather than 5.1v as printed on the outside. Again, its performance does not meet its specifications.

A look at ripple and noise is revealing as well.


At the 500mA loading, the ripple averages 434.6mV peak to peak, which is high (above the 150mV of normal wall chargers, and ~100mV of the genuine Xiaomi) and close to the absolute limit of 500mV peak-to-peak imposed by the 4.75v-5.25v range requirement.


When testing at 1A, the power bank exhibited a ripple with two main frequency components. This increased the peak to peak ripple significantly to 827.8mV, which is extremely high and out of specification for USB. This is likely to cause stress and potential damage to equipment.


While testing at 2A, it seemed it had settled back down to 834.1mV peak to peak, but then I discovered that the power bank actually does alternate between ripple regimes, especially at start-up.


In this case, at the “2A” loading, it produced an average of 1800mV peak-to-peak, for a whopping 36% ripple content. But considering it didn’t actually give out 5v, and closer to 4.4v, the ripple content is actually closer to 41%.

This sort of performance earns it the title of worst powerbank tested to date in ripple performance.


One question I get asked a lot is – given that Xiaomi products are relatively inexpensive, why would counterfeiters try to target such an inexpensive product? The margins are already slim, so why would they bother going after such a product if what they are interested in is profit?

In short, I don’t know. It’s probably got something to do with volume. Initially Xiaomi power banks were hard to get your hands on due to shortages of stock – the manufacturer couldn’t keep up with demand, leaving a demand “gap” which counterfeiters could capitalize on. A good product with good reputation will see good demand, and even with slim margins, it may be possible to make up for it in sales volumes. Maybe they even have a few spare cells they need to get rid of.

But alas, the jig is up, and many pages have identified such counterfeit items. So why do they keep selling it? Well, they’re hoping that you don’t find out because it’s so well made. Maybe they’re hoping that by the time you find out, it’d be too late and you’re past all buyer protection scheme entitlements.

Or more likely, they want to dump the product on you. The scheme works like this – sell a given product, it might be rubbish but if the buyer complains, then offer a partial refund on the guise that a full refund is only available if you send the item back, and knowing that the cost of postage is uneconomical, many people will settle for this. Often the partial refund is small to the point of being offensive (in my case it was), but it will be enough that the seller can still make a tidy profit.

Should you settle?

Which neatly brings me down to this point – should you settle?

In a word, no. Protest. Dig in your heels. Don’t give up. If you’ve purchased something counterfeit from eBay by PayPal, you are covered by buyer protection. The policy has specific clauses in relation to counterfeit products:

Under Section 4.3 Potential Buyer Risks:


If you file a claim about a counterfeit item you may be required to:

  • Obtain documentation from a qualified third party to substantiate your claim and/or
  • Destroy the item and provide evidence of this at your expense; or
  • Make the item available for the seller to collect at their expense.

It is against the law to possess or sell a counterfeit item.

Under Section 4.4 Potential Seller Risks:

ITEMS MAY NOT BE RETURNED If your buyer files a claim about a counterfeit item they may be required to:

  • Obtain documentation from a qualified third party to substantiate their claim and/or
  • Destroy the item and provide evidence of this at their expense; or
  • Make the item available for you to collect at your expense.

The payment may be reversed.

It is against the law to sell a counterfeit item.

As a result, if the item is counterfeit, you should be able to contest this for a full refund. This is all without posting the item back unless the seller pays to collect it or have it posted. The only burden is a need to prove the item as counterfeit, and to destroy the item on their instruction and document it.

Often a clear statement in regards to you knowing your entitlements and threatening negative feedback and PayPal cases will eventually succeed in resolving the matter in your favour, although often at the cost of leaving positive feedback anyway. I suggest you take this route – you really don’t need to waste more of your time.

And afterwards, you might have a product in your hand, but I would suggest you consider disposing of it immediately. It’s a safety risk to both you and your devices (some have exploded and caught fire).


The fake Xiaomi power bank may look similar to the original, but it couldn’t be a more different product. The product does not feature the same level of design and safety as the genuine one, and is composed of low quality components.

The capacity falls below half the stated capacity, while simultaneously being unable to meet the current output rating and producing so much ripple and noise to be a risk and threat to your devices.

If you receive one of these, or have one in your posession, I would highly recommend you stop using it. There is nothing to be gained if it catches fire, and burns down your house. There’s nothing to be gained if the power is so unstable that it damages and kills your $500 smartphone. All for a US$20 counterfeit charger.

As it turns out, there are several different variants of counterfeit Xiaomi powerbanks, and almost certainly each of them will be slightly different in performance. But it seems likely, inferring from this unit, that they all are a long way away from ever meeting the same performance standard of the genuine Xiaomi.

Posted in Electronics, Power Bank | Tagged , , , , , , , , | 3 Comments

Review, Teardown: Xiaomi (Mi) NDY-02-AD Gold 10400mAh Power Bank

This post comes to you thanks to a generous reader from Germany (thanks Tobias!), who decided to contribute a power bank for testing. Unfortunately, the reader didn’t tell me what it was before it was dispatched, and it turns out to be another Xiaomi (Mi) 10400mAh.

In my last review of the Xiaomi, I reviewed the silver coloured edition which is the most popular one. However, the Xiaomi also comes in a range of other colours at a slightly higher cost – this one is a golden one.

While the colour of a product shouldn’t affect its performance, I decided it was worthwhile testing this unit for several reasons. There can be subtle build differences and revisions in products which can affect performance, and there might be quality control issues worth highlighting that only manifest themselves after analyzing several units. Statistically, having just one sample isn’t particularly significant – so lets see if Xiaomi can repeat their stellar performance again.


DSC_8199 DSC_8201

This power bank arrived in a matte cardboard box, identical to that of the silver one. A notable difference with the silver one is the use of the URL, which indicates this is probably an older unit. Newer units are marked with instead. This one also had a stock tracking barcode label which appears to be for the distributor/reseller’s use.


The side of the box has a serial number (top) and UPC barcode (bottom) along with the model number.


The other side is adorned with specifications, and the official verification seal. I did scratch off the silver coat and verify the numbers underneath at, where it correctly verified as genuine, as the first time the code had been entered.


Included is the power bank, a Chinese manual and a USB micro B cable. The micro B cable had no Xiaomi logo, and had all wires connected for both data transfer and charging, which is another indicator of genuine status. The finish of the golden power bank is to a high standard, identical to that of the silver one. The front is adorned with the Mi logo, the rear features the URL.

DSC_8202 DSC_8203

The ports are buff coloured, and a Mi logo is seen inside the USB-A connector indicating, again, its genuine status. The specifications are printed in light grey on the other side of the power bank, and the indicator LEDs are small. All good signs!


According to my not-so-trusty scales, this unit weighs 260gm, which is a little more than the 252gm of my silver unit, but might come down to slight changes internally and some added weight because of the finish.


Breaking apart the power bank, an identical form of construction is seen.


Internally, the circuitry is laid out identically to that of the previous power bank. Some slight exceptions are seen in the PCB silkscreening, which indicate this PCB was manufactured by a different company, to the same design specifications. From the date code, this PCB was made in Week 12, 2014 which is 6 weeks prior to the PCB in my silver power bank.


The same sort of LG cells are visible inside, but the printing on the cells are all “wavy”. I’m not sure why this is the case, but the cells themselves appear to be genuine.


The marking anomaly affects all the cells within this power bank.


On the other side of the PCB, we can see everything is identically populated to that of the silver unit, down to the same unpopulated capacitor spot. The Abov microcontroller is dated Week 9, 2014, which is 5 weeks earlier than the one used in the silver bank.


In terms of construction, some slight improvements can be made with the mylar insulation on the tabbing – there’s a slight gouge on the negative link, and the spot welding seems a little less consistent than in the other unit, but it’s still a good effort in general.

Performance Testing

The same methodology used in the previous review is carried forward, with the main interest in verifying the capacity of the power bank and calculating efficiency.

Load (mA) Run Capacity (mAh)
500 1 9747.957136
500 2 9761.970879
500 3 9764.186531
500 4 9762.0129
500 5 9742.946146
Mean 9755.814718
Range 21.24038552
StDev 9.666239966
Load (mA) Run Capacity (mAh)
1000 1 9276.94628
1000 2 9466.482107
1000 3 9469.5623
1000 4 9475.603401
1000 5 9471.726582
Mean 9432.064134
Range 198.6571208
StDev 86.77706413
Load (mA) Run Capacity (mAh)
2000 1 8761.804708
2000 2 8760.62915
2000 3 8768.616895
2000 4 8735.541467
2000 5 8693.808093
Mean 8744.080063
Range 74.80880164
StDev 30.78258426

At 500mA, the usable capacity is 9756mAh, at 1000mA, the usable capacity is 9432mAh and at 2000mA, the usable capacity is 8744mAh. The recorded value of usable capacity at 1000mA may be slightly reduced because of the first run showing a ~200mA disadvantage possibly due to inconsistent charging (early termination). However, again, a very consistent charge termination was achieved for all other runs except the first.

Calculated efficiency figures are 93.8%, 90.7% and 86.4% at 500mA, 1A and 2A respectively, using 3.7V as the basis. Using 3.6V, the efficiency improves to 96.4%, 93.2% and 86.4% respectively.

Compared to the silver power bank, the resulting capacity figures are all within 50mAh! Likewise, the efficiency figures are virtually identical and the minimum capacity and maximum efficiency claims are all met.


The voltage profiles are slightly different but exhibit the same level of stability over the long term and are all well within the USB voltage specification requirements.

xmgold-500ma-5ms xmgold-500ma-500us

The rather Xiaomi specific “two-step” ripple is also present. At 500mA, the values are 84.16mV peak-to-peak at 5mS/div, and 91.56mV peak-to-peak at 500uS/div which is a commendable result, below 150mV of wall chargers. This is slightly less than the silver unit, but within expected sample-to-sample variation.

xmgold-1a-5ms xmgold-1a-500us

At 1A load, the values are 87.5mV and 90.11mV respectively which is higher than the silver unit, but consistent with prior readings. It’s a good result, in fact, the silver bank’s result is a little suspiciously low.

xmgold-2a-5ms xmgold-2a-500us

Increasing the load to 2A pushes the values to 108mV and 95.15mV respectively, which is almost indistinguishable from the silver power bank, and an exceptionally good value seeing as the ripple output is almost independent of the output current.


In a result I didn’t quite expect, the Xiaomi Silver and this Xiaomi Gold perform almost indistinguishably on capacity (within 50mAh), and ripple performance. This is good news, as it indicates that the quality control is probably quite good, and quality LG cells are consistent for capacity. Each of them has their slight construction quirks, but nothing too major.

I’d have to say this is another good score for Xiaomi. I am very happy with its performance.

Posted in Electronics, Power Bank | Tagged , , , , , , , | 1 Comment

Visited: Floriade 2014

This week, I had the opportunity to go on a short day road-trip to Canberra, and on the way, we stopped off at the Big Merino. After that, I spent some time admiring the month-long Floriade display, in its final week. It’s normally best to visit Floriade in the earlier weeks, as the freshly produced floral displays tend to be better presented. Visiting later risks seeing the flowers at the end of their run, although crowds seem to be drastically reduced.

This post is mainly going to be a photo-gallery style posting, as there’s really not much commentary to go with it. I did spend quite a lot of time trying to chase some bees and insects, but being late in the day made it rather challenging. I know my photographic composition is probably, a one-out-of-ten at best with boring framing, but I do love the colours and textures of the flowers. Enjoy.

Posted in Photography | Tagged , , | 1 Comment