A7N8X: USB Card Reader not working

[Loom]

Hi,

my internal 6-in-1 card reader is giving me headaches:

after installation is install the 4 removable card deivces correctly,
eveything is running smooth. But after the next boot the card reader isn't
recognized anymore. Windows XP is halting for 10 seconds during boot, then
the power LED of the cardreader goes off and that's it. Every once in a
while it works, let's say every 10 bootup cycles.

Strange enough, but if you unplug and the replug the usb cable while running
windows, eveything works again. Up to the next boot, then it breaks again.
The 2 USB and the firewire port of the reader are still working all the
time.

Then I hooked the device to a powered USB Hub where it get recognized
everytime. Just not with the motherboard USB plugs. No matter which ones.

It's definatly no software problem, since it also doesn't work on a freshly
installed Windows XP with the recent nvidia drivers.

Any help is greatly appreciated

 

[Tony Cue]

Then I hooked the device to a powered USB Hub where it get recognized

everytime. Just not with the motherboard USB plugs. No matter which ones.

Surely this would indicate that there is not enough power to the device?
What size/brand PSU are you using? If it works fine with a "powered" USB
hub, there's your problem!

My IntellMouse occasionally powers off, I just unplug it, put it back in and
all is fine.
--
Regards,
Tony. (Remove Q to reply via e-mail)

"Tell me, Mr. Anderson, what good is a phone call if you're unable to
speak?"
Agent Smith, The Matrix...

 

[Loom]

Surely this would indicate that there is not enough power to the device?

What size/brand PSU are you using? If it works fine with a "powered" USB
hub, there's your problem!

My IntellMouse occasionally powers off, I just unplug it, put it back in and
all is fine.

Thanks, but I don't think it's the PSU, it's an 350W Enermax.

And don't forget, it's working if you unplug it and plug it in again.
The device itself shouldn't consume more power then let's say a mouse.

From all I read now, the Nforce2 chipset is not only very picky about
memory, but also
on USB.

Michael

 

[bongdaddyboo]

Hi,

my internal 6-in-1 card reader is giving me headaches:

Ditto, but both a combined CF/SM reader, a single SM reader and a Creative
Muvo.

after installation is install the 4 removable card deivces correctly,
eveything is running smooth. But after the next boot the card reader
isn't recognized anymore. Windows XP is halting for 10 seconds during
boot, then the power LED of the cardreader goes off and that's it.
Every once in a while it works, let's say every 10 bootup cycles.

Same again here.

Strange enough, but if you unplug and the replug the usb cable while
running windows, eveything works again. Up to the next boot, then it
breaks again. The 2 USB and the firewire port of the reader are still
working all the time.

And again....

Then I hooked the device to a powered USB Hub where it get recognized
everytime. Just not with the motherboard USB plugs. No matter which
ones.

Bingo. That's how you solve it (more below)

It's definatly no software problem, since it also doesn't work on a
freshly installed Windows XP with the recent nvidia drivers.

Any help is greatly appreciated

Here goes:

It's apparently a know problem that Nvidia are aware of and their published
solution is to use a USB 2.0 compliant hub. A USB 1.1 hub won't do the
trick.

Here's why:

For some reason, certain hardware - USB card readers, USB thumb drives, USB
MP3 players, USB hard discs (all USB 1.1 btw, not newer USB 2.0 versions) -
doesn't seem to fully adhere to the specification for timing signals as per
the USB spec, or at least Nvidia's interpretation of the USB spec that
they've implemented in the Nforce2 chipsets, and as such the timing signals
for these devices exhibit some "jitter", thus causing some of the behaviour
you've described above.

Without going into too much more detail, the way to get around this problem
is to connect a USB 2.0 hub inbetween the motherboard and the USB device
causing problems. USB 2.0 hubs apparently have some additional curcuitry
that reclocks the timing signal, and by feeding the 1.1 device through the
2.0 hub, the supposed "jitter" is removed and the motherboard receives a
clean and jitter-free timing signal from the 1.1 device. Result = the USB
1.1 kit works without a single problem now.

I'd recommend you try that solution as it's definitely worked for me. I
know there's a little extra expense involved initially, but it's definitely
worth it just to have the attached kit working all the time.

HTH

 

[Loom]

my internal 6-in-1 card reader is giving me headaches:

Ditto, but both a combined CF/SM reader, a single SM reader and a Creative
Muvo.

It's apparently a know problem that Nvidia are aware of and their published
solution is to use a USB 2.0 compliant hub. A USB 1.1 hub won't do the
trick.

Thanks!

I read that NVidia explanation before. From my point of view it's just
ridiculous that it's
not working. I'm beginning to regret that I bought the Nforce Board. A lot
of strange things
going on beside that. Not to speak from the infamous BIOS death symptom.
Don't want to hold
breath everytime I save CMOS settings. Why didn't they include a second BIOS
chip in the retail
package? Well, let's wait for the KT600.

But back to topic: The external hub where the cardreader is working is just
a plain USB 1.1 hub,
and not a 2.0 as Nvidia explained.

So what's going on here? I definately won't buy another hardware because
it's not that comfortable to
connect the cardreaders backpanel cord to an external hub because it's only
15 centimetres short.

Damn! :-(

One could think that USB is a brand new cutting edge technology. Shouldn't
happen.

Michael

Michael

 

[bongdaddyboo]

Thanks!

I read that NVidia explanation before. From my point of view it's just
ridiculous that it's
not working. I'm beginning to regret that I bought the Nforce Board.

Other than this particular problem my A7N8X-Deluxe has been a peach to be
honest.

But back to topic: The external hub where the cardreader is working
is just a plain USB 1.1 hub,
and not a 2.0 as Nvidia explained.

It's quite odd. I've actually tried it with a 1.1 hub and it doesn't work.
Whack the 2.0 hub in and Bingo, it works!

The 2.0 hub's definitely running 2.0 mode as well, as I've got an Epson
scanner through it as well as a 2.0 external drive enclosure.

It's got me stumped really, but the FUD via smoke and mirrors that Nvidia
feeds us has me convinced they're right!! Either that or they've arsed the
design of the chipset slightly and are trying to pass the blame to the
manufacturers of the "faulty" hardware.... because after all, Nvidia's

One could think that USB is a brand new cutting edge technology.
Shouldn't happen.

I know. Doesn't happen on an old MSI KT3-UltraARU board I've got either....
make your own conclusions.

 

[bongdaddyboo]

Paul wrote:

Interesting. I searched in Google, and found a reference to this
document:

http://www.epox.de/_boarddetail/8rda+/nForce2_MCP-T_MCP_USB1.1_AppNote.pdf

<SNIP: quoted contents of above link>

That's the info I came across recently.

<SNIP> For example, if you build your chip after other
companies have released their products, some companies would take
note of the non-compliant devices or "tough" devices, and would
add test cases to their new design, to make it more tolerant of
real world situations.

You'd think they would do that.

Maybe the difference between the MSI KT3-Ultra and a A7N8X, is
the clocking ? Does the KT3-Ultra have a clock chip, and does
the Ultra BIOS have any "Spread Spectrum" settings. I remember
early in the introduction of the A7N8X, people couldn't get
a BIOS setting of "disabled" for Spread Spectrum, to stick after
it was set.

Ah. Now from what I can remember, I've disabled spread spectrum on the KT3.
I'd have to make a journey to confirm though as my sister's got the board
now. Can't remember offhand how spread spectrum's set on this Asus board
but it's probably something worth checking I think, although I definietly
remember it had the 1003 BIOS when installed and SS wouldn't "stick" as
you've said above. Now it's got the latest 1004 BIOS for the 1.06 board
I'll drop in, check it out and report back.