NotebookForums.com › Forums › General Notebook Discussions › Linux & Other OS's › SIIG 1394b expresscard
New Posts  All Forums:Forum Nav:

SIIG 1394b expresscard

post #1 of 22
Thread Starter 
This one:
http://www.newegg.com/Product/Produc...82E16839150010

Is anyone using this card with a recent Linux kernel? I can't find any chipset info on this card. How about ExpressCard in general - easy to use? Kernel just sees it as a PCIe-hotplug slot, correct?

Just want to make sure it'll work before I buy. My laptop model is in my sig fwiw.

TIA!
post #2 of 22
moved to Linux
post #3 of 22
Quote:
Originally Posted by lupin..the..3rd View Post
This one: http://www.newegg.com/Product/Produc...82E16839150010 Is anyone using this card with a recent Linux kernel? I can't find any chipset info on this card. How about ExpressCard in general - easy to use? Kernel just sees it as a PCIe-hotplug slot, correct? Just want to make sure it'll work before I buy. My laptop model is in my sig fwiw. TIA!
I don't have a laptop with an ExpressCard, but it's supposed to be just a PCIe x1 lane and I'll bet that the regular "PCI Express Support" in the kernel will make the ExpressCard work. BTW, IEEE 1394 is standard like USB. The ohci1394 and sbp drivers should have the card work fine, too. I have 1394a stuff, no 1394b, but I think it's pretty similar. Just make sure your kernel supports IEEE 1394 and you'll be fine.
post #4 of 22
Thread Starter 
Cool, that's kind of what I figured.

Why the heck are most new notebook models still using 1394a anyways?? That's like shipping a notebook with USB 1.1 ports instead of USB 2.0.
post #5 of 22
Quote:
Originally Posted by lupin..the..3rd View Post
Cool, that's kind of what I figured. Why the heck are most new notebook models still using 1394a anyways?? That's like shipping a notebook with USB 1.1 ports instead of USB 2.0.
I'm actually using a notebook with 1394a and USB 1.1. Trust me, 1394a is MUCH better than USB 1.1. USB 2.0 is roughly as fast as 1394a and since people seem to be happy with USB 2.0, there's no push to put a faster interface like 1394b on machines.
post #6 of 22
Actually there is, in the professional markets which I believe is what Lupin is looking for.

Unfortunately very few people ship with 1394b, and even less ship a notebook with a powered firewire port, so it kinda sucks either way.

1394b is great for a small RAID array external so that you can take it with you where ever.

Seablade
post #7 of 22
Thread Starter 
Quote:
Originally Posted by imnotdrphil View Post
I'm actually using a notebook with 1394a and USB 1.1. Trust me, 1394a is MUCH better than USB 1.1. USB 2.0 is roughly as fast as 1394a and since people seem to be happy with USB 2.0, there's no push to put a faster interface like 1394b on machines.
USB 2.0 and 1394a can deliver roughly 40 MB/s of real world throughput. A single modern IDE disk drive can do about 80 MB/s in sequential throughput (either read or write). So the short answer is that USB 2.0 and 1394a reduce your external hard drive performance by HALF. Not good. 1394b delivers 80 MB/s real world performance - a perfect match for a single external drive or a RAID1 mirror (I'm using a Venus DS3R). Why would you want to only get half of the disk performance you've paid for? Now that laptops come with high-end gaming video cards, maybe the mobile gamer crowd will join the choir and start asking for this overdue feature.
post #8 of 22
Thread Starter 
Quote:
Originally Posted by seablade View Post
Actually there is, in the professional markets which I believe is what Lupin is looking for.
You got it, most of the better pro equipment favors the firewire interface. Besides, firewire protocol is similar to SCSI and is much better suited to bulk data transfer than USB.
post #9 of 22
Quote:
Originally Posted by lupin..the..3rd View Post
USB 2.0 and 1394a can deliver roughly 40 MB/s of real world throughput.
Not quite. Most USB 2.0 controllers can only throughput roughly 23-28 MB/sec, if my experiences and the experiences of guys like Tom's Hardware Guide and Anandtech are any indication. Not that many people have benchmarked IEEE 1394 stuff, so my only experience with it is in a USB 2.0/IEEE 1394a 2.5" drive enclosure. The USB 2.0 speeds were around 24 MB/sec from an NVIDIA CK804 USB 2.0 controller and IEEE 1394 is 26-28 MB/sec or so from a TI TSB43AB22/A onboard IEEE 1394a controller. The drive itself is a 100 GB Hitachi Travelstar 5K100 PATA unit that was in my old, now-dead notebook and had maximum sustained read speeds of about 40 MB/sec.
Quote:
A single modern IDE disk drive can do about 80 MB/s in sequential throughput (either read or write).
Some of the very largest-capacity 3.5" 7200 rpm PMR PATA drives can hit 80 MB/sec sustained in read speed, but they're more like 60-65 MB/sec in writes. Many people use 2.5" drives as external drives for notebooks instead of 3.5" units as the 2.5" ones generally can be USB-powered, unlike 3.5" units. The very best 7200 rpm 2.5" HDDs can reach somewhere around 60 MB/sec in peak read speeds and in the 50s for write speeds. But most units inside of the drive enclosures are 5400 rpm or even 4200 rpm units, with the 5400 rpm units getting between 40-50 MB/sec in sustained reads and the 4200 rpm units being in the 30s. There are also some 1.8" portable drives out there; those are all 4200 rpm units and generally are also in the 30s MB/sec read speeds.
Quote:
So the short answer is that USB 2.0 and 1394a reduce your external hard drive performance by HALF. Not good. 1394b delivers 80 MB/s real world performance - a perfect match for a single external drive or a RAID1 mirror (I'm using a Venus DS3R).
I'd like to see a benchmark of the 1394b drive because I've not seen anything on 1394b. The ports are pretty rare to find outside of Macintoshes and a few really high-end enthusiast motherboards. Another even better option for external HDD interfaces is external SATA. eSATA is much easier and less expensive to implement for most laptops as there is already SATA support in the chipset. Not true for 1394. eSATA has an interface speed that's 1.5 Gbps, yielding 150 MB/sec in transfer ability. An eSATA drive functions just like an internal SATA drive as far as the OS is concerned- hotplugging and all. But eSATA only works for disks, other things that are not mass-storage devices won't work with eSATA and would be better off with USB or 1394. Gigabit Ethernet is also an option as the theoretical bandwidth is 125 MB/sec point-to-point. It's not commonly seen on external disks as actually throughputting anywhere near 1 Gbps takes a pretty decent NIC and embedded chip in the disk as well as GbE from the disk to the computer. Unless somebody either has a gigabit switch at home or uses a crossover cable to tie computer directly to disk, it'll be 100 Mbps and much slower than USB 2.0, or 1394a/b.
Quote:
Why would you want to only get half of the disk performance you've paid for? Now that laptops come with high-end gaming video cards, maybe the mobile gamer crowd will join the choir and start asking for this overdue feature.
The technology is somewhat available on the external disks, but yeah, it's mostly USB 2.0-only. If your notebook lacks the 1394b ports, I believe that there are ExpressCards and PCMCIA cards that have the interfaces on them. Ditto with eSATA, although eSATA is faster than the entire PCI bus that PCMCIA is based on- better to go with an ExpressCard notebook if you want to run eSATA.
post #10 of 22
Thread Starter 
Quote:
Originally Posted by imnotdrphil View Post
Not quite. Most USB 2.0 controllers can only throughput roughly 23-28 MB/sec, if my experiences and the experiences of guys like Tom's Hardware Guide and Anandtech are any indication. Not that many people have benchmarked IEEE 1394 stuff, so my only experience with it is in a USB 2.0/IEEE 1394a 2.5" drive enclosure.
True, the USB 2.0 controllers, while they have a higher bitrate than 1394a (480 Mbps vs. 400 Mbps), they are slower at bulk data transfer (as I mentioned in my previous post). Not sure how you determined that "not that many people have benchmarked 1394". 1394 is the industry standard for external pro A/V equipment, and external storage on Mac. Since it's the defacto standard, there are a plethora of benchmarks. You'll find them on web sites geared towards professional audio/video editing though, not on consumer oriented sites like "Tom's".
Quote:
Originally Posted by imnotdrphil View Post
Some of the very largest-capacity 3.5" 7200 rpm PMR PATA drives can hit 80 MB/sec sustained in read speed, but they're more like 60-65 MB/sec in writes.
Not according to your favorite web site, here: http://www23.tomshardware.com/storag...2=676&chart=36 Set the drop-down to "Maximum write performance" and you'll see five different models that achieve greater than 80 MB/s (actually, very close to 90 MB/s). Now change the drop-down to "Maximum read performance" and the numbers are almost identical to the "write" numbers, with several entries greater than 80 MB/s. And HDD's are only getting faster each year.
Quote:
Originally Posted by imnotdrphil View Post
I'd like to see a benchmark of the 1394b drive because I've not seen anything on 1394b. The ports are pretty rare to find outside of Macintoshes and a few really high-end enthusiast motherboards.
Google is your friend, there are plenty. You're right, they are rare outside of Macs and enthusiast desktops. (That was pretty much what I complained about two posts ago). But there are plenty of add in cards that provide it; PCI, PCIe, ExpressCard.
Quote:
Originally Posted by imnotdrphil View Post
Another even better option for external HDD interfaces is external SATA.
Faster than 1394b, yes, but even more rare. Good luck finding ANY laptop OR desktop that has integrated eSATA.
Quote:
Originally Posted by imnotdrphil View Post
But eSATA only works for disks, other things that are not mass-storage devices won't work with eSATA and would be better off with USB or 1394.
Which, IMO, is why its just easier to standardize on a single interface. 1394 gives 800Mb/s today and 1600Mb/s with the next generation (once it becomes available). Plus, most pro A/V equipment today has already standardized on 1394.
Quote:
Originally Posted by imnotdrphil View Post
Gigabit Ethernet is also an option as the theoretical bandwidth is 125 MB/sec point-to-point.
With Gig-E, you're lucky if you even see HALF of the theoretical bandwidth. Ethernet is even worse for bulk data transfer than USB is. There is Waaaaaay too much latency in ethernet, plus the host CPU has to process all those TCP/IP frames, chewing up valuable CPU cycles. The latency alone kills ethernet as a viable bulk storage transport for all but the most mundane applications. Add to that the size of the TCP/IP packet overhead, and you've lost a good chunk of your bandwidth just to the protocol! The 1500 byte MTU does you no favors either. Ethernet and TCP/IP is probably the worst interface there is for bulk data.
Quote:
Originally Posted by imnotdrphil View Post
If your notebook lacks the 1394b ports, I believe that there are ExpressCards and PCMCIA cards that have the interfaces on them.
I believe so too. (See the title of this thread).
post #11 of 22
For the record it is very common to see 7200RPM RAID Arrays in use in external drives on 1394b for multimedia production.

I am considering buying a bus powered RAID-0 array with 2 2.5" laptop drives in it in fact for my own use.

Seablade
post #12 of 22
Quote:
Originally Posted by seablade View Post
For the record it is very common to see 7200RPM RAID Arrays in use in external drives on 1394b for multimedia production. I am considering buying a bus powered RAID-0 array with 2 2.5" laptop drives in it in fact for my own use. Seablade
You will need to probably have four USB cables attached to the thing as each USB cable can provide 2.5 watts (5.0 VDC, 500 mA) and most 2.5" drives pull roughly 5 watts apiece. Some pull a little more, especially 7200 rpm units, and some pull less, especially 4200 rpm units.
post #13 of 22
That is what Firewire is for One cable, no power supply.

PS It is a VERY bad idea to try to pull power from multiple USB ports on a laptop as there is nothing saying that those usb ports actually have seperate power supplies.

Seablade
post #14 of 22
Thread Starter 
Quote:
Originally Posted by seablade View Post
That is what Firewire is for One cable, no power supply. PS It is a VERY bad idea to try to pull power from multiple USB ports on a laptop as there is nothing saying that those usb ports actually have seperate power supplies. Seablade
Quite true. USB can supply at most just 2.5W of bus power per port, while 1394 provides 7.5W, 3x as much.
post #15 of 22
Quote:
Originally Posted by seablade View Post
That is what Firewire is for One cable, no power supply. PS It is a VERY bad idea to try to pull power from multiple USB ports on a laptop as there is nothing saying that those usb ports actually have seperate power supplies. Seablade
USB ports don't have separate power supplies from each other as they just get their +5 VDC from a trace in the motherboard from the +5 VDC leads from the PSU (desktops, servers) or from a +12 VDC to a +5 VDC regulator (laptops.) It doesn't matter if it's one trace or many traces because the power supply is all the same. Each port must be able to supply up to 500 mA of current to be compliant with the official USB spec. So I'd wager a guess that the ports on your computer really can output 500 mA each unless you have installed a really shady USB PCI or PCMCIA card. The reason why it's bad to pull power from a USB port is that the power may not be there. The USB spec says that all devices start at 100 mA current and then have to tell the USB host controller that they need 500 mA. Almost all bus-powered hard drives have one standard USB cable (data + power) and then one power-only cable. The power-only cables don't interface with the host controller's logic and thus should only get 100 mA. Most powered external USB hubs simply provide 500 mA of power to the +5 volts no matter what, but not all onboard USB host controllers do. I had trouble with a USB 2.0 2.5" HDD enclosure as my computer's host controller would only supply 600 mA of power to the HDD instead of the 1000 mA that it needed. It would commonly power-cycle when seeking and I'd see warning messages in my dmesg about USB power usage. I returned it and got a more expensive USB 2.0/IEEE 1394a enclosure. The new enclosure still uses the two-cable setup, but since one is 1394, there's 7.5 watts of power available just from the 1394 connector. The USB power connector is just gravy at that point and the disk runs just fine.
post #16 of 22
Thread Starter 
Quote:
Originally Posted by imnotdrphil View Post
The new enclosure still uses the two-cable setup, but since one is 1394, there's 7.5 watts of power available just from the 1394 connector. The USB power connector is just gravy at that point and the disk runs just fine.
So you're using a drive that's connected both to a USB port, and to a 1394 port on the same pc?? Why not connect it just to the 1394 port and be done with it? Am I missing something here?
post #17 of 22
Quote:
Originally Posted by lupin..the..3rd View Post
Quite true. USB can supply at most just 2.5W of bus power per port, while 1394 provides 7.5W, 3x as much.

Firewire spec has 15W, just some implementations of it(Apple on their laptops) only provide the 7W.

Seablade
post #18 of 22
Quote:
Originally Posted by imnotdrphil View Post
USB ports don't have separate power supplies from each other as they just get their +5 VDC from a trace in the motherboard from the +5 VDC leads from the PSU (desktops, servers) or from a +12 VDC to a +5 VDC regulator (laptops.) It doesn't matter if it's one trace or many traces because the power supply is all the same. Each port must be able to supply up to 500 mA of current to be compliant with the official USB spec. So I'd wager a guess that the ports on your computer really can output 500 mA each unless you have installed a really shady USB PCI or PCMCIA card.

It does matter actually. For instance if you are connected to a USB port via PCI(Leaving realm of just notebooks) you have a set amount of power that spec handles, whereas the amount of power that your motherboard can handle is something completely seperate.

While the end power supply is the same, so is the end power supply of two outlets in your house on seperate breakers.

Quote:
The reason why it's bad to pull power from a USB port is that the power may not be there.

Overgeneralizing there. It is a bad idea to pull more power than the spec allows for yes, but pulling power in general can be very good and is quite often done by many things(Keyboard, mice, etc).

Quote:
The new enclosure still uses the two-cable setup, but since one is 1394, there's 7.5 watts of power available just from the 1394 connector. The USB power connector is just gravy at that point and the disk runs just fine.

Like Lupin I don't understand the logic of an enclosure that requires both 1394 and USB. If you need that much power, you may as well just make it 1394.

Seablade
post #19 of 22
Quote:
Originally Posted by lupin..the..3rd View Post
So you're using a drive that's connected both to a USB port, and to a 1394 port on the same pc?? Why not connect it just to the 1394 port and be done with it? Am I missing something here?
The USB cable only carries power and doesn't even have any data wires inside of it. The drive has four connectors in the back: two 6-pin 1394s, one mini-USB (data), and a round +5 VDC input. A USB power-only cable with a ~3.5 mm round male power connector plugs into the +5 VDC input. So the FireWire carries the data and the USB power cable carries some extra juice in case the drive needs it. I could probably use the drive with just the 1394a cable on my desktop, but why not connect the extra power? My laptop has a 4-pin 1394a connector and that does require the extra power from the USB power connector to function. Note that I do not have the USB data cable attached- it sits in a drawer and does not get used.
post #20 of 22
What type of drive is this?

Personally it sounds like a horrible design. If it can't comfortably be run off Firewire, the limited power from USB will not help much, and if you are going to be that close to overdrawing, it shouldn't exist IMO.

Seablade
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Linux & Other OS's
NotebookForums.com › Forums › General Notebook Discussions › Linux & Other OS's › SIIG 1394b expresscard