Review & Teardown: Kogan USB 3.0 Universal Memory Card Reader

Last week, Kogan had a free shipping day, and not being one to let such opportunities slip by, I decided to pick up a Kogan USB 3.0 Universal Memory Card Reader for AU$19. It’s a fair price to pay for such a product, similarly priced to my favourite Transcend RDF8. I needed another reader, so I decided to place a gamble on this one to see if it’s any good.

Unboxing

This one was shipped from Kogan’s Australian arm via eParcel and arrived in a few days. The package itself is in a plastic clamshell which comes apart without needing any tools. The cardboard insert is very plain all around, and eschews colour or even feature-blurbs as they really aren’t useful for an online shopping experience. Maybe they saved a few dollars too.

In the package, all that is provided is the reader unit itself, and a thick, high-quality 0.6m USB3.0 cable. I wish they would provide some that were longer, especially for those who have rear mounted USB 3.0 ports. In true Kogan style, the manual is only provided online and is a short one-pager. No drivers or CD are included or provided online as this works as a USB Mass Storage class device and is driverless under most modern operating systems.

The unit itself is mostly black panels, with a white centre (similar to an Oreo). The top features a clear plastic cut-out which lets LEDs shine through.

Cards are accepted “front and back” with no slots on the side. The CompactFlash slot is next to the microUSB-B connector, which can be a little inconvenient for some users. There are six slots in total – the manual appears to list the format support although it doesn’t seem quite right – for example, one port is labelled MS Slot, and the other is labelled Memory Stick slot. Likewise, one port is listed as MMC slot, and the other is listed as SD slot. It may be incorrectly labelled, maybe they mean XD and M2 perhaps?

I don’t have that many cards anyway – SD, microSD and CF are the most important in general.

Teardown

Couldn’t resist trying to find out what the reader is made out of, so I pried it open. Note that this is constructed with glue and plastic clips which crack, so once you take it apart, it’s going to need some glue to be reassembled. To open it, you need to pry the rear black solid plate away from the white plastic surround.

Inside, we see that many connectors are plastic, as expected. The crystal is visible on this side. The PCB is secured with a dab of hot-glue which must be removed to solve the mystery of what is the chipset?

The PCB itself is marked with HC-RST5301-VCR55820120316 (sic). The chipset used is a Realtek RTS5301 which doesn’t have much information available online, although is used in several different products online.

There are many locations for LED mounting, but most of them are unpopulated bar two. This may have been for indication of a “per-slot” access status, but to save cost, they’ve decided not to populate them.

There is an EEPROM – a Pm25LV010 1Mbit SPI Flash, likely for firmware. The soldering of some of the components looks moderately poor.

The top lid has nothing special to offer us, although the black residue around the side may be due to the glue used to bond the case together – the solvent may have actually eaten the plastic or the paint from it.

Using the Reader

To use the reader, one merely has to connect it to an appropriate USB 3.0 (or 2.0, or even 1.1) port. The OS is able to detect it as a “generic” device, indicating that the device ID strings have not been changed from defaults.

It’s as simple as plugging cards in and using it, right? Unfortunately not for this reader is a little bit of a problem child!

I found that my Kingston 128Gb SDXC and my Toshiba Exceria 64Gb SDXC card will not work with the reader at all. It causes the light to extinguish on the reader then periodically singly-flash while disconnecting from the USB bus altogether. The reader will not reactivate until unplugged and replugged.

Plugging the reader in with either card in a slot will cause it to fail to be detected. The ONLY SDXC card in my collection that works is the Sandisk Ultra 64Gb microSDXC.

All my SDHC and SDSC cards perform as expected with the reader. My CF cards also perform as expected.

I’m not sure if it’s a dud or not, but judging from some other people’s experiences, there were reports that readers based on the RTS5301 chipset have problems with Mac machines running Mavericks as well. As I have no such machine, I can’t tell whether there will be problems with this particular device.

Unfortunately, since I had cracked it open, and due to needing to shoulder return mail cost if taking advantage of the money-back guarantee, it looks like I’ll be stuck with it like this.

Testing it with the Sandisk Ultra 64Gb card produced similar results to my Transcend RDF8 reader, thus indicating that it’s likely that the RTS5301 is no slouch, when it works. On the CrystalDiskMark, it seems that it doesn’t have the same anomalous 512k result, but the 4k results are slower possibly indicating more command overhead or latency with this reader.

Can it be Fixed?

Of course, that got me thinking – can it be fixed? Unfortunately, datasheets on the RTS5301 are non-existent, and Realtek doesn’t have any available firmware updates, but what happens if I try to apply the update for the Sandisk SDDR-289 to this reader? The reader is the same chipset after all …

As it turns out, the reader is recognized and it seems that the firmware version is actually older. Clicking on update …

… does actually result in a successful update. The device is then re-detected as a Sandisk.

I recommend against doing this as there is no way within the software to back-up and restore the original firmware, and the new firmware alters the number of slots available! This may affect your ability to use the reader.

Sad to say, the reader is detected correctly, and functions as before – still no luck with my Kingston 128Gb SDXC or my Toshiba Exceria 64Gb SDXC.

UPDATE: It can be fixed!

After some playing around with the firmware updater, I found out it is a Realsil chipset (a subsidiary of Realtek). I accidentally flashed an incorrect firmware from a related product (RTS5307) to it which bricked the reader.

So, I thought, I might as well remove the EEPROM. I was pretty much ready to throw away the bricked reader, so I didn’t care about being delicate. With a screwdriver, I lifted the chip along with most of the traces. If you have the tools, desolder it properly, or clip it at the leads but ensure you don’t leave them shorted.

Lo and behold, the reader was detected – it was unbricked and is now running with the mask ROM defaults in the chips. I assumed this would happen as many of these chipsets can be built into readers without EEPROMs to save cost. As a result it may show slots that aren’t existent.

Interestingly, the mask ROM has no problems with my Kingston 128Gb SDXC or Toshiba Exceria 64Gb SDXC! It works! And none of the other cards have problems with the reader as it is. With the Exceria, we can see speeds of close to 90Mb/s which indicates this reader is fast, although similar to the RDF8 (~4Mb/s faster).

Why did I have to sabotage the product to get it to work? I’m not sure. I would assume there’s either a firmware bug (the more likely cause), or maybe somehow the EEPROM shares the bus with the SD cards, and just aren’t happy with certain cards attached to it or running at UHS-I clock rates.

Running the firmware update program results in the following:

So, it appears, without the EEPROM, the original mask ROM firmware is version 0126, which is several revisions earlier than the one on the EEPROM supplied. As a result, it may now be buggy under certain conditions (e.g. Mavericks). I still can’t recommend the reader, although if you have one and you can’t get your SDXC card working and you’re almost about to bin the reader, you might as well give this a try at your own risk.

Conclusion

Unfortunately, I cannot recommend this reader to anyone at the present moment. As it stands, I’m not sure if I have received a dud or not, but it certainly has compatibility issues with my cards, whereas the Transcend RDF8 does not. Both are a similar price, and the two chipsets currently make the bulk of the USB 3.0 card reader market.

Pressed to make a decision, I would prefer a Genesys Logic based reader based on this experience.

This used to be a common issue, and still seems to be, despite the increasing standardization of flash cards. Compatibility issues can range from no detection to slow performance all the way up to card corruption. I’ve experienced them all at one point or another, so I’d advise you choose your readers with care.

With the update, I’ve been able to make it work by sabotaging it intentionally. While it’s a strange finding, at least it’s not an entire loss for me!