Thinkpads Forum

Posted: **Sat Nov 12, 2011 12:29 pm**

EOMtp wrote:Won't one of these media readers that fits (fully) in the PC Card slot do the job?
http://www.ebay.com/sch/i.html?_nkw=%28 ... m270.l1313

No, these are the very old slow 16bit PCMCIA card versions which should be avoided at all cost, especially since these most probably won't support any modern SDHC cards with reasonable capacities. Check out wikipedia.org/wiki/PCMCIA to learn about the technical differences between PC-Card/PCMCIA and Cardbus. These 32bit CardBus cards can be distinguished from old 16bit PC-Card/PCMCIA cards by the presence of a gold band with eight small studs on the top of the card next to the pin sockets. The devices found with the above cited eBay search lack the typical signs of a proper 32bit Cardbus adapter and thus are obsolete.

Better make sure to choose some 32bit Cardbus Adapter SDHC card reader for the T61 PC-Card slot like this one here: ebay.com/itm/300467818479. I bought multiple of these very devices for quite a few T61 machines of my own and my friends, and we are all very happy users of these. The biggest SDHC card i used so far with these has 16GB. I would have tried a 32GB one, but as long as buying two separate 16GB cards is still cheaper than buying one single 32GB card, testing one of these is out of question for me.

Posted: **Sat Nov 12, 2011 1:01 pm**

This one here clearly states it supports 32GB SDHC, and posts reasonable benchmark speeds. Do you think they would be downright lying?
http://www.ebay.com/itm/SDHC-XD-SD-MS-P ... 43aafd9e4a

Posted: **Sat Nov 12, 2011 2:18 pm**

dr_st wrote:This one here clearly states it supports 32GB SDHC, and posts reasonable benchmark speeds. Do you think they would be downright lying?
http://www.ebay.com/itm/SDHC-XD-SD-MS-P ... 43aafd9e4a

I wouldn't assume them to be lying, but since it can be doubted that an old 16bit PCMCIA-Adapter (on the photo there is no gold band with eight small studs on the top of the card next to the pin sockets) would be able to manage any such SDHC card, i'd rather say they are plain wrong. Even if this adapter would be able to manage these cards, which i doubt, it would be far too slow to be practical. This particular seller also offers the same Cardbus card i mentioned before: ebay.com/itm/220661093554. Better get that one instead and don't waste any money on any old 16bit PCMCIA adapter card.

Posted: **Sat Nov 12, 2011 7:37 pm**

rumbero wrote:... these are the very old slow 16bit PCMCIA card versions ...

The bandwidth of even 16-bit PCMCIA is 20MBytes/sec, which is higher than the maximum read/write bandwidth of all except the fastest flash memory cards (e.g., SanDisk Extreme can sustain 25MB/s; all others are under 20MB/s). Therefore, even if these card readers were 16-bit cards -- which they are not; they are 32-bit cards which support SDHC, as advertized -- speed should not be an issue.

Posted: **Sat Nov 12, 2011 7:51 pm**

EOMtp wrote:Therefore, even if these card readers were 16-bit cards -- which they are not; they are 32-bit cards which support SDHC, as advertized -- speed should not be an issue.

May i take it for granted that you own one of these card readers (lacking the gold band with eight small studs on the top) and thus are able to practically validate these assumptions? You wouldn't be the first one to regret wasting money for obsolete hardware based on false information.

I'd suggest to play safe and make sure to only acquire a proper Cardbus adapter with this gold band with eight small studs on the top, in order to avoid unpleasant surprises. The direct eBay links i provided in my former replies are known good Cardbus SDHC card readers with the mentioned properties.

Posted: **Sun Nov 13, 2011 2:07 am**

Is there a way to validate, really validate, whether a given adapter is 16-bit or 32-bit PCMCIA?

I have some AKE USB2.0 CardBus adapters which lack the gold strip but say 32bit PC Card.

Edit:

OK, I found a simple way, using CrystalDiskMark benchmarking tool.

I tried my Corsair Voyager Mini USB drive on my T60, once through the integrated Intel USB controller, and once through the AKE PCMCIA card (which uses NEC chipset).

There are noticeable difference in read rates - 20-25MB/s on the onboard, 10-12MB/s on the PCMCIA.

I would not assume the NEC controller to be so bad. It seems to suggest that indeed the PCMCIA card is not full 32-bit, and that in practice 16-bit PCMCIA cannot reach the full 20MB/s speeds.

Now, that does not mean that 16bit PCMCIA cards will not support high capacity SDHC at all. In fact, if they say they support them, they probably do. However, you will probably not realize the full potential of the SDHC cards as far as speeds go.

Posted: **Sun Nov 13, 2011 10:30 am**

I realized that we have gone somewhat off-topic.

Mods/admins, could you split the thread so that everything starting from EOMtp's post is in a separate topic (probably something along the lines of "Card readers and Cardbus adapters", and probably in the general HW/SW forum)?

Posted: **Sun Nov 13, 2011 5:42 pm**

dr_st wrote:Is there a way to validate, really validate, whether a given adapter is 16-bit or 32-bit PCMCIA?
I have some AKE USB2.0 CardBus adapters which lack the gold strip but say 32bit PC Card.

If it appears as a PCI device behind the CardBus bridge, then it is a 32-bit CardBus device. (To check in Windows, use devmgmt.msc.)

There do in fact exist USB host controllers in ISA cards, but I do not know about PCMCIA PC Card.

dr_st wrote:I would not assume the NEC controller to be so bad. It seems to suggest that indeed the PCMCIA card is not full 32-bit

Yes, the NEC controller, in a CardBus package, IS that bad. You have a CardBus card (32-bit). (Even if the design was so cheap, that it left out the gold band.) For these nameless (in my eyes "AKE" reads "Generic") CardBus-to-USB cards, the speeds you observe are normal. "Normal", where "normal" is reports from just two other people. Although countless of these things were made and sold, just frustratingly few people write about actual performance.
http://forum.thinkpads.com/viewtopic.php?f=28&t=93951

EOMtp wrote:The bandwidth of even 16-bit PCMCIA is 20MBytes/sec, which is higher than the maximum read/write bandwidth of all except the fastest flash memory cards

Is that so? I did not know PC Cards can be so fast. Please cite a source for that number.
Maybe the cheap mass-storage sticks used like business cards as give-aways are slow. I think that the average retail USB flash stick in production today can reach 30 Mo/s transfer speed.

Posted: **Mon Nov 14, 2011 10:04 am**

automobus wrote:If it appears as a PCI device behind the CardBus bridge, then it is a 32-bit CardBus device. (To check in Windows, use devmgmt.msc.)

It definitely appears as a PCI device behind the CardBus bridge, but does it really mean it's 32bit? I believe that the CardBus is backwards compatible with 16bit PCMCIA cards, and I would still expect any such device to appear behind the CardBus controller when viewed "by connection".

Furthermore, I did not see any reference to PC Card being implemented behind an ISA bus. The reference I found is "The original standard was built around an 'enhanced' 16-bit ISA bus platform" (http://en.wikipedia.org/wiki/PC_Card). But so was ATA for that matter (http://en.wikipedia.org/wiki/Industry_S ... urrent_use). Nonetheless, it doesn't mean that ATA devices anywhere sit behind an ISA bus.

Furthermore, what Windows displays in the PCI section is mostly related to the way that these devices expose themselves to the OS. Everything that exposes a PCI config space will be there (including PCI-e, AGP, and even some devices with a dedicated bus of their own, like some network controllers on some desktop boards).

What I'm saying is that the way the Device Manager or lspci display devices is probably not a good indicator of what the devices actually are.

automobus wrote:Yes, the NEC controller, in a CardBus package, IS that bad. You have a CardBus card (32-bit). (Even if the design was so cheap, that it left out the gold band.) For these nameless (in my eyes "AKE" reads "Generic") CardBus-to-USB cards, the speeds you observe are normal. "Normal", where "normal" is reports from just two other people. Although countless of these things were made and sold, just frustratingly few people write about actual performance.
http://forum.thinkpads.com/viewtopic.php?f=28&t=93951

With this I agree, and I now retract my previous statement. In fact, some pictures in that thread show the same AKE USB card with the CardBus gold stripe, and reporting similar speeds. I can't imagine they actually have separate production lines for 16bit and 32bit cards. More likely that they are all 32bit (otherwise they would be really lying), but sometimes they "go cheap" as you said, and skip the gold stripe.

It is a bit disappointing to know that these USB controllers are so much inferior to the integrated ones, but I guess it does make sense to use only the cheapest for this kind of devices. Generally, NEC USB controllers are quite good (in fact, they were pioneers both in USB2 and USB3), but obviously the high-end ones will not be used in such adapters.

automobus wrote:Is that so? I did not know PC Cards can be so fast. Please cite a source for that number.

One source would be this: http://www.quatech.com/support/comm-over-pcmcia.php. Of course it is only theoretical, and I do not know how much it can actually deliver. With USB2 for instance, even the best controllers out there can barely sustain just over 50% of the theoretical bus limit.

Posted: **Mon Nov 14, 2011 10:42 am**

dr_st wrote:It definitely appears as a PCI device behind the CardBus bridge, but does it really mean it's 32bit? I believe that the CardBus is backwards compatible with 16bit PCMCIA cards, and I would still expect any such device to appear behind the CardBus controller when viewed "by connection".

If it appears as a PCI device, then yes, it is 32-bit. PCI is 32-bit. If it appears as something else, as a non-PCI device behind CardBus bridge, then one might conclude it is not CardBus, not PCI, not 32-bit.

dr_st wrote:Furthermore, what Windows displays in the PCI section is mostly related to the way that these devices expose themselves to the OS. Everything that exposes a PCI config space will be there (including PCI-e, AGP, and even some devices with a dedicated bus of their own, like some network controllers on some desktop boards).

Exactly. AGP, PCIe, and CardBus are based on PCI. Just like PCI devices, they have a PCI class, vendor ID, and device ID. PC Card is not based on PCI—a card does not have PCI IDs

edit to point out: PC Card pre-dates PCI.

Posted: **Mon Nov 14, 2011 11:47 am**

dr_st wrote:In fact, some pictures in that thread show the same AKE USB card with the CardBus gold stripe, and reporting similar speeds. I can't imagine they actually have separate production lines for 16bit and 32bit cards. More likely that they are all 32bit (otherwise they would be really lying), but sometimes they "go cheap" as you said, and skip the gold stripe.

I'd rather assume that these sellers simply sell old stock, for which they never participated in any production. I have been burned enough by some eBay deals to distrust any vendor description that doesn't match the expected item in every regard. So, if the pictures used to illustrate the offer are not showing the expected technical characteristics of a gold strip with 8 dimples for something sold as a proper CardBus card, the rest of the description simply does not match it anymore. I would not bother knowingly buying anything like that, especially if an offer with a coherent description like in 300467818479 or also 220661093554 is readily available.

Posted: **Mon Nov 14, 2011 12:56 pm**

dr_st wrote:I have some AKE USB2.0 CardBus adapters which lack the gold strip but say 32bit PC Card.

I wish I thought to ask you this sooner: what is the voltage key on that card?

Posted: **Mon Nov 14, 2011 2:53 pm**

rumbero wrote:I'd rather assume that these sellers simply sell old stock, for which they never participated in any production. I have been burned enough by some eBay deals to distrust any vendor description that doesn't match the expected item in every regard. So, if the pictures used to illustrate the offer are not showing the expected technical characteristics of a gold strip with 8 dimples for something sold as a proper CardBus card, the rest of the description simply does not match it anymore. I would not bother knowingly buying anything like that, especially if an offer with a coherent description like in 300467818479 or also 220661093554 is readily available.

"They" in this case are not the sellers, but whoever manufactured the devices.

I am not saying your advice to play safe on eBay is bad, in fact it's very good.

All I am saying is that it is probably wrong to assume that just because a device does not have the gold strip, it is not 32bit Cardbus. Because evidently there exist cards that show the same hardware characteristics, even though some have it, and some don't.

automobus wrote:I wish I thought to ask you this sooner: what is the voltage key on that card?

How would I go about checking that?

Posted: **Mon Nov 14, 2011 8:13 pm**

dr_st wrote:How would I go about checking that?

One picture says it better than many words, and Wikimedia has a good picture.
http://commons.wikimedia.org/wiki/File: ... _notch.jpg

Posted: **Tue Nov 15, 2011 12:33 am**

Well, to tell you the truth, it looks somewhere in between the two, but more like the 32bit version.

Posted: **Fri Feb 10, 2012 3:57 am**

I read as much as I could comprehend about PC Card. I wanted to understand why real world performance is so low. I am not an expert, and I have no formal education of computer engineering. Do not take my words as facts. If you know better than I do, then please correct my errors.

Quiz yourself! Spend a minute or two thinking about these questions.

What is the correct name: PCMCIA, PC card, 16-bit PCMCIA Card…?
How fast (or slow) is 16-bit PC Card, really?
Why does 16-bit PC Card i/o have such high CPU load?
Can performance be increased?

What is the correct name: PCMCIA, PC card, 16-bit PCMCIA Card…?

I think it is important to know the names of these technologies. PC Card, as a generic term, refers to the standards and all sorts of devices. The first few versions of the standard were named "PCMCIA Standard". In 1995 the CardBus interface was created, and the name of the standard was changed to "PC Card Standard".

PC Card Standard, paraphrased wrote:PCMCIA and JEITA have developed a standard for a credit card-sized adapter, called a 'PC Card' that does for notebook and other portable computers what the AT bus did for desktop PCs.

When referring to products that support 16-bit operation, the term "16-bit PC Card" should be used. "CardBus PC Card" is the correct term that can be used when referring to the 32-bit bus master specification of the PC Card Standard.

How fast (or slow) is 16-bit PC Card, really?

As EOMtp and dr_st pointed out, 16-bit PC Card has a theoretical maximum speed of 160 megabits per second (20 Mo/s). I did not believe that number was true at first. Based on my experience, I thought 2 Mo/s was more likely! I can believe that some specific combinations of card, host, and software might exist and have high performance. But in fact, 16-bit PC Card usually has very low performance. It is slow and hogs processor time.

David Hinds, Linux PCMCIA HOWTO wrote: http://pcmcia-cs.sourceforge.net/ftp/do ... html#ss3.3

16-bit PCMCIA cards have a maximum performance of 1.5-2 MB/sec.

Why does 16-bit PC Card i/o have such high CPU load?

16-bit PC Card can operate in two basic transfer modes: programmed input/output (PIO), and memory mapped input/output (MMIO).

During PIO, the driver software reaches out to the PC Card continuously. In a multi-threading computer, other instructions can still be executed. In a single-threading computer, all other tasks suffer.

During MMIO, the PC Card Host Adapter does i/o to a slow PC Card. When a PC Card is accessed, the driver software communicates data with the PC Card Host Adapter. The Host Adapter does the "dirty work", while the system CPU can smoothly do other work.

Let me make an analogy. Imagine a small town that represents a computer. The main street is named PCI Lane. The daily newspaper is delivered by a youth named Driver. Driver can deliver most of the papers quickly: ey rides a bicycle on PCI LN, pulls a paper from eir bag and throws it to the door of every home and office. The bicycle is fast and does not stop any traffic in town. This efficient way of working represents MMIO. But Driver needs to make one annoying delivery, to a senior citizen named Peacey. Peacey's house is much further away from the street than the other buildings. Driver cannot just toss the paper into Peacey's yard, because Peacey is too old to bend down and retrieve paper from the ground. And Peacey forbids Driver from riding eir bicycle on the lawn! So Driver parks the bicycle in the street, walks the path to the front door, and deposits newspaper in the mailbox. During this delivery, street traffic is halted. This slow way of working represents PIO.

Can performance be increased?

To improve performance by offloading work from the CPU, MMIO is needed. Then, when the CPU is not doing the dirty work, i/o frequency should be increased (a little closer to that mythical 10 MHz). I found failed attempts at improving performance, again for Linux PCMCIA. If they cannot get it working, then I doubt anyone will ever do it for Windows or Macintosh.

http://lists.infradead.org/pipermail/li ... 04473.html
http://lists.infradead.org/pipermail/li ... 05249.html

I end this write-up with comic relief. A quote by Linus Torvalds on drivers.

Linus Torvalds wrote:For a driver writer, there is one rule above _all_ other rules:

"Reality sucks, deal with it"

That rule is inviolate, and no amount of "I wish", and "it _should_ work
this way" or "…but the documentation says" matters at all.

Bonus: where can one find a copy of the spec?

On the disgusting polluted Internet of today, good information can be more difficult to find than to understand! Here is goods. Pass it on. To the people at affron.narod.ru: Thank you for hosting the Specification!

http://affon.narod.ru/pcmcia.html

content revised 2012-02-11 to add the Linus quote
wording edited 2012-05-02

Posted: **Fri Feb 10, 2012 4:19 am**

Thank you for the very good write-up!

It's amazing how some information is so hard to find and organize even in today's internet.

Posted: **Thu Oct 18, 2012 4:13 pm**

Reviving this old thread with some more information.

I went ahead and purchased one of the cheap PCMCIA-SDHC adapters for testing purposes:

Analysis:

Looking at the card, it lacks the gold stripe with 8 bumps, however the notch is that of a 32bit CardBus, as shown on Wikipedia Cardbus Reference.

It appears in Device Manager (Thinkpad T60) as follows:

Code: Select all

PCI Bus --> Intel 82801 PCI Bridge --> Texas Instruments PCI-1510 Cardbus Controller --> VIA USB Universal Host Controller + VIA USB Enhanced Host Controller

So it looks to be a simple USB Hub, with one of the ports being hardwired to the SDHC interface.

Unlike the Sandisk ExpressCard SDHC Adapter, which always registers as a connected USB Mass Storage Device and gets a drive letter assignment, even without a card installed, this CardBus controller just looks like an empty hub, until you connect a card. At that point, the card shows as a Mass Storage Device connected to one of the USB ports in the hub, and only then gets a drive letter assignment.

Performance:

I have the Sandisk ExpressCard adapter in the same computer to test against, and the only SD cards I have accessible is some old el-cheapo 128MB MicroSSD with an adapter, and a Sandisk Ultra II 4GB SDHC card.

Testing Sequential R/W using CrystalDiskMark I got the following results:

El-Cheapo 128MB + Sandisk ExpressCard Adapter: Read = 10MB/s, Write = 3MB/s
El-Cheapo 128MB + VIA CardBus Adapter: Read = 5MB/s, Write = 2MB/s
Ultra II 4GB + Sandisk ExpressCard Adapter: Read = 18MB/s, Write = 10MB/s
Ultra II 4GB + VIA CardBus Adapter: Read = 11MB/s, Write = 8MB/s

Connecting the said cards directly to the onboard USB ports of the machine (through a USB-SDHC adapter) leads to speeds comparable to the Sandisk ExpressCard.

So it looks like the Via CardBus adapter is quite a bottleneck. I don't know if the cause is the interface or the USB controller, but assume the latter. Interestingly, it can break 10MB/s read speeds with the 4GB SDHC card but not with the 128MB SD, even though the SD card clearly supports such speeds. I believe the performance is therefore depends on a combination between the theoretical throughput of the controller and the flash density.

When I get around to it, I may try some higher capacity SDHC cards, and I am also considering ordering one of the more expensive CardBus adapters for comparison. Maybe also someone who owns one of these (rumbero?) can post some performance benchmarks and information (for instance, I'd like to know whose controller they are using).

Posted: **Tue Jan 08, 2013 3:37 am**

An update:

I ended up buying the SDCBA-C01 Cardbus Adapter, which has been highly recommended here and in other places.

One immediately sees that it is a higher grade product. The packaging and the rigidity of the card itself speak for it. Performance is also higher, and the adapter and SD card sit truly flush with the chassis, which was not quite the case with the previous adapter (and in fact, was the cause that forced me to buy this one).

It is also a pure PCI device - it does not implement a USB controller between the CardBus and the SDHC, but appears in the Device Manager as a "SDA Standard Compliant SD Host Controller" under "SD Host Adapters", much like the built-in card readers on modern Thinkpads.

However, all the previous numbers I reported in this thread regarding the performance of the cheap Cardbus SDHC/USB adapters should be taken with a grain of salt. The reason is that they were heavily influenced by the T60 Cardbus performance issues under WinXP, which I only later noticed and managed to solve.

In the process, I learned a good deal about some of these adapters and controllers, and will post a more detailed analysis of these old toys at a later time.

Edit: Forgot to update this post with the link to the analysis. Here it is:
http://forum.thinkpads.com/viewtopic.php?f=18&t=111973

Thinkpads Forum

PC Card, Cardbus, SD Card adapters, readers & performance

PC Card, Cardbus, SD Card adapters, readers & performance

Re: PC Card/4-in-1 media reader slot in 4:3 LCD T61 - possible?

Re: PC Card/4-in-1 media reader slot in 4:3 LCD T61 - possible?

Re: PC Card/4-in-1 media reader slot in 4:3 LCD T61 - possible?

Re: PC Card/4-in-1 media reader slot in 4:3 LCD T61 - possible?

Re: PC Card/4-in-1 media reader slot in 4:3 LCD T61 - possible?

Re: PC Card/4-in-1 media reader slot in 4:3 LCD T61 - possible?

Re: PC Card, Cardbus, SD Card adapters, readers & performance

Re: PC Card, Cardbus, SD Card adapters, readers & performance

Re: PC Card, Cardbus, SD Card adapters, readers & performanc

Re: PC Card, Cardbus, SD Card adapters, readers & performance

Re: PC Card/4-in-1 media reader slot in 4:3 LCD T61 - possible?

Re: PC Card, Cardbus, SD Card adapters, readers & performance

Re: PC Card, Cardbus, SD Card adapters, readers & performance

Re: PC Card, Cardbus, SD Card adapters, readers & performance

16-bit PC Card performance

Re: PC Card, Cardbus, SD Card adapters, readers & performance

Re: PC Card, Cardbus, SD Card adapters, readers & performance

Re: PC Card, Cardbus, SD Card adapters, readers & performanc