CARDSPEED - Card Readers and Memory Cards

History

2007-01-25 - Test Results for the new Hama Readers
The tests are not completely finished, but it's time to show some results:
Hama USB 2.0 UDMA CF Reader
Hama FireWire 800 CF Reader
Hama CardBus CF Adapter
Hama CardBus 30 in 1 Adapter

The results with the Hama CardBus adapters are about the same as with other CardBus adapters using the same controller (ID 1145-F021). People at Hama seem to get much higher transfer rates with the CardBus 30 in 1 Adapter, so the performance might depend a lot on the chipset of the Notebook. I will try to get more information about this issue.

A special note for people who take a very close look at the test results and asking for the fastest card reader: The results for the two FireWire CF readers with the Extreme IV CF look very different at first glance, 40.27MB/s vs. 39.93MB/s, but the actual difference is only 0.85%.

2007-01-08 - Hama: Four New Card Reader Products
Hama expand their card reader segment with four new devices in two special categories: Fast Single-Slot CF Readers and CardBus Adapters.

Hama 00055352: Hama Digital Tech-Line FireWire Traveldrive CF Card I/II MD
Hama 00055353: Hama Digital Tech-Line USB 2.0 Traveldrive CF Card I/II MD
Hama 00055354: 32bit PC-Card Adapter 30in1
Hama 00055355: 32bit PC-Card Adapter CF I/II

The Single-Slot Tech-Line CF Readers are supposed to be released mid January.
More details will follow...

2006-12-11 - Memory Stick PRO-HG
SanDisk and SONY announced the Memory Stick PRO-HG format. This new standard raises the clock rate from 40MHz to 60MHz and introduces an 8-bit data interface by adding four data lines. This will most likely be done by introducing a second row of pads in a similar way as with the MMCplus interface. I just expect the new pads to be in line with four of the existing pads, most likely four in the middle. Memory Stick PRO-HG will be compatible with older devices and work the same way as a regular MS PRO would do when inserted in a 4bit slot.

After introducing new modulation schemes and differential signal paths, parallel interfaces were considered outdated. But step by step, the new memory card interfaces go back to parallel interfaces. Serial interfaces have advantages if you have to cross distances over a wire, but from socket to card there isn't much distance or wire.
Currently, we have CF with 16bit at 2x33MHz, MS PRO-HG with 8bit at 60MHz, MMCplus with 8bit at 52MHz and SD with 4bit at 50MHz. This means that the next candidate for an upgrade would be the SD standard that recently suffered from the SDHC upgrade - after successfully doubling the clock rate. I can't wait to see whether they will use one step to 8bit at 64MHz or if they do the SDHC8 upgrade first, and then the SDHC8HS upgrade - or the other way round.

2006-11-30 - Super SD: Interpreting the Results
In the meatime, I did a quick check (8MB read and write) in about half of the card readers I have available. If the card really runs faster in MMC 4 mode than in SD mode, then I haven't found any card reader where the card runs in MMC 4 mode. But there are several card readers where the card runs in the slow MMC 3 mode, although the reader would suppport at least SD, and in some cases even MMC 4. With this behaviour, the Super SD qualified for a column in the SD Compatibility Table.

Step by step, the regular test results will be added. But this will take some time, especially for the card readers where the card runs in MMC 3 mode. For the moment, the test results can be found at the top of the SD 1.01 Test Results and at the bottom of the MMC 4 Test Results.

2006-11-28 - 1GB Pretec Super SD Card
After about 12 weeks of waiting, the 1GB Pretec Super SD was finally available. The card turned out to be as I expected. It looks exactly like an MMCplus card: 1.4mm thickness and 13 pads. In terms of SD it would be a Thin SD without write protect switch, but four additional pads.
First checks showed that the maximum read reate seems to be about 10.6MB/s, and the maximum write rate is about 3.4MB/s. It is very hard to tell which modes the card actually supports, because the maximum read rate is still within SD 1.01 and MMC 4-bit 26MHz performance. Further tests in different card readers might reveal a little more detail.

2006-11-21 - PIXO 28 in 1 PC-Card Adapter
What's so special about this PCMCIA adapter? Well, some time ago I added test results for USB 1.1 and PCMCIA for reference purposes. Then I realized that my old PCMCIA adapter does not support MS PRO or xD-Picture Cards, so I bought a new one. But the new one had no support for SmartMedia...
Together with the PIXO MSR-PRO, I received the PIXO 28 in 1 PC-Card Adapter, which supports all media types except CF - even xD-Picture Cards of Type M&H and SDHC! Due to strange limitations in the PCMCIA-ATA interface, the transfer rate is only up to 1.26MB/s, but the tests with some of the largest cards showed that the adaper works properly.

2006-11-20 - Test Results for the CompuApps USB 2.0 UDMA CF Reader
The test results for the retail version of the CompuApps USB 2.0 UDMA CF Reader are now available. As expected, the results are the same as those taken from the sample within the usual tolerances.
For a better overview concerning the different types of CF readers, I prepared a CF Reader Comparison table.

2006-11-16 - More SDHC Compatibility Results
In the meantime, the 4GB Transcend Class 2 miniSDHC arrived, and the results of the compatibility check are added to the Card Reader Compatibility Table.

Although the card is labeled as Class 2, the best write rate of the card seems to be about 1.75MB/s. This means that I have to contact Transcend and ask what's going on here.

2006-11-16 - CompuApps USB 2.0 UDMA CF Reader Released
Yesterday I received the retail version of the Omniflash Uno CF 2.0! / USB 2.0 UDMA 40 Uno CF Reader/Writer from CompuApps, Inc. The reader has a matte black rubberized finish with silk screen label instead of a sticker. Photos will be available in a few days.

2006-11-14 - PIXO MSR-PRO Multi Card Reader
Several days ago, I received the PIXO MSR-PRO Multi Card Reader from PIXO GmbH. In the meantime, the tests are almost complete, only a few unimportant cards have not been tested yet. The test results are excellent, thanks to the use of the powerful AU6375 controller. The reader did not show any problems, especially none of the problems I noticed with a card reader from a different manufacturer using an earlier version (0547 vs. 0608) of the AU6375 controller. Photos will follow later, unfortunately the reader arrived one day after the last photo session. Yes, sometimes the memory cards are abused for temporary storage of photos.

2006-11-13 - First SDHC Tests Finished
The tests with the Transcend Class 6 SDHC are finished. The Transcend cards aren't such good writers, so the best write rate was 10.64MB/s. But this is fast enough to show that there is a difference between minimum (6MB/s for Class 6) and maximum (10.64MB/s in a fast card reader). For the moment, the test results are added to the SD Card Tests.

For card readers, the result of the SDHC compatibility check is a little poor. I don't know when the specification was available, but at least some Alcor Micro controllers from the second week in 2006 support SDHC properly. Unfortunately, I haven't found a way to read out the true firmware level for Alcor Micro controllers, so the only hint is the manufacturing date printed on top of the controller.
A special guest in the SDHC league is the PIXO MCA-28 PCMCIA Adapter. Due to general limitations in the PCMCIA-ATA interface, it's not fast, but it works properly. The SD Compatibility List is almost complete, some readers managed to hide during the tests, and checking the adapters takes more time, since they tend to report wrong card sizes due to strange CHS mappings.

2006-11-10 - Busy...
Right now I don't have the time to write stories. To sum up the work in progress:
The first SDHC card arrived: Compatibility tests results are added to the SDHC special.
PIXO MSR-PRO, a current AU6375 reader.
And a few other things...

2006-10-23 - Pretec e-Disk II Test Results
A few days ago, I received a sample of the Pretec e-Disk II. This is a SD/MMC card reader stick with a closed card slot, so that it can also be used as USB memory stick, whenever it is equipped with a suitable memory card. Unlike many other SD/MMC sticks, the e-Disk II has support for SD 1.1 and MMC 4, and can keep up with current card readers. Based on my observations, the card should only be changed when the reader is unplugged to ensure reliable and proper detection of the card.

2006-10-20 - CompuApps UDMA CF Reader Delayed
Just a note to those who are waiting for the UDMA CF Reader: The realease was originally announced for September, but due to delays in the production process, the reader is now expected to be available late November or early December.

2006-10-10 - SD Compatibility Information
Who would have guessed that SD cards will become the problem child number two in the flash card family, after all the trouble with xD-Picture cards? I finally decided to start a page with information about SD Compatibility. This is just a first version with a quick overview, more information will be added step by step.

2006-10-09 - More SDHC Players
More and more companies are listing SDHC cards. In addition to previous announcements from Toshiba, Panasonic, SanDisk and Pretec: This means that right now there are most likely more companies offering SDHC cards than there are customers willing to buy these cards...

2006-10-05 - NEXTO CF Ultra ND-2525
The first notes about the NEXTO CF Ultra ND-2525 are available.

2006-10-04 - OnSpec Inc. xSil251 Based USB 2.0 UDMA-4 CF Reader
OnSpec Inc. provided a sample of a CF Card Reader using their xSil 251 Hi-Speed USB Extreme Flash Drive R/W Controller. The card reader is distributed by CompuApps, Inc. and is named OmniFlash USB 2.0 UDMA 40 UnoCF Reader/Writer.

This card reader sets new speed standards for USB 2.0 CF readers with a top performance of 30.9MB/s read and 22.9MB/s write for the SanDisk Extreme IV CF card. This reduces the gap between the fastest FireWire and fastest USB 2.0 readers from 40.2 vs. 18.4MB/s to 40.2 vs. 30.9MB/s. With fast PIO-6 CF cards, the reader is even a little faster than current USB 2.0 readers, that already leave FireWire readers behind. With fast PIO-4 CF cards, FireWire readers are a little faster, but for fast PIO-2 CF cards, the USB solutions are again a little faster - just to explain the different rankings with different cards. Enough talking, now check out the Test Results. The results for the Microdrives will follow soon, but you can be sure that there are no problems.

2006-09-29 - Layout Changes
Due to the increasing number of cards and test cases, the Test Results by Cards had to be split up in smaller units. More explanations later.

2006-09-29 - 2GB xD-Picture Card Type M
I could not resist and ordered a 2GB xD-Picture Card. After the first tests finally finished, I was quite surprised: In terms of xD, this card hits "astronomical" rates of about 5.5MB/s read and 2.1MB/s write on a compatible, fast and error-free card reader! More results will follow, but this will take several days.

Speaking of xD-Picture Cards: Due to the high random access rates and the very low read rates, the duration of a complete test run for a 1GB xD-Picture card is about 2-3 hours(!). Therefore, I will run all future xD-Picture Card and SmartMedia tests on my secondary test system. The results might be a few percent less than on a faster system, but speaking of speed and fast in conjunction with these card types is rather ridiculous.

2006-09-27 - SanDisk Extreme III SD
After the good experience with the Extreme III MS, it was obvious that I have to try the Extreme III SD as well. I was watching the price for about 10 days, and the day after I placed the order, it dropped a few Euros - that's life. The special thrill was that I already had three high-speed SD cards that performed very well compared to the limit of the current SD specifications. But the SanDisk Extreme III 2GB SD (SDSDX3-2048-902) was able to beat these cards in both, read and write rate. The first results with the fastest SD readers are already included in the Test Results by SD.

Just to explain this Extreme III triple-feature, and as a special note to all manufacturers: The moderate results with the SanDisk Ultra II CF and Ultra II SD, and especially the bad random access performance of these two cards kept me almost two years from buying another SanDisk product. But the current Extreme III card series are really well done!

2006-09-19 - SanDisk Memory Stick PRO Duo Ultra II and Extreme III
The test results with the Extreme III CF and Extreme IV CF cards showed that SanDisk finally learned their lessons in speed and access time, so it was worth trying their Memory Sticks. The SONY Memory Sticks showed the strange behaviour that they could be read fast right out of the box, but dropped to about 2/3 of that speed after the first data was written.
First tests with the SanDisk Memory Stick PRO Duo Ultra II 4GB (SDMSPDH-4096-902) showed a read performance of about 16MB/s and a write performance of about 10MB/s. The SanDisk Memory Stick PRO Duo Extreme III 2GB (SDMSPDX3-2048-902) shows about the same read performance around 16MB/s, but a higher write performance of about 16MB/s. The results for the Extreme III using the fastest readers are already included in the Test Results by MS, first test results with the Ultra II and the results using the slower readers will be added step by step.

Some notes on the filesystems: The SanDisk Memory Stick PRO Duo Ultra II 4GB comes formatted to FAT32 with 32kB cluster size, which is also the default format of my SONY DSC-T3. But the DSC-T3 also accepts the 4GB media formatted to FAT16 with 64kB cluster size. The SanDisk Memory Stick PRO Duo Extreme III 2GB came with a "dirty" FAT where 10 clusters (320kB) are already marked as occupied on the blank media. Further examinations showed that the start of the FAT seems to be wrtten every 48 sectors (24kB), where the FAT entries for the non-existing clusters 0 and 1 are interpreted as "occupied cluster, end of cluster chain" entries, and therefore do no harm. The same effect happens when formatting the blank media with the SONY DSC-T3, and seems to be a bug in the SONY formatting procedures. This effect does not happen with the SONY Memory Stick PRO Duo 2GB that has a slightly different available size (2,015,887,360 bytes versus 2,045,771,776 bytes).

2006-09-18 - More SanDisk Extreme III and Extreme IV CF Test Results
I am still busy running tests with the Extreme IV 4GB CF on the slower card readers. But I managed to finish the tests with the card readers that are able to cross the 10MB/s barrier. The fastest of the below 10MB/s readers achieves 8.31MB/s with the Extreme IV CF. This means that this formerly fast reader achieves 20% of the read performance with todays fastest CF card. See the Test Results by CF/MD for details.

I also had some aces up my sleeve, and this was the right time to make use of them. I added a complete set of test results using a CardBus CF reader. This reader is supposed to use the same hardware as the Delkin labeled Aska CF32A I used during my first tests. Because the adapter did not work properly after some time, I decided to get a replacement some time ago. But the SanDisk Extreme III and Extreme IV CF are the only cards so far where the CardBus CF Adapter shows good performance, but still not as good as current USB 2.0 readers.
Another surprise was the A-Tec CF reader, which is using exactly the same controller as the Vivanco RW CF, and therefore did not get closer attention. I already noticed that the Vivanco RW CF is not able to read the Transcend 120x 8GB CF properly, but did not find the time to analyze this in detail. I assumed that the problems would be related with the 8GB size, but the tests with the 1GB Extreme III CF and the 4GB Extreme IV CF showed problems as well. The A-Tec CF reader is able to read these cards without any errors, so I would say that the board layout of the Vivanco CF reader causes problems with cards faster than about 100x.

2006-09-13 - New Test Result Pages and Test Procedures
During the last years, I gathered about 60 Memory Cards with a total capacity of about 50GB, and about 55 Card Readers (some of them are only listed in the Card Readers and Controllers cross-reference) with a total of about 180 card slots. Due to the extensive test procedures, I can't test all combinations of cards and readers - especially if there are slow cards or readers involved, or even worse, if errors occur that require further tests and investigations.

I have carefully selected a set of nine different cards that are fast and not too big for quick-testing multi-card readers. These cards are suitable to analyze the capabilities of a card reader in terms of speed across the current range of card formats and interface standards, and to be able to do a proper comparison between different card readers. If a card reader shows good results with these cards, it qualifies for futher tests. If a card reader shows outstanding results with certain media types, further tests will be done with these types, but other tests will be omitted. The procedure for Single-Slot card readers is a little different: Depending on the card types and the performance of the reader, more or less cards will be used.

The Test Results, Quick Comparison shows a wider range of card readers with a reduced set of cards, while the Test Results, Detailed Comparison shows results with more cards, but only for selected card readers. The tables now also show the key data of the controller used: Model and either the firmware level, or the manufacturing week (format: YYWW) in case there are no distinct firware levels available. The full test details for a specific card reader are still available on individual pages an can be selected from the Test Details, by Readers.

And last, but not least: Don't forget that the detailed results from my first extensive tests with 24 different card readers and several card adapters can still be looked up in the Test Archive.

2006-09-11 - SanDisk Extreme III CF, Extreme IV CF and Extreme FireWire Reader
In order to keep up with the development in flash card speeds, I was more or less forced to make some little investments: A SanDisk Extreme III 1GB CF (SDCFX3-1024-902) and the SanDisk bundle SDCFRX4-4096-902 consisting of the SanDisk Extreme IV 4GB CF (SDCFX4-4096-902) and the SanDisk Extreme Fire Wire Reader (SDDRX4-CF-902). The first test results are already included in the tables. The formerly huge test results tables are in a transition phase to a new scheme, that will be explained here later.

Currently, the only way to make use of the full speed of the Extreme IV CF cards is using the SanDisk Extreme FireWire reader. The achieved read rate of 40.27MB/s is even above the transfer rate of the fastest USB 2.0 hosts, which means that even future USB 2.0 card readers will not be able to catch up. But the test results for the Extreme III CF show that current FireWire readers are not able to make use of the full potential of the previously fastest CF cards, such as the Extreme III CF or Transcend 120x CF. In technical terms: Current FireWire readers have a small advantage with fast CF cards offering good PIO-4 performance (as for example the SanDisk Ultra II CF), but stay behind the performance of current USB 2.0 card readers and cards offering faster modes than PIO-4, but no UDMA modes.

2006-09-11 - SanDisk Extreme IV 4GB CF: Filesystem Warning and Information
My filesystem checks revealed a serious configuration error in the factory default filesystem of the SanDisk Extreme IV 4GB CF. I can't tell whether other sizes and/or card types are affected in a similar way. The specified FAT size is one sector too small, leaving the last 34 clusters (1,114,112 Bytes) without valid FAT space. This means that filling up the card to the specified size could cause loss of data and/or serious corruption of the filesystem. The Canon EOS-1D Mark II, for example, recognizes this error and reports "Card1 not formatted / Card Err 1". Windows XP mounts the media with reduced size adjusted to the available FAT size, however without reporting any problem to the user. But this methos allows read and write access to the safe portion of the media. Please make sure to format the card in your camera before using it.

The details: The card contains a FAT32 filesystem with a specified size of 7,999,425 sectors with 32 reserved sectors, 976 sectors per FAT, 2 FAT copies, 64 sectors per cluster. This leaves space for 124,960 clusters, while the specified FAT size offers space for only 124,926 clusters. If a device does not check the FAT size, the last sector of FAT 1 will overlap with the first sector of FAT 2, and the last sector of FAT 2 will overlap with the first sector of the root directory.

2006-09-01 - Pretec News: 8GB SDHC and New Super SD (SSD) Cards
Pretec just introduced their new 8GB SDHC, but the recently introduced "Super SD Card" is the more spectacular news.

Actually, the Super SD cards would be Super MMC, but most likely for marketing purposes, they were named after the more widespread and popular SD cards. MMC 4 could offer more than twice the speed of SD 1.1 cards, but they have one serious drawback: Although they can still be used in SD slots, the cards fall back to 1bit 20MHz compatibility mode with a theoretical limit of 2.38MB/s if the host device does not suport MMC 4. The reality is even worse, because many host devices only run at 1.6MB/s in this mode.
The Super SD cards have the physical dimensions of regular MMC, already defined in the SD Card Specification as "Thin SD". It is not stated whether the Super SD cards will have the 9-pad SD layout or the 13-pad 8bit MMC 4 layout, and whether they can support SD 1.1 or only SD 1.01. The card is supposed to detect the fastest mode of the host device and act either as MMC 4, or as SD card. This means that the card could be able to perform at up to 17-21MB/s with SD 1.1 support and up to 9MB/s with SD 1.01 support instead of the horribly slow compatibility mode. Please note that in case of a 9-pad layout, the card could only use the 4bit MMC 4 mode. The details remain unknown until the arrival of the first Super SD card.

2006-08-25 - SDHC Launch and Comments
According to the Panasonic Online Shops, the Panasonic 4GB SDHC (RP-SDR04G, RP-SDR04GE1K for Europe, RP-SDR04GU1K for the USA and RP-SDR04GJ1K for Japan) is now available. The next companies will be SanDisk and Toshiba: SanDisk announced the standard series 4GB SDHC (SDSDBR-4096) bundled with their new MicroMate SD/SDHC USB 2.0 reader (SDDR-113). Toshiba announced the production start for the 4GB SDHC SD-HC004GT4.

There seems to be a lot of misinformation on SDHC, mostly due to the lack of public information from the SD Card Association. The new SD card series from Panasonic sheds a little more light on the subject: The RP-SDR series consists of Class 2 SD Cards from 512MB to 2GB and one Class 2 SDHC Card with 4GB. This means that the new SD 2.00 standard introduces several new features, and that SDHC is only the extension of the capacity. SD 2.00 cards of up to 2GB will still be SD cards being compatible with standard SD devices, while SD 2.00 cards of more than 2GB and up to 32GB will be SDHC cards and therefore require SDHC compatible devices.

But there seems to be some mystery concerning the SD 2.00 speed classs specifications: Panasonic announces their RP-SDR series as "...up to 5MB/s ... compliant with the Class 2 SD Speed Class specification...". Up to 5MB/s would still be no contradiction to a guranteed minimum sustained data rate of 2MB/s for Class 2, but the previous Panasonic RP-SDK series has a claimed read/write rate of 20MB/s. The real-world sustained read rate of the RP-SDK with the Panasonic CardBus SD adapter is 20.95MB/s and the real-world sustained write rate of the RP-SDK with the GL819 controller is 16.7MB/s. The question is, why the cards from the new series seem to offer significantly less data rate than the cards from the previous series. For SanDisk, this is similar: While there are Ultra II and Extreme III SD cards available, the new SDHC will be released as "standard" series card of Class 2. Only Toshiba announced their SDHC card as Class 4.

The SDHC cards are already listed in some regular internet shops that show prices of about 190 Euro for the Panasonic 4GB SDHC RP-SDR04G and about 110 Euro for the SanDisk SDSDBR-4096 bundle of 4GB SDHC card and SDHC reader.

2006-08-24 - New Cardspeed Scale
While preparing the "old" Cardspeed Scale, I already realized that there are two views necessary. But the CF 4.0 specification was the final reason for a redesign. Now there is one Scale showing the interface standards up to 25MB/s and the Memory Card Speed Ratings up to 166x. This range can be perfectly covered with USB 2.0 card readers, and Memory Cards with speed ratings up to 150x are commonly available. The second view shows the full range including the CF 4.0 UDMA-5 and UDMA-6 modes together with the theoretical and real-world transfer rates of different host interfaces.

This is the detailed view on the lower end of the full scale...
Image: CARDSPEED Scale: Details up to 25MB/s
In the range up to 25MB/s, CF, SD, MMC and MS are equally good represented.

This is the full range of current memory card interface standards. The top row shows the memory card speed ratings on the left side and the theoretical and real-world data rates of popular host interfaces on the right side.
Image: CARDSPEED Scale, full range
The bottom row shows the memory card interface standards. Above 25MB/s there's only the 8bit 52MHz mode of MMC 4 and the CF-ATA modes from UDMA-2 to UDMA-6.

USB 2.0 Card Readers should be fine for up to 30MB/s (200x). The fastest FireWire 400 (1394a) CF readers should just be able to keep up with the claimed 40MB/s (266x) speed of the SanDisk Extreme IV CF, while their FireWire 800 (1394b) Extreme CF reader should be able to show significant improvement with the next generation of Extreme CF cards. CardBus adapters should have even more potential than that, but the controller chips used often limit the data rates well below the performance of current USB 2.0 readers - especially for media other than CF.

2006-08-22 - CompactFlash Strikes Again!
One of my readers dropped me a note that the CompactFlash Association released Revision 4.0 of the CompactFlash Specification (CF 4.0). This new specification adds the missing two ATA modes UDMA-5 and UDMA-6 doubling the theoretical rate of the previous CF 3.0 standard. To sum up the history: CF 1.0 allowed 8MB/s, CF 2.0 allowed 16MB/s, CF 3.0 allowed 63.5MB/s and CF 4.0 now raises the limit to 127MB/s.
Does this mean that we will see 32GB CF operating at 80-100MB/s in the next few months? This would push portable devices and Ultra-Compact Notebooks to a new level, if these CF would be availble at reasonable prices.
Right now, 120x CF are available at about 20 Euro per GB, while SanDisk takes about 40 Euro per 20MB/s Extreme-III-GB and about 55 Euro per 40MB/s Extreme-IV-GB. And I have no doubt that there are enough photographers to support these prices.

2006-08-20 - Cardspeed Scale
To give a better impression of the different card interface speeds, I created the Cardspeed Scale:

Image: CARDSPEED Scale

This shows that only CF 3.0 and MMC 4 offer speeds of more than 25MB/s. The theoretical data rates are 63.57MB/s for CF 3.0 UDMA-4, 49.59MB/s for MMC 4 8bit@52MHz, 23.84MB/s for SD 1.1 4bit@50MHz and 19.07MB/s for MS PRO 4bit@40MHz.

The green arrows show popular memory card speed ratings. Although SanDisk does not give speed ratings, Ultra II means ~66x, Extreme III means ~133x and Extreme IV means ~266x.

The diagram can now be found on the Memory Cards Information page.

2006-08-16 - Flash Summary
It's time for a little summary. Besides the incredible price drop, there was also some development going on...

Capacity: SanDisk now offers 2GB microSD, which indicates that there is much room for the larger card types. SONY offers the Memory Stick Micro M2 up to 1GB, while the MMCmicro seem to be available at poor 512MB only. SanDisk also offers 4GB Memory Stick PRO Duo as regular and Ultra II type. Olympus and FujiFilm announced 2GB Type M xD-Picture Cards. SD cards are in some kind of a crisis since the SD Association declared 4GB "regular" SD cards as illegal, and introduced the new SDHC standard instead. CF cards are available at 8GB for quite a while now, but there are no 16GB CF announced yet. But it looks like the Pro photographers are not happy with the performance of the 4GB and 8GB FAT32 media, mostly due to the default cluster size of 4kB, which leads to huge FAT sizes. Some heavy-hitters even switched back to 2GB media.

Speed: SanDisk is just about to release the Extreme IV CompactFlash series with a claimed read/write rate of 40MB/s, about 266x in terms of X-Rating. FireWire 400 can just keep up with that rate, but even the best USB 2.0 host implementations are maxed out by these rates. The regular SD 1.1 cards have reached the maximum allowed by current standards, mostly 150x. This only leaves room for improvement for the miniSD and microSD. SanDisk now offers Ultra II microSD and Transcend introduced 80x microSD. MMCplus/MMCmobile would allow higher rates up to 333x, but besides some 266x announcements, the regular types seem to be stuck at poor 66x-80x. Although the MMC 4.x standard would be license-free and even faster than SD 1.1, there seem to be no serious MMCplus/MMCmobile implementations besides card readers. The main problem with MMCplus/MMCmobile cards is that they fall back to MMC 3.31 compatibility mode with a maximum of 2.3MB/s (16x) when used in older devices. SD 1.1 run at least at acceptable 66x when used in older SD 1.01 devices.

Prices: Starting in January, there was an incredible price drop that ended at 150 Euro for 8GB 120x CF and 77 Euro for 4GB 150x SD. Memory Stick and xD-Picture Card prices also dropped in price, but not as much as CF or SD. Noname SD cards can be found at 15 Euro for 1GB, but other card types and/or speed ratings can go up to about 60 Euro for 1GB.

2006-08-13 - More Slot-Icons
The Slot-Icons are now also added to the Card Readers Overview.

2006-08-08 - New Feature: Slot-Icons
In order to give a better overview concerning the supported card formats, Slot-Icons have been added. For a start, the icons are added to the Test Results Card Reader Selection. The Icons will be added to the full card readers overview in a few days.

The icons will show whether the card has to be inserted in normal orientation or upside down, or whether a passive adapter is needed. Some special cases are also indicated:
xD-Picture cards can be used in SmartMedia slots with xD to SmartMedia adapters, but support for cards of more than 128MB or Type M/H cards cannot be guaranteed.
Some Memory Stick slots in older card readers are short enough for using Memory Stick Duo, but the guides and contacts were not designed for the smaller Memory Stick Duo. In many cases, the Memory Stick Duo can be inserted if they are handled with care, but there's always the danger of bent contacts inside the slot.
Some SD slots have a tight fit. When using RS-MMC or MMCmobile with snap-on extender part, the extender might come off and the card can get stuck in the slot.

2006-07-19 - Tests finished
The tests with the Hama CardReaderWriter 1000&1 are finished - except for the 1.8" harddisk, due to the long duration of the test. As expected, the tests showed state-of-the-art performance. Unfortunately, the tests showed that there are problems writing to xD-Picture Cards of Type M and Type H, but this should not be taken too serious. I will report here if I get more details.

A few notes about the adapter-free concept: The combo slot for SD and miniSD has one drawback: It takes more force than usual to insert a regular SD card. Please make sure that the labels are properly attached and remove any labels that are not attached in a recessed area, or labels that exceed the recessed area. miniSD cards have to be inserted upside down along the left edge of the slot.
The combo slot for microSD and MMCmicro might take some practice: The smaller microSD have to be inserted straight and along the right side of the slot.

2006-07-11 - Hama CardReaderWriter 1000&1 / Hama TravelDrive Mobile 1000&1
A few days ago, I received two new card readers from Hama.

The Hama CardReaderWriter 1000&1 is the latest model of the famous "blue" card reader series. This is a multi card reader supporting all media types - except Memory Stick Micro M2 - without the use of adapters. In addition to the four "regular" slots at the front, there is a combo slot for microSD and MMCmicro at the right side. The controller is a Hama GL819 with firmware 9323, delivering top-notch performance for all media types except for an average rate with the old MMC 3.31 cards. Detailed test results will follow.

The Hama TravelDrive Mobile 1000&1 is a "USB stick" type card reader for cards of the SD/MMC/MS families. It supports all subtypes of these card families - except for Memory Stick Micro M2 - without the use of adapters. It's a three-slot two-drives reader with one SD / miniSD / MMC 8-bit slot and one microSD / MMCmicro slot for the SD drive and one MS / MS Duo slot for the MS drive. The controller is again a Hama GL819, but with firmware 9325. Detailed test results are already available.

2006-07-11 - New Table Gridline Layout
The table gridlines were originally formatted as empty cells with black background to ensure compatibility with older browsers. The drawbacks were large file sizes and slow processing.
The gridlines of most tables are now formatted using stylesheets, which means that they will not display properly on older browsers. The display should be OK for IE6, Netscape 7, Opera 7 and Firefox. The tables look really bad with Netscape 4.x and the look a bit odd with IE 5.0. But in general, you will benefit from faster loading and processing.

2006-06-21 - SONY Memory Stick Micro M2 1GB
In order to complete my set of memory card types once again, I ordered the SONY Memory Stick Micro M2 1GB MS-A1GA. Quick tests showed a maximum read performance of about 9.3MB/s and a maximum write performance of about 5.0MB/s. More test results will follow soon.

2006-06-01 - First Test Details for the NEXTO CF OTG ND-2300
The key figures for the performance of the NEXTO CF OTG ND-2300 are available. It will take some time to prepare the tables with the datailed test results.

2006-05-09 - First notes about the NEXTO CF OTG ND-2300
First notes about the NEXTO CF OTG ND-2300 are available.

2006-05-04 - NEXTO CF OTG ND-2300
Coming soon: Information on NEXTO CF OTG ND-2300.

2006-05-02 - Back to business...
After a long break, I finally managed to put my mind back on this topic.

2006-01-16 - Some Test Results
With increasing card speed, testing has become more and more frustrating. Currently, only very few card readers can break the 10MB/s barrier for CF, SD, MMC and MS. And as cards get bigger, the tests take more time. A complete test run for the 8GB CF on a formerly fast reader takes about two hours!
I added one test result for the 8GB Transcend 120x CF and three test results for the 2GB SONY MS PRO Duo. Some of the test results for the Memory Sticks are from card readers with a higher firmware version than the "officially" listed ones, which results in better performance for this media type.

2006-01-07 - SONY Memory Stick PRO Duo 2GB MSX-M2GS
Finally, the 2GB MS PRO Duo arrived. I already have the SONY Memory Stick PRO Duo High Speed 1GB (MSX-M1GN), but the results with the SONY Memory Stick PRO 256MB (MSX-256S) made me choose the non-high-speed version.
Amongst all my card readers, there are only three readers with outstanding MS performance, all of these using Genesys controllers. Quick tests showed that the regular MS Pro reads faster on these card readers than the High Speed version. The GL819 datasheet lists two modes for MS PRO: 15MHz and 20MHz. Maybe the readers only use the 15MHz mode. The read performance is around 13MB/s, more details will follow.

2005-12-28 - 8GB Transcend 120x CF
Today the order with the 8GB Transcend 120x CF arrived. Transcend claims read rates of 20MB/s and write rates of 18MB/s. This means that the card will be a big challenge for the card readers. As announced, the card also supports DMA in IDE mode, but it turns out that the supported IDE Modes are only PIO 4 and MDMA 2, which allow theoretical rates of 15.89MB/s in binary arithmetics.
There were three candidates amongst my card readers that could provide acceptable performance: The Hama 19 in 1 V.2, the NEXTO CF, and the Vivanco RW CF. It turned out that the Hama 19 in 1 V.2 using the GL819 controller is the only one that comes close to the advertized speed: Maximum rates of 17.55MB/s read and 13.65MB/s write. With the filesystem involved (FAT32, 32kB cluster size), these values drop to 17.50MB/s read and 12.51MB/s write.
According to the datasheet, the GL819 controller should be able to support 16-bit memory mode at 100ns and 80ns timing as listed in the CF 3.0 specification. These timings should give theoretical limits of 19.07MB/s and 23.84MB/s in binary arithmetics. The big question is whether the card does not go faster in real world, or other components affect the performance. But 17.5MB/s read is already a big improvement compared to the current 80x cards.

2005-12-14 - 512MB MMCmicro and a fast 128MB microSD
One week ago, the order with the 512MB MMCmicro arrived. But I was on a photo trip, so I unpacked it yesterday. The MMCmicro is number two in the race, the estimated delivery date for the 2GB MS PRO Duo is listed as end of January.
The order also included another microSD, which shows much better write performance than my current cards. Now I have suitable micro card types and can test the card readers properly.

2005-11-22 - 512MB microSD
The 512MB microSD arrived yesterday. Quick tests showed a read performance of about 10MB/s and a write Performance of about 3MB/s. Detailed results will follow next week, I am too busy at the moment.

2005-11-17 - Flash Card Memory Sizes
It's time for a little summary on flash card memory sizes. For Compact Flash, 8GB have become a standard size. Transcend offer their 8GB 80x CF at slightly less than twice the price of the 4GB version. CF-II seems to be a thing of the past, 8GB easily fit in a CF-I card.
SD cards have reached 4GB, which is the maximum capacity with the current SD card standard. However, my 4GB Transcend 150x SD even exceeds the maximum thickness of 2.1mm+0.15mm tolerance, and there are some slots where the card can only be violently removed. Physical size seems to be more of a problem than memory size for the 4GB versions.
MS PRO are advertized up to 4GB, but at very high price and low availability. MS PRO Duo are available up to 1GB, the 2GB versions have been announced long time ago, but the cards are not yet available. Chances are low that they will be delivered this year.
For xD-Picture Cards, there are no news after the 1GB Type M disaster. Availability is still critcal and prices went up again. I will not make any efforts getting cards or dedicated xD-Picture Card readers unless I get proper support from FujiFilm or Olympus. This card type is by far the worst and the most worthless amongst all current card types.
The smallest cards, MMCmicro and microSD are available at 256MB. For both types, 512MB versions are announced. But especially for microSD, space is critical. The cards are specified with a thickness of 1mm, but this includes the pull bar. The main body of the card has a thickness of only 0.7mm.
Currently, I have three cards on order: A 2GB MS PRO Duo, a 512MB microSD and a 512MB MMCmicro. Let's see which one wins the race.

2005-11-16 - NEXTO CF Firmware 1.3
There's no new firmware available right now, I'm just a little late. I did not find the time to prepare a new version of my NEXTO CF Special, so I decided add at least a few notes.

2005-11-15 - SD Card Size Limit Mystery
After having read many rumors about SD card size limits, I decided to investigate. It looks like all the problems with SD card sizes up to 4GB come from not having read the specifications at all, not having read them properly, or simply not being able to understand what is written there. Another problem might be that the specifications are not public, but there are enough SD card datasheets from card manufacturers available, where all the necessary details are listed.
The size of an SD card is given in the CSD register as block count. This block count is calculated from a card size base value and a card size multiplier. But there is also a value specifying the block size. With the default block size of 512 bytes, the maximum size is indeed 1GB. But the specifications allow a block size of up to 2kB (in powers of 2), that allows specifying a card size of up to 4GB. Those who missed the ellipsis dots between the values 9 and 11 might not accept a block size of 1kB, which is the common block size for SD cards of more than 1GB and up to 2GB.
The proper way to calculate the card size is the "simple" formula (C_SIZE+1) * 2C_SIZE_MULT+2 * 2READ_BL_LEN   where C_SIZE is a 12-bit value, C_SIZE_MULT is a 3-bit value, and READ_BL_LEN is from 9 to 11. That's 12 bit size, 2-9 bit from the multiplier and 9-11 bit from the block size, a total of 32 bit. The card itself uses a 32-bit byte address, which limits the size to 4GB as well, assuming unsigned values as usual.
The only true size limit for current SD cards is 4GB. All other "size limits" are due to improper implementations of the current specifications. Upcoming standards will introduce a block addressing mode instead of a byte addressing mode to extend the address space without having to change the card register layout.

2005-10-26 - SD Compatibilty Check, Part 2 finished
After my trip I was finally able to finish the last test. As expected, everything went fine.
Device  Adapter   2GB SD  4GB SD
NEXTO CFDataFab 5 in 1 CF OKOK
NEXTO CFMinolta SD CF OKOK
NEXTO CF4 in 1 PCMCIA & PCMCIA CF OKOK

2005-10-18 - SD Compatibilty Check, Part 2
To continue my SD compatibility test, I checked the SD cards with some SD to CF adapters and the NEXTO CF over FireWire. Here are the results:
Device  Adapter   2GB SD  4GB SD
NEXTO CFDataFab 5 in 1 CF OKOK
NEXTO CFMinolta SD CF OKOK
NEXTO CF4 in 1 PCMCIA & PCMCIA CF OK??*
* Unfortunately, I ran out of time while testing and I have to leave for a short trip now. The card is properly detected and I am pretty sure that it works OK. But the test to prove this runs about 1 hour 45 minutes.

Because there are still many statements about incompatbilties with many devices, I will prepare a detailed SD compatibilty page.

2005-10-16 - New Cards: miniSD 1.1 512MB and MMCmicro 256MB
These two cards complete my "card pool" for now. But Sony and SanDisk already announced the M2 - Memory Stick Micro, so the next challenge is already waiting.

Due to the growing number of cards, some changes in the layout were necessary. The test results for SD and MMC are split up on individual pages. The test results for multi card readers only list selected cards from each family, the other test results will only be shown on the detail pages. Colors have been added to the test results: Grey indicates that the speed was taken, but the verify (data integrity check) was not yet performed. When the read or write test was completed, but the verify showed errors, the figures will be shown in red. Whenever the test results table shows grey or red entries, you should check the detail pages for more comments.

2005-10-12 - ATP Pro Max SD 512MB
ATP Electronics, Inc. provided an ATP Pro Max SD 512MB. This is the first memory card contribution since I started this project. By the way: Don't forget to check the Credits Page from time to time.

2005-10-10 - SD Compatibility Check
After all the rumors about incompatibility with SD cards of more than 1GB, I decided to do a quick check with some of the older card readers that did not qualify for the new tests. There were less problems than I expected.

Reader  Controller   2GB SD  4GB SD
Anubis Typhoon 8 in 1smsc USB 97C223; FW 1.95 - - -- - -
CarryDrive FotobarCarry Jupiter Rev. 1.0 OKOK
DeLOCK 12 in 1NeoDio ND3260LA; FW 010D OKError
Hama 9 in 1 UpgradeICSI IC1210; FW 1.6A OKOK
Lexar RW018DataFab DF-UG-26; FW 0128 OKOK
PixoMedia 14 in 1 MSR-14Carry Winter Rev. 1.4 OKOK

2005-10-09 - Panasonic Pro High Speed SD 2GB RP-SDK02G
I finished the tests with the Panasonic Pro High Speed SD 2GB RP-SDK02G. The card shows about the same read performance as the Transcend card. Although the Panasonic card has a higher maximum write speed of 15.58MB/s (106.3x), it shows a significantly lower write speed when writing to the filesystem. This might be due to the slightly higher random access time. I tried to improve the FAT alignment, but this did not increase the write speed. In case I find out more about this speed drop, I will post a message here.

2005-10-08 - Vivanco RW CF USB2 and Vivanco RW SD USB2
Vivanco GmbH provided the single-slot plugin-type card readers Vivanco RW CF USB2 (18252) and Vivanco RW SD USB2 (18253). These card readers are not as small as other single-slot readers, but very stylish and very fast.

2005-10-07 - Test Results for the Transcend Ultra Speed 150x SD and the Panasonic CardBus SD adapter
The Panasonic CardBus SD adapter BN-SDPC3E is currently the fastest device for SD cards. A top transfer rate of 91.5% compared to the theoretical maximum does not leave much room for improvement. All test results for the CardBus adapter were taken on the old test system, a Notebook with native CardBus slot. At the moment, I have no comparison to the performance of the adapter with current Notebooks, so the write rates might be some percent lower than possible. Unfortunately, the PCI CardBus adapter I tried did not work in the new PC based test platform.

The full test procedure with the the Transcend Ultra Speed 150x SD TS4GSD150 revealed that there are random read errors with the pre-series sample of the Hama 19 in 1 V.2 reader. All the other USB 2.0 card readers shown here only suport SD 1.0 and therefore don't get above 10MB/s.

The reason for the delay in testing the card was because I used the Transcend 150x SD in my 1D Mark II during an event. I only shot 648 photos with a total of 1.72GB, but all of the images downloaded without errors. For a final statement about the FAT16 workaround solution, I have to wait until I get the chance to fill the whole card.

2005-10-01 - Transcend Ultra Speed SD 150x 4GB TS4GSD150
I received the order for a 4GB Transcend Ultra Speed 150x SD two days ago, 273 Euro including shipping and COD fees. Here are the results of my first tests:
Panasonic SD CardBus BN-SDPC3E
   Read: 21.81MB/s = 148x
   Write: 12.35MB/s =  84x
Hama 19 in 1 V.2 USB 2.0
   Read: 18.45MB/s = 126x
   Write: 12.20MB/s =  83x

I tested the Transcend Ultra Speed SD 150x 4GB TS4GSD150 in my Canon EOS-1D Mark II (Firmware 1.2.3). The camera won't accept the card right out of the box (FAT32, 4kB cluster size) and display "Card 2 not formatted". Formatting fails with the message "Cannot format / Change card2". Formatting the card on Windows XP to FAT16 with 64kB cluster size seems to do the job: The camera accepts the card, and a quick test across the whole card size with 16 test shots - 4 photos on each 1GB segment - was OK and the data used to fill the card was left untouched. But I can't give any guarantee for data integrity right now.

2005-09-14 - Finally: The New Test Results
This is only the basic structure of what I have in mind, but for the moment it is more important to present results rather than perfection. Especially the time-consuming and rather unspectacular test results for the microdrives and the mini-harddisk are skipped and will be added later. A quick summary of the most important topics:

The Hama 19 in 1 V.2 is a pre-series sample. I assume that the few problems that occured are only glitches of this specific sample, or that they will be fixed in the final version. This card reader is the first to support SD 1.1, MMC 4.0, and also shows very good read and write results for Memory Stick PRO.

The Hama Mobile MediaSafe VSV-40 is only a loan, but because of the good card reader test results, I decided to include them in the list. Attention: These are the test results for the card reader hardware. The card copy feature is handled by the internal CPU and is much slower. Detailed results for the copy performance will follow.

Actually, I wanted to include a big overview of USB stick sized card readers. But in the meantime, it has become very difficult to get test samples. But I am proud to present the Lexxington Mini-Card-Reader SD / MMC and MS / MS PRO. The card readers from this series are not only small, they are also by far the most beautiful card readers available. If these readers seem to be the slowest in the list, you should remember that all "slow" card readers have been removed from the new set of contestants. Compared to all card readers, they would be upper middle class in speed.

The C-enter 22 in 1 has the smallest base area amongst all multi card readers with about 65x67mm². The thickness of 20.6mm is due to the card compartment in the top cover, which can hold one SD oder one Memory Stick and one xD-Picture Card. This space can also be used to store either one MS Duo adapter or one miniSD adapter. Space enthusiasts could just remove the cover and the hinges to save space, but female photographers might prefer to use it as makeup mirror, thanks to the shiny surface. Based on the test results, this is the only card reader that shows good results for all media types. It is also the fastest card reader for the "old" MMC 3.31 cards by reaching the physical limit of 2.5MB/s.

A few notes concerning the website: For the test series with the new test system, only a selection of the fastest card readers was chosen. All previous test results can be found in the Test Archive pages.
The large Cross-Reference Table is now split up in one table for the multi card readers and one table for single-slot card readers. For a better overview and faster loading, the lists By Readers and By Cards have been split up: The results for the card readers can be found on individual pages, the results for the cards are split up by card families on four pages.

Step by step some attributes were added to the Memory Cards Overview. Brand new are the best values for read and write. These values are not necessarily the maximum values for the given card, but these are the best test results with the available equipment. Unlike the normal test results, which are taken by actually writing to and reading from the cards logical filesystem, these values are taken by diretctly writing to and reading from the media. Because this way of testing is not part of the usual procedure, the values will be taken step by step.
Next to the total storage capacity, the true CHS mapping is listed. In many cases, this is not equal to the CHS mapping used in the partition table. The values for sector offset and sector length of the partition, and the filesystem type and cluster size are usually the fatory settings. In some cases, these values have been modified for good reasons, and are then marked with * to mark the modified values. Another brand new attribute for CF compatible media is the list of supported ATA modes. This finally explains the non-linear behaviour of certain card readers, which don't work as fast as expected with PIO 2 cards.

Some cards have been added: Transcend 80x CF 4GB, SONY MS PRO 256MB, SmartMedia of 128MB and 64MB, and three MMC 4.0 cards: Two at first glance different MMCplus and one MMCmobile, all three from the same eastern manufacturer. Now I only need a fast SD 1.1 (120x-150x), a really fast MMC 4.0 (200x-266x) and one MMCmicro. A comparison between 512MB xD Type S and Type M would also be interesting, but currently there is a serious supply shortage. In the meantime, I spent my private budget for MS PRO and MMCplus instead.

2005-08-30 - Work in Progress
It will still take one or two weeks until I can present the latest test results taken on the new test system. Due to the increasing amount of data, the test results will be split up on several tables and pages instead of a few pages with huge tables. But this also requires lots of changes for processing the results.

2005-08-12 - New Test Results: Hama 19 in 1 V.2, Hama MMS VSV-40, NEXTO CF
Actually, I am in the transition phase from the old test system to the new test system. However, there are three important sets of test data available and I decided to add those to the old list before I completely switch over to the new test system and also to a slightly different layout. The three contestants are:

Hama USB 2.0 Card Reader 19 in 1 V.2 - Pre-Series Sample
This is a pre-series sample of the upcoming Hama card reader. The biggest problem is that I don't have any really fast cards available, but I think the test results already show that this reader has lots of potential. The only problem was that the SONY MS PRO Duo could not be read properly. The read rates were so low that the test tools got timeouts while reading. I think that this is due to the fact that this is a pre-series sample with lots of modifications on the board. According to other tests, the SanDisk MS PRO can be read at 15-16MB/s.

Hama Mobile MediaSafe VSV-40
This is a PSD with 40GB 1.8" Toshiba harddisk, 3-slot Multi-Cardreader and 2" colour display. Up to now I only tested the card reader part. It shows moderate results for CF, but excellent SD performance and very good performance with other media.

NEXTO CF ND-2500
When connected via FireWire, the CF slot of the NEXTO CF can be used as FireWire card reader. It shows excellent performance for Microdrives and the fastest CF cards, but only moderate perfomance for medium speed CF.

2005-08-05 - Hama 19 in 1 V.2 Card Reader
This USB 2.0 card reader will become the FireWire and CardBus killer.
SanDisk 2GB Ultra II CF: Read 10.99MB/s, Write 8.85MB/s.
Panasonic 512MB SD (RP-SDH): Read 9.77MB/s, Write 6.07MB/s.
Panasonic 512MB Pro High Speed SD (RP-SDK): Read at about 17.4MB/s.
SmartMedia Read ~6.5MB/s, xD-Picture Card Read ~6MB/s.
The only weakness is standard MMC Read/Write at about 1.6MB/s.
MMCplus is reported at 30MB/s read with this controller.

2005-08-04 - A-Data launched 4GB 150x SD Card on 2005-08-02
According to a news release from 2005-08-02, A-DATA Technology launched the 4GB 150x Turbo SD Card claiming 22.5MB/s read and 15MB/s write performance.
The step from 2GB to 4GB was very quick compared to how long it took from 1GB to 2GB.

2005-08-02 - Trying to Catch Up...
Well, a lot of things happened. Let's start with the flash watch. CF prices dropped to 50 Euro per GB. 4GB CF can be considered a standard size now, but 8GB CF are still hard to find and quite expensive. The 4GB Microdrive is available at 150 Euro, the 6GB at 185 Euro. 8GB Microdrives should be available soon. SD prices dropped to 55 Euro per GB. There are more 2GB SD cards available now, and Pretec even announced a 4GB SD card. MS Pro are available at 100 Euro per GB, MS PRO Duo at 110 Euro per GB. 2GB MS PRO Duo will be released soon. The xD-Picture Cards are available at 80 Euro per GB. 512MB xD are available as regular (SLC) and Type M (MLC), 1GB xD are only available as Type M (MLC).

I received a sample of the upcoming Hama USB 2.0 Card Reader 19 in 1 V.2. This card reader will set new marks for CF and SD speed. I don't have any fast CF availabe, but a quick test with a Panasonic RP-SDK SD showed 17.4MB/s read performance. More test details will follow soon. This card reader will be available in about 4 weeks at a street price in the range from 8-15 Euro. The package and the card reader will be marked "V.2".

I also received the Hama Mobile MediaSafe VSV-40 for testing. This is a mobile video player with 40GB 1.8" harddisk, multi card reader and colour LCD. A short review will follow.

And now something slightly off-topic. Panasonic came up with a new type of storage: The P2 Card at the size of a PCMCIA-II card contains 4 SD cards and a CardBus Interface. Like this, you get 8GB storage having 4 times the bandwidth of a SD card. Based on the transfer rate stated above, this would be 70MB/s! The new AG-HVX200 Camcorder offers two P2 slots with hot-swap feature for continuous recording.
Calculating only the price for the SD cards, 8GB would be about 400 Euro with standard SD media or about 800 Euro with high-speed media. However, actual P2 Cards will be much more expensive. Based on current technology, Toshiba could build a 30GB CardBus harddisk for less than 200 Euro. The question is whether a P2 Cam would accept such a harddisk, and whether the harddisk would provide the required sustained write rate. The slowest media zone of the MK3006GAL offers at most 9.56MB/s.

2005-07-05 - NEXTO CF Special Extended
The latest extensions of the NEXTO CF Special are worth mentioning. Now there are several photos added and reports about the 5 in 1 CF adapter as well as the external battery pack.

2005-06-02 - NEXTO CF Special Extended
I updated the NEXTO CF Special with the test details. Some details will follow, but these will become silent updates without mentioning them here.

2005-05-30 - NEXTO CF Special
The first version of my NEXTO CF report is online. Test details will follow.

2005-05-18 - Time Problems
I am back, but besides my reular job, I have to manage 2500 photos and 5 hours of video. In the meantime the order for new harddisks (one of those the Toshiba MK-6006) and the NEXTO CF including the 80GB Hitachi Travelstar 5K100-80 (HTS541080G9AT00) arrived. Now only the two enclosures are missing, then my mobile disk pool is up to date again. All in all it means that this project will slow down for a while.

2005-05-09 - Tevion 6 in 1 USB 2.0 Card Reader
Just a short note before I have to leave for a trip. Starting from today, it is available at some outlets of one of the most famous food discounters in Central Europe for sensationally low 5.99 Euro. The Tevion 6 in 1 Card Reader uses the ICSI IC1210 controller as M-Version with firmware 1.6E. If I interpret the datasheet correctly, the M-Version uses ROM instead of flash, which would mean that the firmware cannot be updated. The performance will be very close to the SilverCrest 16 in 1 and Anubis Typhoon 16 in 1, which use the same controller with firmware 1.6B.

The card reader is compact and lightweight. Notes on the slots: The xD-slot is not wired, so it does not work. The SD slot is deep, so you need an adapter if you want to use RS-MMC. The MS slot also takes MS Duo, but the card does not insert smoothly.

I did some research: The card reader is "manufactured" by SUPRA Foto-Elektronik-Vertriebs GmbH, www.traveler-service.de. There is a Tevion 8-in-1 card reader shown, which just seems to count CF-II and Memory Stick PRO as additional media. In the download are you can find broken links to a firmware update for an old 6 in 1 model from October 2003. The controller inside the "current" 6 in 1 is labelled week 50 of 2004, and the PCB is labelled 2005.02.17. So, this is not just to get rid of some old overstock items.

2005-05-06 - The Truth about Formatting
The first version of my Article about Formatting is available now.

2005-05-02 - Comments on the Panasonic Pro High Speed SD - And a bad surprise
As reported before, I had the Panasonic 1GB Pro High Speed SD for testing for a while. I was already prepared that the results won't be that thrilling. Some time ago, I had the chance to do a quick test with the 512MB version of that series and the fastest of my card readers, so I knew that the transfer limit of the readers was reached. Right now, the Panasonic CardBus SD Adapter seems to be the only way to make use of the actual speed of these cards - at least according to other test results. However, at a high price of 50 Euro.

What gave me cause for concern was the moderate write rate of the card. I already exchanged the USB 2.0 controller for a faster model some time ago, but the write rate stayed below my expectations. Since I have a PC with current components available now, I decided to do some cross-checks again. Reading at 8.694MB/s with the new PC compared to 8.633MB/s using the Notebook confirmed the accuracy of my read tests. However, the write rate of 7.558MB/s with the PC compared to 5.964MB/s using the Notebook was way off. While the Notebook writes to external USB 2.0 harddisks at 14MB/s, it seems to have problems with flash media. Further checks revealed that the difference is not that big with all cards, and that this problem only affects write rates of about 5MB/s or higher. This means, that I have to move over to a new test platform sooner than expected. Further information will follow.

A striking detail is that both cards offer even less capacity than usual. The 1GB card has a total of exactly 945MB (CHS 960/32/63) available, the 512MB card exactly half of that, i.e. 472.5MB (CHS 960/16/63) available. The FAT16 filesystems with 16kB cluster size offer 60460 clusters (944.6875MB) or 30224 clusters (472.25MB) of space available. So much for the urban legend that the filesystem takes up so much precious space.

2005-04-28 - xD Special Completed
The comments for the tests are added, which means that the 1GB xD special is now complete.

2005-04-27 - Panasonic 1GB Pro High Speed SD
I had the Panasonic 1GB Pro High Speed SD from a colleague for a while. One test series is finished. More comments follow.

2005-04-25 - Overview Controller Manufacturers and Controller updated
The Controller Manufacturers and Controllers Overview is updated. There's still a lot to discover, and for the second half of this year, transfer rates in the range from 10-20MB/s are announced.

2005-04-24 - 1GB xD-Picture Card Test Completed
The test series are completed. The comments on the test results will follow. Check out the 1GB xD article in the specials section.

2005-04-11 - 1GB xD-Picture Card
I could not resist the temptation and ordered a 1GB xD-Picture Card. I knew that the tests will be exciting, but I did not expect things to become that thrilling. At first I noticed that the card came with an ambiguous partition table, which means that the CHS and LBA entries did not match. This was not really a problem, since the partition definied by the smaller LBA entry was large enough to hold the file system, but usually both entries should match.
First tests showed that the card reads a lot slower than the 16MB and 64MB cards from my collection, but offers a slightly higher write speed. In order to get a quick overview, I tested some combinations of card, adapter and reader by checking only 32MB.

The fastest rates were about
2.6MB/s read and 1.9MB/s write for native xD slots,
3.0MB/s read and 2.2MB/s write with xD to SmartMedia adapter, and
2.9MB/s read and 1.6MB/s write with xD to CF adapter.

The card was not recognized by some readers with dedicated xD slot, but instead worked properly in some card readers without dedicated xD slot using the xD to SmartMedia adapter. During my quick tests, the card worked well in all but one card reader using the xD to CF adapter. This means that the xD to CF solution is the most compatible solution right now. Almost all card readers showed similar results with the xD to CF adapter as stated above, except the ones with GL816E or Carry Win II controller, which performed at only 1MB/s read and 0.6-0.9MB/s write.

Detailed results will follow. But due to the low transfer rate and the combinations with adapters, the test series will take a very long time.

2005-04-11 - News Flash: Dropping Prices
I just noticed that the 1GB SONY Memory Stick PRO Duo HS dropped from 195 Euro down to 129 Euro, most likely due to the availability of the 1GB SanDisk Memory Stick PRO Duo at 115 Euro. Another price drop happened with the release of the 1GB Olympus xD-Picture Card, which sells at 95 Euro now. This means that the two "proprietary" card formats are now available in high capacities (compared to their physical size) and at reasonable price. Please note: All prices are taken from reputable sources that are able to actually deliver the cards in time and at no or reasonable shipping costs.

Now it looks like the SD card manufacturers take over the former role of the Memory Stick PRO manufacturers and stay way behind current demands in memory size. I read that the first 2GB SD cards have been shipped, but I would guess that it will take some months until 2GB SD cards will be available from more manufacturers. And many high-speed cards are still only available at up to 512MB. What a shame if you see that the regular Memory Stick PRO is available with 4GB and that even 1GB miniSD cards are annouced. Those people who declared the CompactFlash "outdated" should shut up until 8GB SD cards are available.

2005-02-28 - Test results for Hama 19 in 1 (55114) and Hama 22 in 1 (55116)
I was able to finish the tests with these two card readers. They come with firmware 1.8D installed, which fixes some compatibility issues with certain USB 2.0 host chipsets. Unfortunately, there is a little loss in performance due to this compatibility tuning. Hama is working on a new firmware that is supposed to ensure compatibility and performance.
I had the chance to test a beta version of Firmware 1.9C which already increases SD performance, but we have to wait until a final version is released.

2005-02-16 - ICSI Firmware Multitude and more Translations
I finally found out that there are more ICSI Firmware Releases than I thought, and that the different releases can show quite different performance. The good part is that I was able to find out where to check the firmware revision of a card reader, if the information is provided. This means that I added a new colums to the Card Readers Overview. And I finally get to know why the RW020 with the DataFab controller shows the same performance as the readers with NeoDio controller.
I also finished the translations for the individual comments, which means that the translation process is finished.

2005-02-11 - MS Duo Update and Translations
I finished most of the test series with the 1GB SONY MS PRO Duo HS, except for the readers using slow 1-bit access. This card was a big disappointment, since the name indicates high speed. The read performance is almost acceptable: The Enermax 7 in 1 offers the top rate with 6.89MB/s, the next best bunch of readers is at around 6.5MB/s. But there seem to be problems with the write performance, the top rate is at poor 1.76MB/s. The readers that obviously only supprt 1-bit transfers read at about 1.4MB/s and write at about 0.8MB/s.
In the meantime I finished the German translations for all of the pages, except for the comments on the individual test notes. I hope that there are no significant errors due to the changes.

2005-02-08 - SD Tests: Trouble and Surprise
I borrowed the SanDisk SD "Standard" and SanDisk SD Ultra II from my colleagues again to test these cards with the new readers. The testing took more time than I thought because of two incidents.
First the good, but strange news: The cards did not produce the random errors I noticed during my last tests. I repeated the tests several times with both, the old and the new CardBus USB 2.0 controller, but everything was OK.
Now the bad, but interesting news: I had severe problems with the write protect "feature" with some readers. Some time ago I already super-glued the write protect switch of the Transcend 1 GB SD to the unlock position, because it got loose. But now it turned out that even some SD slots can get worn out after only a few cycles of operation and therefore cannot detect the position of the switch properly. I took a closer look at the different SD slots and found out that there are four basic types of write protect contacts used: One type at the lock position and the other at the unlock position. One type has the contacts open and will short the contacts, when there is no notch. The other type has contacts shorted, and will open them when there is no notch. Depending on what type is used, you can try to bend the contacts in the appropriate direction or shortcut or open the pins permanently.
I always knew that this type of write protection is lousy engineering work, so let's hope that at least the camera manufacturers use proper SD slots. Just imagine that the SD slot in your camera gets stuck at the lock position...

2005-02-02 - Hama Mobile DataSafe On-The-Go
I also received a pre-series model of the upcoming Hama Mobile DataSafe On-The-Go. This is a battery powered USB 2.0 20GB harddisk with USB 2.0 Host feature that can connect to virtually any USB "Storage Class" device using FAT file system (FAT 12, 16 and 32 are supported). The special feature is that you can copy in both directions to or from any drive and therefore can use multi card readers, too.
This means that you can copy files from a Digital Cameras that connects as storage device, copy files from any slot of a card reader, and copy from the internal disk to an external USB disk.

2005-02-02 - Hama TravelDrive 22 in 1 USB 2.0
The Hama TravelDrive 22 in 1 is the latest "deluxe" model. It has a more appealing finish with its rubberized shell and the package also includes a CD with Ulead PhotoImpact XL SE.

2005-01-26 - Hama 19 in 1 USB 2.0 Card Reader
The Hama 19 in 1 is the latest "budget" model. It features the improved Memory Stick slot (with perfect fit for regular sized MS and MS Duo) and the latest SD slot that takes SD, miniSD, MMC and RS-MMC in one single slot. The miniSD has to be inserted upside down along the left edge of the slot.

2005-01-14 - eMail from Hama
I knew that this would happen sooner or later. I received an eMail from a product manager of Hama concerning my website. Up to now I had to contact the manufacturers, and the response quota was absolutely disappointing.

2005-01-25 - MS PRO Duo HS 1GB
I received the SONY Memory Stick PRO Duo HS 1GB that I ordered for my new camera. The first tests were very disappointing, because the card only showed acceptable write performance during the first write cycle. All subsequent write operations were and are slow. I observed this effect with other media, too. However, the difference between first and subsequent writes was small (less than 10%). I have no explanation for this phenomenon, but for this Memory Stick the ratio was 2.75. The poor write performance also means that it will take a lot of time to finish the test with all my card readers.

2005-01-20 - MS Duo CF Adapter
As a preparation for the Memory Stick Pro Duo that I ordered for my DSC-T3, I bought the SONY Memory Stick Duo CF-I Adapter (MSAC-MCF1N). This adapter has the size of a CF-I card, and the Memory Stick Duo can be inserted fully into the adapter. This means, that with the help of this adapter, you can theoretically use MS (Pro) Duo in any CF-I device. My experiments will show, whether there are any compatibility issues.

2005-01-20 - Results of the "Battle of the Giants"
Finally, I found the time to finish the last two tests. This means, that all test results for the five new card readers are available. Those are my first test results using the new CardBus USB 2.0 Controller. Previous test results might differ slightly due to the new controller.

Summary
I have five new card readers which use three different controllers.

Analysis of the Results

An interesting fact is that none of the five contestants shows top performance with MMC media. They all stay about 20% behind the readers using ALi, NeoDio or smsc controllers. If you have many MMC, but limited time, you have to get a card reader using one of these controllers.

Summary

Except for not supported media types or lack of current for the non-standard 1.8" harddisk, none of the readers failed with any of my cards. I will do the tests with the borrowed cards some time, at least for the cards that I can get easily. But the only interesting test results will be those with the SanDisk standard and Ultra II SD cards.

Those of you who don't need maximum SD performance have free choice amongst these readers. Then you can choose according to the specific details like the supported media types. Especially the DeLOCK 19 in 1 and the Anubis Typhoon 16 in 1 support a wide range of media types. The retec 10 in 1 from my previous readers has the unique miniSD and Memory Stick Duo slot.

My personal favourite is the C-enter 12 in 1. It is by far the smallest multi card reader, and takes all supported cards in their natural orientation (i.e. not upside down). It has excellent build quality: The two aluminium shells are snapped to the solid base frame and are attached with double-sided tape. The size is achieved by spreading the four slots across three sides of the frame. This is a little unusual, but easy to handle, once you are used to it. As special accessory, there's a docking station included for quickly changing from on-desk to mobile use.

It looks as if the battle for the top position is more about supported media taypes and ease of use than about speed. Now we only need individual status LEDs for each slot (as the DeLOCK 12 in 1 has), then we only have perfect card readers?

2005-01-12 - New Adapters
Thanks to a hint in a forum, I was able to get two adapters at very low price: A Memory Stick adapter in shape of an extended CF-I card, and a SmartMedia adapter at the size of a CF-II card (however, the SmartMedia sticks out). The price was only 5 Euro each.
Besides that, I am busy with other stuff, and there are still two test runs with the 4GB MD missing.

2005-01-03 - Memory Stick Duo
From now on I have a Memory Stick Duo available. Now I can fully judge the media type compatibility of the card slots.

2005-01-02 - Three Sets of Test Data online
What does a crazy man do on new year's eve and new year's day? Testing card readers!
Although the DeLOCK 19 in 1 with the Carry WIN II Rev. 1.0 and the C-enter 12 in 1 Aluminium with the Genesys GL816E are using different contollers, the results are surpsrisingly close to each other. For CF, the Anubis Typhoon 16 in 1 with the ICSI IC1210 is sometimes a little faster, but due to the new Firmware version, it is slower with the fast CF cards. As expected, it stays behind the opponents with SD cards, and seems to read the TransFlash card in MMC mode. Surprisingly, none of the three contestants shows outstanding performance for MMC. The card readers with NeoDio ND3260-LA, ALi M6535 A1 or smsc USB97C2xx are about 20% faster.
More details will follow when I'm done with the other test series.

Just in case some of you did not notice: From time to time I add new columns to the tables. Since the card manufacturers are more and more abusing the CD-ROM speed rating, I also display this value in the detailed tables (by readers / by cards).

2004-12-31 - 128MB TransFlash Card
I received the order with the 128MB TransFlash card. Now my collection of tiny cards is complete: RS-MMC, miniSD and TransFlash.

2004-12-30 - Preparations for the "Battle of the Giants"
Right now I am in the middle of the preparations for the "Battle of the Giants". There will be a total of six or seven contestants with three different controller types. It will become a tough competition and they will make the other readers turn pale.

2004-12-28 - Four new card readers
I received a package of four card readers from Reichelt Elektronik.

2004-12-24 - A new card reader: C-enter 12 in 1 Aluminium
I received the C-enter 12 in 1 Aluminium USB 2.0 from PEARL. It uses the powerful GL816E and is currently the smallest multi-card reader due to the slot arrangement on three sides. It also features a docking station for on-desk use.

2004-12-23 Links to the new Frame Layout added
I decided to add a link to the new frame layout without changing all of the current pages first. This means, that the headings will be inconsistent, but you can benefit from the new layout right now.

2004-12-22 New Frame Structure
Yesterday I completed and uploaded the first version with frame layout and fixed navigation menu.

Previous Highlights
I did not make any specific notes during the last year, but I will try and summarize the important steps based on my postings.

Hans-Jürgen Reggel   ·   http://www.hjreggel.net/cardspeed/   ·   2004-12-21 ~ 2007-01-25