Quick Review: Sandisk Ultra 128Gb UHS-I microSDXC Card

I was pretty disappointed to have been dudded in my first attempt to get a hold of a 128Gb microSDXC card, but I won’t let an eBay fake stop me. After all, Sandisk products are supposed to be high quality. This time, I decided to high-tail it to a brick-and-mortar store (ARC Computers) to grab one from the shelf for AU$121.26 – and managed to grab the only one. This also provides a solid set of comparison images for direct comparison in case there is doubt.


Despite the item listing at ARC claiming to have 30Mb/s read speed, it seems that the cards that were stocked are the updated version of the package, now claiming up to 48Mb/s.


Specifically, it is noted that the tray and card sits neatly in its cutout, the printing is sharp and the colours aren’t oversaturated. The reflection in the phone screen is fine, the centering of the logo is good. The hanger-tag cut-out is the special “squared off” version rather than a generic hanger cut-out. Everything’s all good so far.


The rear seems to have a new re-designed printing with significant reduction in printed text by the removal of other languages. This is probably because this packaging is specific to Australia and New Zealand market. Important is to note that there is no inconsistencies in the use of the SDXC logo – you won’t find any SDHC logo hiding here.


The package also feels a little different – this is because there’s an electronic article-surveillance tag (metal strip acousto-magnetic/sensormatic type) pre-installed in the package. This can make for a difficult experience sliding the card carrier out.


The card carrier has a very large area at the top, where a specific warranty statement for Australia and New Zealand is provided. The package also extends significantly left and right of the cut-out as well so there’s little chance of popping it out.


The card itself is directly printed upon and is solid edge to edge. No signs of white edges or remarking should be apparent.


The rear contacts are nice and clean, with a green PCB substrate (which is very similar to their Sandisk 32Gb Extreme card) along with place of manufacturing markings on a black painted area.

DSC_9782 DSC_9783

The adapter is also up to the quality expected – straight pins, laser etched date and place of manufacture at the back, and “rails” to guide the card into the adapter. There really isn’t any doubt about the authenticity of this product at this stage.

The read-out of the card’s characteristics are as follows:

Capacity: 127,864,930,304 bytes
CID: 035344534c31323880504583a700ea5f
CSD: 400e00325b590003b8aa7f800a40400d

The vendor ID code of 03 matches the expectation for Sandisk products.

Performance Test

Unlike the former Ultra products which only claimed up to 30Mb/s (which was actually conservative), this one claims 48Mb/s. Lets see if it’s actually capable of reaching this mark.

HDTune Pro with Transcend RDF8


Reading with the RDF8 fell short of the 48Mb/s claimed on the packet, averaging 42Mb/s. This is, however, in the ballpark of what the Ultra cards have performed in past testing, and is still miles faster than the counterfeit stuff.

HDTune Pro with Kogan RTS5301


No compatibility issues were had with the card in the Realtek RTS5301 reader (which is often picky about which cards will run at UHS-I). The average read speed is faster, at 43.9Mb/s, still a little short of the claimed “up to” speed.

CrystalDiskMark with Transcend RDF8


The card seems to put out relatively expected speeds for an Ultra product, noting that the sequential write performance is actually pretty good for a Class 10-rated Ultra card. The small block performance is not too bad either, compared to others in the database.

CrystalDiskMark with Kogan RTS5301


It performs a little faster on the read with the Realtek based reader, but is slightly slower/the same on writes. It’s good to see the card maintains good performance across the two tested readers.

H2testw with Transcend RDF8


Is this a surprise? Hardly. Sandisk is a byword for quality – and that can only be the case if their products live up to their specifications. In the H2testw benchmark, no failures of the flash were found, with the write averaging at 14.9Mb/s, and read averaging at 41.2Mb/s. This isn’t quite the 48Mb/s on the packet, but it’s still quite a bit above the 30Mb/s on their “older” packet designs. It’s probably just a marketing choice so as not to be unfairly compared to other manufacturers who may also put slightly inflated “up to” figures on their packages.


Go to a decent retail establishment, pay the right price, get the right stuff. What a shame it isn’t this simple when it comes to eBay sellers. At least, here, I have documented what a genuine Sandisk 128Gb Ultra UHS-I MicroSDXC card looks like, along with the packaging and validated its performance. While it doesn’t quite meet the new 48Mb/s “up to” label, it blasts past the 30Mb/s claims of the older range of packages and performs consistently with the rest of the Ultra series of cards. It easily surpasses Class 10 requirements for sequential writes, and is the highest capacity card available to date. This makes it easy to recommend, provided you are buying a genuine product.

The performance test, CID and CSD databases will be updated shortly with the results of this review.

Posted in Computing, Flash Memory, Photography | Tagged , , , | Leave a comment

Quick Review: *FAKE* Sandisk Ultra 128Gb UHS-I MicroSDXC Card

The holiday periods around Christmas are a big time for shopping and spending, and one of the hottest items is undoubtedly electronics. Unfortunately, for those who aren’t as wise or careful, it’s also an opportune time to be scammed with the supply of sub-standard counterfeit products at reasonable looking prices. Part of the whole idea stems from the fact that gift-givers often don’t keep receipts, won’t return products, and won’t test them until it’s too late. By then, the seller has probably absconded with the money, never to be seen again.

It’s definitely advisable to practice some caution – don’t fall into deals which are too good to be true. I really needed a 128Gb microSDXC card to pair with another gift to myself, that will have to wait to be reviewed. Knowing the price for flash memory is ~$0.50/Gb for SSDs of that capacity, the fact a 128Gb microSDXC card was available for AU$80-ish didn’t seem too outlandish. Pressed for time, and looking for a deal, I decided to go with an undisclosed Australian eBay seller with very positive (99.5%) feedback. Having learnt from my last trip with fake flash, I decided to opt for something in full retail packaging to avoid the “bulk” packaged scams. This is what I received.UnpackagingI received a single padded envelope with the retail package inside. A quick glance at the retail packaging might look convincing, but it had me worried.

DSC_9749Having previously reviewed several Sandisk genuine products, this review of a 32Gb Ultra provides you a good idea what the genuine package should look like.

Significant differences include:

  • Different package hanging cut-out tab in the top.
  • Poor colour printing in the background with changed colour and saturation on the smartphone and almost invisible “reflection” of the clock text. Greys are too dark in the printing.
  • Slightly thin text in the Full HD logo on the top.
  • Crooked cardboard cutting at the top edge, with it tapering upwards towards the corner.
  • Slight font thickness differences in the blurb text.
  • Changed size of “speed” label underneath the capacity.
  • Poor fit (crooked) of the plastic card carrier in relation to the box – card carrier appears to be different with only very little “excess” plastic off the edge, so that the edge of the carrier is visible from the front.
  • Sandisk logo is not properly centred on the bottom.



A flop over to the rear confirms the fear of a fake with poor quality printing that’s too thick and not centred. It’s even on the verge of being cut-off on the right edge. This is not characteristic of Sandisk quality gear. Another key issue is the use of the SDHC logos on this package, when it should be the SDXC logos as the card is >32Gb capacity. (The comparison Ultra 32Gb has SDHC logos, as appropriate because it’s an SDHC card).

I was pretty sure the card wasn’t genuine at this point, but to prove it beyond reasonable doubt, it has to be unpackaged and tested.


A close look at the plastic carrier doesn’t seem to inspire confidence. From unpacking other Sandisk products, normally, the carriers have larger “wings” around the sides to stop them from moving around in the package.


In this case, the plastic only just covers the card storage locations, which seems a little suspect.


DSC_9753Close examination of the card reveals an unusual white fabric like cloth edge, and what appears to be a glossy “label” applied over the base card. This results in a white edge all the way around the card, which implies it has been remarked.

The rear of the card isn’t marked in normal Sandisk markings, which normally include the country of manufacture, in a neat small font further away from the card edges. The contacts also appear to be scuffed, implying it may have been tested, used or modified prior to packaging. This is at a level above normal expectations.

DSC_9755 DSC_9756

A careful examination of the adapter reveals that it is of sub-standard quality. Regular Sandisk adapters have their printing well aligned, and guide rails inside the microSD slot void itself. Furthermore, they have sharply defined notches, instead of the ragged edges seen above. Another deviation is the bend in the adapter’s contacts in the top right hand corner of the second image, and a lack of laser etch text which adorns the rear of genuine Sandisk adapters.

Further examination was performed by retrieving the card’s manufacturer defined registers (CID and CSD). These hold information pertaining to the card itself and should be programmed at manufacture.

Size: 128,849,018,880 bytes
CID: 0020204e2f412020103144a22300ecfd
CSD: 400e00325b590003bfff00000a4000b7

The CID values can be compared with other genuine Sandisk cards in my collection:

035344534430324780019acc7600844b SANDISK 2Gb Blue SDSC
035344535530314780401c751300637d SANDISK ULTRA II 1Gb MicroSDSC
0353445355363447801013d98600d6d3 SANDISK ULTRA 64Gb UHS-I MicroSDXC
035344534533324780231f9dcd00dc7b SANDISK EXTREME 32Gb UHS-I microSDHC
03534453553332478000e4b46c00e123 SANDISK ULTRA 32Gb UHS-I microSDHC
30534453553034470027901eaa00b97f SANDISK 4Gb Class 4 microSDHC

Note that most if not all Sandisk genuine cards have their first 2 octets (the Manufacturer ID) set to 03 or 30 (although the 30 may have been a typo on my behalf). In fact, checking the manufacturer ID of 00 in my database matches none of the cards I own, and therefore does not come from a reputed manufacturer, let alone Sandisk. The value of 00 may even be reserved and not allocated to any manufacturer – this is something only the SD Foundation would know.

This alone is another indicator that the card is not a quality item, and not a Sandisk item. Along with this, it seems the CSD has been tampered with to indicate the card as 128Gb in hardware, thus it doesn’t appear to be the wrong capacity even when re-formatted (as opposed to repartitioning scams which are easily undone).

Performance Testing

Another sure-fire giveaway of a bad card is a bad performance result. Due to the size of the card, it takes quite a while to test, so it took a whole day to collect the following results.

HDTune Pro with Transcend RDF8


The card shows an unusual difference in read speed near the beginning which is not repeated in a pattern, which implies something wrong with the flash. The read speed is only a paltry 21.6MB/s which implies it’s operating in Class 10 transfer modes, rather than UHS-I as indicated in the package. Traditionally, Sandisk cards under-promise and over-deliver, as can be seen in the table of previous readback tests where Ultra cards consistently delivered 44Mb/s+ on the same equipment.

HDTune Pro with Kogan RTS5301


You might scream “compatibility issue” but it’s not. Here’s it tested with another USB 3.0 reader which has qualified many cards, even up to 128Gb size, and the performance is within a whisker. The performance doesn’t show the variance in performance, but just because the whole card is read does not mean it’s working properly.

CrystalDiskMark with Transcend RDF8


Testing with CrystalDiskMark only utilizes the first 1Gb of the card, and was successful. The results achieved are uncharacteristic of Sandisk Ultra cards, with the worst 512kb and 4kb write performance delivered to date. No other card, even un-classed, has achieved such poor write results at both 512kb and 4kb transaction lengths. Again read performance sequentially falls far below the 42Mb/s achieved by other genuine Ultra cards.

CrystalDiskMark with Kogan RTS5301


These sentiments are echoed when tested with a second card reader. This is not what you should be seeing from a Sandisk Ultra card.

H2testw with Transcend RDF8


The proof is in the pudding. This card is not a genuine 128Gb card and only has about 5Gb of workable flash media. The rest of the data is “overwritten” in an endless loop, with the first 64Mb protected so the file-system doesn’t appear corrupted but the files definitely are. The full write and read exposes the poor performance of the card which fails to meet 10Mb/s on either writes or reads and cannot be considered Class 10!

The full log is below:

Warning: Only 122847 of 122848 MByte tested.
The media is likely to be defective.
5.1 GByte OK (10771712 sectors)
114.8 GByte DATA LOST (240818944 sectors)
Details:114.8 GByte overwritten (240818944 sectors)
0 KByte slightly changed (< 8 bit/sector, 0 sectors)
0 KByte corrupted (0 sectors)
64 MByte aliased memory (131072 sectors)
First error at offset: 0x0000000148ba0000
Expected: 0x0000000148ba0000
Found: 0x0000001df77a0000
H2testw version 1.3
Writing speed: 7.86 MByte/s
Reading speed: 8.17 MByte/s
H2testw v1.4

This test took over 8 hours of my time to complete. It’s conclusive. I’ve even verified this with a hex editor which shows data repetition/overwriting, and by filling the card completely with other files and checking their contents.

After the test filled the card, an HDTune Pro test was run with the card, and the speed inconsistency can be seen to align exactly with where you would expect it to be based on the H2testw result. This is not a coincidence.


Counterfeit Item Protection

It is at this time I should remind readers of their rights under Paypal’s Buyer Protection. 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.



Through diligent documentation, it is proven beyond a doubt that this item is a counterfeit. The packaging was poor and did not match up to that of a genuine product with colour differences, layout differences and poor alignment. The card itself was not correctly marked and showed signs of remarking, the adapter was below standards and did not have the expected mark text on it. Even the manufacturer’s data in the CID register did not define the Vendor ID correctly to be Sandisk (03-hex). The card underperformed for speed, and was also not capable of handling the amount of data it was advertised for.

At the moment, I am in discussions with the seller to resolve this problem. I hope it will be resolved to my satisfaction – I will not shy away from opening a Paypal case if necessary.

ADDENDUM: The seller was very apologetic and offered to resolve the problem immediately to my favour. The items have been removed from the market and the supplier is being contacted.


Well, I couldn’t get to sleep. I got my money back, and I have the item on hand, but it’s utterly useless to me. One niggle was in the back of my mind – what lies underneath the label like covering on the card? Only one way to find out – a sharp hobby knife.


Savaging the card with a knife isn’t as easy as it seems. Whatever glossy print material they used seems to be quite like the coke-bottle labels – they’re very thin and slippery. Attacking it at different locations with a sharp knife allows us to remove some of it. On genuine cards, marks are always applied to the surface of the card itself, and no knife can remove them. The fact a knife has this property is a dead giveaway that this is NOT genuine.


Slowly, and persistently, more of the plastic overprint is removed. Unfortunately, nothing seems to show underneath. Better finish the job anyway …


Et voila! The remarking has been removed off the card. The identity of the original card is still unknown. The reason for this is that the card has been carefully ground down so as to roughen the surface to improve adhesion of the remarking label, as well as to ensure the whole remarked card is not too thick for insertion. The grinding away of the material has taken away any original manufacturer prints on the card, if any!

And now … I can probably go to sleep.

Posted in Computing, Flash Memory | Tagged , , | 3 Comments

Teardown: A 9V Carbon Zinc Battery

DSC_9594I suppose this particular post belongs to the “everyone should have done this at one point in their lives” bin, but strangely, they probably haven’t. Just the other day, one of my multimeters decided to let me know that it needed a battery change. Out came the old 9v battery, and in went a new one. So, what do we do with the old battery? Throw it away … right?

Well, that’s only one possibility. Another is to take it apart to understand what’s inside a 9V carbon zinc cell. After all, there’s nothing really chemically dangerous in these cells as long as you don’t eat it, so why not?

This particular cell comes from Thumbcells, a Chinese company whose logo is a little scary …


The battery is fairly recent, so it complies with RoHS regulations for the protection of the environment. The first step is to locate the seam that runs down the side of the battery – this is where the metal “casing” for the battery is crimped together. This particular casing is only for mechanical rigidity and doesn’t serve any electrical function. Once the seam has been located, take a small flat-bladed screwdriver and begin prying along the seam.


The way the seam has been made, the can is actually folded in on itself in an interlocking way, which explains why it is so difficult to take apart. A pair of pliers could help. Already, we can see some of the structure – lets peel away the outer can.


This reveals the internal structure of the battery. The 9v battery is composed of six 1.5v carbon zinc cells connected in series. The six cells are seen here, looking suspiciously like cellophane-wrapped Werther’s Original candy … but trust me, they’re not for eating. The stack is wrapped in clear plastic to hold it together and insulate it. The plastic top card holds the terminals, whereas the bottom plastic card just completes the cell. The folded edge of the outer casing applies mechanical pressure to the end plastic cards to “squeeze” the cells together to make contact.


The terminals themselves are only little pieces of metal with “spikes” to bite into and make contact with the cells. The positive terminal is connected to the bottom of the stack and the metal contact is further insulated with paper to avoid short circuiting under mechanical stress (e.g. crushing). Since these cells give so little current due to high internal resistance, even a short circuit is unlikely to cause major issues aside from a dead battery.

DSC_9599 DSC_9600

Here we can see the “bite” marks from the top and bottom terminals into the cells themselves. The end cells seem to be constructed specially, with soft zinc plate contacts for the terminals to bite into. The stack of cells can be further disassembled by scoring the outside plastic shrink wrap and peeling it away.


DSC_9602It’s now when you realize the “gooey”, almost plasticine like nature of the cells. As they’re made of electrolyte-soaked carbon powder and what appears to be cardboard, you’re acutely aware not to squeeze them too much or you’ll get a mess … so out comes the outer layer of clear plastic.

This allows us to separate cells from the stack itself, and we can already see the difference with the cells inside the stack – their ends are not nice and shiny and are instead very much dark and carbon-coloured. Also, surprisingly, we find that we’re not free from the clear plastic shrink wrap just yet, as each individual cell is also plastic wrapped.


Squeezing the cell, it becomes evident that it’s not all soft. There’s something inside …


There’s lots of carbon soaked in some electrolyte, and that makes a pretty nice mess if you stake it apart. It will also rust some metals as it’s slightly corrosive too. But removing the wrap entirely …


… reveals an embedded thin zinc plate electrode. So it seems that there is no carbon rod in the cell, as it might not be required due to the small cell size itself.


After making a ripe old mess, we managed to get a good look at the lozenge-like cells that make up a carbon zinc 9V battery, and look into those cells too. Unlike alkaline 9v cells, there are no AAAA-sized cells inside … so don’t buy carbon zinc cells expecting to extract AAAA cells! Maybe I’ll break one of those apart in the future just to show the difference.

Posted in Electronics | Tagged , | Leave a comment