quick question about USB cards...

[Eric]

My friend has a relatively new Win7pro system (I don't have the specs at the
moment) and we added a USB 2.0 card (Rosewill NEC 4+1 Port USB2.0 PCI CARD
Model RC-101 ). Now if he has anything plugged into that card on boot the
system doesn't make it past the post screen before it locks up, but if he
leaves the card free when he boots the system starts normally and everything
works just fine. My guess is that the pull from the card and additional
devices attached is just a little too much for the power supply (I'n not
sure what he has for a power supply and won't know until tonight), but I
thought I'd check here to see if there is some sort of common problem that
I'm unaware of...

 

[Paul]

My friend has a relatively new Win7pro system (I don't have the specs at the
moment) and we added a USB 2.0 card (Rosewill NEC 4+1 Port USB2.0 PCI CARD
Model RC-101 ). Now if he has anything plugged into that card on boot the
system doesn't make it past the post screen before it locks up, but if he
leaves the card free when he boots the system starts normally and everything
works just fine. My guess is that the pull from the card and additional
devices attached is just a little too much for the power supply (I'n not
sure what he has for a power supply and won't know until tonight), but I
thought I'd check here to see if there is some sort of common problem that
I'm unaware of...

The 4+1 card has only 4 physical ports on the NEC chip. One of the external ports
is shared electrically with the internal port. Don't use both
of those at the same time.

The information about sharing, isn't in the manual.

http://www.rosewill.com/Mgnt/Uploads2/AttachmentForProduct/RC-101_user_manual.pdf

Looking at the layout, the bottom connector on the faceplate, has a different
component configuration than the other three connectors. So I'd have to
*guess* that the bottom connector is the shared one. If you use the internal
connector, then avoid using the bottom one at the same time, and vice versa.

There is some feedback here.

http://www.newegg.com/Product/Product.aspx?Item=N82E16815166002

"2/12/2011 5:58:52 PM Wanted it to work

Non-functional. Have tried on 3 different motherboards with the
same result each time...failure to boot. Regardless of which pci
port it is inserted in, the system failed to boot every time.
Removal of the card resulted in a normal boot cycle. First
Rosewill product to let me down. As others have stated, not
worth returning for RMA."

"10/1/2008 5:27:25 PM Won't handle large video files

Data corruption when copying large video files (eg, > 10M in size).
Notice that user "ssorg" reported same problem in his/her review from
5/7/2008. My files were recorded from a hidef camcorder to an SDHC card.
When files copied to my laptop via a USB connection and then compared
to original, no problems. When files copied to my desktop PC via new
Rosewill RC-101 USB card, copied files had data corruption. Files copied
did NOT compare successfully with original files"

At least one problem report, was consistent with the card nob being
fully seated in the slot. If the computer is a home built, verify
the alignment of the motherboard to the add-in slots. I've had a
sound card fail to work right, because one of the 32 bus bits
wasn't touching, and the Plug and Play identifier was wrong in
one bit as a result. (That would be just the start of problems, as
any function on the card would also end up broken because of that.)

So there is some kind of issue, but exactly what, only a schematic
for the thing would admit to a mistake. The NEC chip itself, has
a history of working well (as the very first USB2 driver was tested
against one of those chips). It's been around a while.

The only known issue with things like NEC chips, is the ports
"blow out" easily, via static electricity. So that's the only
warning against owning them. Lots of NEC owners, may have a port
or two that no longer works, and the failure is believed to be
due to static sensitivity.

Try another card, and see what happens ? Perhaps look for a
competing brand of USB chip (like a card based on VIA
Technologies chip), and see if it works any better.

Paul

 

[Flasherly]

My friend has a relatively new Win7pro system (I don't have the specs at the
moment) and we added a USB 2.0 card (Rosewill NEC 4+1 Port USB2.0 PCI CARD
Model RC-101 ). Now if he has anything plugged into that card on boot the
system doesn't make it past the post screen before it locks up, but if he
leaves the card free when he boots the system starts normally and everything
works just fine. My guess is that the pull from the card and additional
devices attached is just a little too much for the power supply (I'n not
sure what he has for a power supply and won't know until tonight), but I
thought I'd check here to see if there is some sort of common problem that
I'm unaware of...

Could be a conflict in the BIOS hardwired input/output IRQ or address
schemes, although not something I'd expect offhand from Rosewill's
(Neweggs bread&butter brand) design compatibility & engineering
standpoint. A PS swap can't totally be ruled out, but I'd first check
BIOS boot option devices and order, possibly then IRQs, although the
latter may make a mess of the Windows config. Boot priority &
assignment in the BIOS on/off switches to see when/if it'll boot past
with an attached device recognized by the OS. There's only so much
you can do to a MB past its BIOS and a few pinjumpers as far as
testing by trial and error, after all, unless you're lucky and have
magnificent spare power supplies unused and ready, new in their boxes.

 

[Loren Pechtel]

My friend has a relatively new Win7pro system (I don't have the specs at the
moment) and we added a USB 2.0 card (Rosewill NEC 4+1 Port USB2.0 PCI CARD
Model RC-101 ). Now if he has anything plugged into that card on boot the
system doesn't make it past the post screen before it locks up, but if he
leaves the card free when he boots the system starts normally and everything
works just fine. My guess is that the pull from the card and additional
devices attached is just a little too much for the power supply (I'n not
sure what he has for a power supply and won't know until tonight), but I
thought I'd check here to see if there is some sort of common problem that
I'm unaware of...

I've seen this but only with motherboard ports. The scenario I have
worked out is that something plugged into the card leaks voltage
backwards. I don't know what happens but it sticks things in the POST
sequence. Note that it's *NOT* a lockup--removing the offending
device will allow it to continue on.

The reason I think it's leaking power is that on the machine I first
encountered it the motherboard had one of those I'm-powered lights.
That light would persist even when it was unplugged from the main
(yes, I know to push the power button to discharge the caps), but it
would quickly fade if the offending USB cable were unplugged.

Beware that my experience was that the USB ports in question would be
destroyed in a few months.

 

[larry moe 'n curly]

The only known issue with things like NEC chips, is the ports
"blow out" easily, via static electricity. So that's the only
warning against owning them. Lots of NEC owners, may have a port
or two that no longer works, and the failure is believed to be
due to static sensitivity.

I don't know about current NEC USB 2.0 chips, but older boards with
the original 160-pin NEC chips often failed because the 8-pin LM3526-H
chip meant to protect the NEC chip would blow, and I fixed a couple by
replacing that small chip. According to the National Semiconductor
data sheet for that chip, p. 8:

http://www.national.com/ds/LM/LM3526.pdf

there's supposed to be a minimum 120uF tantalum capacitor across the
+5V, and a 0.01uF ceramic in parallel is recommended, but I've never
seen a USB board like that, NEC or not, and only one board even had
ceramic capacitors (Maxtor, based on NEC chip, also my only USB card
with a fuse for each port).

 

[Sjouke Burry]

My friend has a relatively new Win7pro system (I don't have the specs at the
moment) and we added a USB 2.0 card (Rosewill NEC 4+1 Port USB2.0 PCI CARD
Model RC-101 ). Now if he has anything plugged into that card on boot the
system doesn't make it past the post screen before it locks up, but if he
leaves the card free when he boots the system starts normally and everything
works just fine. My guess is that the pull from the card and additional
devices attached is just a little too much for the power supply (I'n not
sure what he has for a power supply and won't know until tonight), but I
thought I'd check here to see if there is some sort of common problem that
I'm unaware of...

Maybe bios legacy support is off, or not an option on your computer?

 

[larry moe 'n curly]

My friend has a relatively new Win7pro system (I don't have the specs at the
moment) and we added a USB 2.0 card (Rosewill NEC 4+1 Port USB2.0 PCI CARD
Model RC-101 ). Now if he has anything plugged into that card on boot the
system doesn't make it past the post screen before it locks up, but if he
leaves the card free when he boots the system starts normally and everything
works just fine. My guess is that the pull from the card and additional
devices attached is just a little too much for the power supply (I'n not
sure what he has for a power supply and won't know until tonight), but I
thought I'd check here to see if there is some sort of common problem that
I'm unaware of...

Is your friend's system set up in the BIOS to boot first from a USB
device or plug-in disk controller? That's the only thing I can think
of that would cause boot problems. What if your friend makes a
bootable USB flash drive, such as with this utility from HP?

http://www.pcworld.com/downloads/file/fid,64963-page,1/description.html

If it makes the computer boot to a DOS prompt, I'd say it's proof that
changing the boot order or another boot option in BIOS is the
solution.

 

[Paul]

I don't know about current NEC USB 2.0 chips, but older boards with
the original 160-pin NEC chips often failed because the 8-pin LM3526-H
chip meant to protect the NEC chip would blow, and I fixed a couple by
replacing that small chip. According to the National Semiconductor
data sheet for that chip, p. 8:

http://www.national.com/ds/LM/LM3526.pdf

there's supposed to be a minimum 120uF tantalum capacitor across the
+5V, and a 0.01uF ceramic in parallel is recommended, but I've never
seen a USB board like that, NEC or not, and only one board even had
ceramic capacitors (Maxtor, based on NEC chip, also my only USB card
with a fuse for each port).

The purpose of 100uF caps, in USB port designs, is to handle inrush
surge (charging the capacitor inside the USB peripheral itself). An
Intel report, showed an oscillogram of the current flow transient,
and there can be a 5 amp spike when a USB peripheral is plugged in.
At that time, Intel recommended the host port be protected with
100uF bypass, as that holds sufficient charge such that the host power
bus doesn't dip too much. Motherboards typically use electrolytics
for the purpose, which are fine for the job. If I was designing
it, I'd put in an electrolytic without a second thought, whereas
with tantalum, I'd need to do further research.

At least some flavors of tantalum, are sensitive to current spikes.
Adding a series limiting resistor, as this article suggests, would
defeat any advantage a tantalum might offer.

http://www.avx.com/docs/techinfo/surgtant.pdf

There is some upper limit, on the size of bypass cap inside a
USB peripheral, and the host side cap has to be correspondingly
larger. And from that, that is where the 100uF comes from - the
ability to plug a second USB device into a USB stack, without
the first device being upset and losing state.

With respect to the NEC USB2 port failures, this is purely from
anecdotal reports, of cards where one or more ports no longer
functioned, and the others worked fine. The USB cards they make
now for $10, generally only have the main chip, and a few electrolytics.

The National part is one of those "current policeman" type devices.
The datasheet says it trips at 1 amp, which means a sustained
(after the transient) current flow of 1 amp, cause the flag
to be asserted, and the external port to have +5V power cut.
That part doesn't touch D+ and D-, and if you wanted to add
static discharge protection, that would be an entirely
different research project.

The static protection solutions here, are rated in terms of
stray capacitance, and that is what degrades high frequency
signal performance.

http://www.semtech.com/esd-protection/USB-ESD-Protection/index.html

There is a nice eye diagram on page 14, showing what a USB port
with transient protection on it looks like. The PESD5V0X1BL
are $0.16 a piece, and you'd need two per port for a total
of $0.32 added per port (on a $10 retail USB card). You'd have
to bump up the price at retail to $15 to cover it. I think
most fly-by-night USB card makers, would rather let the
NEC ports blow, than do that.

http://www.nxp.com/documents/application_note/AN10753.pdf

Paul

 

[larry moe 'n curly]

larry moe 'n curly wrote:

The only known issue with things like NEC chips, is the ports
"blow out" easily, via static electricity. So that's the only
warning against owning them. Lots of NEC owners, may have a port
or two that no longer works, and the failure is believed to be
due to static sensitivity.


The purpose of 100uF caps, in USB port designs, is to handle inrush
surge (charging the capacitor inside the USB peripheral itself). An
Intel report, showed an oscillogram of the current flow transient,
and there can be a 5 amp spike when a USB peripheral is plugged in.
At that time, Intel recommended the host port be protected with
100uF bypass, as that holds sufficient charge such that the host power
bus doesn't dip too much. Motherboards typically use electrolytics
for the purpose, which are fine for the job. If I was designing
it, I'd put in an electrolytic without a second thought, whereas
with tantalum, I'd need to do further research.

At least some flavors of tantalum, are sensitive to current spikes.
Adding a series limiting resistor, as this article suggests, would
defeat any advantage a tantalum might offer.

http://www.avx.com/docs/techinfo/surgtant.pdf

There is some upper limit, on the size of bypass cap inside a
USB peripheral, and the host side cap has to be correspondingly
larger. And from that, that is where the 100uF comes from - the
ability to plug a second USB device into a USB stack, without
the first device being upset and losing state.

With respect to the NEC USB2 port failures, this is purely from
anecdotal reports, of cards where one or more ports no longer
functioned, and the others worked fine. The USB cards they make
now for $10, generally only have the main chip, and a few electrolytics.

The National part is one of those "current policeman" type devices.
The datasheet says it trips at 1 amp, which means a sustained
(after the transient) current flow of 1 amp, cause the flag
to be asserted, and the external port to have +5V power cut.
That part doesn't touch D+ and D-, and if you wanted to add
static discharge protection, that would be an entirely
different research project.

The static protection solutions here, are rated in terms of
stray capacitance, and that is what degrades high frequency
signal performance.

http://www.semtech.com/esd-protection/USB-ESD-Protection/index.html

There is a nice eye diagram on page 14, showing what a USB port
with transient protection on it looks like. The PESD5V0X1BL
are $0.16 a piece, and you'd need two per port for a total
of $0.32 added per port (on a $10 retail USB card). You'd have
to bump up the price at retail to $15 to cover it. I think
most fly-by-night USB card makers, would rather let the
NEC ports blow, than do that.

http://www.nxp.com/documents/application_note/AN10753.pdf

Thanks, Paul. Great information.

Some of those anecdotal failures of the National LM3526-H chips
happened with IOgear brand NEC-based cards of mine, and IOgear
repaired the cards by soldering in new National chips. I knew they
didn't simply replace the cards because I had marked them on the
edges, and the solder flux hadn't been removed.

 

[Paul]

Some of those anecdotal failures of the National LM3526-H chips
happened with IOgear brand NEC-based cards of mine, and IOgear
repaired the cards by soldering in new National chips. I knew they
didn't simply replace the cards because I had marked them on the
edges, and the solder flux hadn't been removed.

You'd need pretty cheap "hired help" to make that pay off

Once the card was repaired, would it fail again ? Was it
actually a design defect in the National chip ? You would
think a "current policeman" design would be pretty thoroughly tested
(as the function is so simple).

Paul

 

[Nobody > (Revisited)]

You'd need pretty cheap "hired help" to make that pay off

Depends..

This could be slack-time and/or batch work, and that does make sense,
especially when it's a known problem. It's even easier if it's a
consistent fail-mode; after you've troubleshot the first few you may not
even need to do more than a quick pass-fail check.

Since it's pretty obvious which *same* parts are bad, the shop doesn't
fix the boards as they come in. They let them stack up for a while, then
set up a rework station to replace said bad parts and do them in
batches. Same part locations, same preheat settings, etc means that all
you have to do is feed boards and parts.

Rework station setup is still about 80% of the time used in circuit
board repairs, once the troubleshooting is out of the way.

Once the card was repaired, would it fail again ? Was it
actually a design defect in the National chip ? You would
think a "current policeman" design would be pretty thoroughly tested
(as the function is so simple).

The best designs in the world are always trumped by bad manufacturing
and/or testing processes.

Think of a 1N4148 silicon diode. That's about as simple as anything
electronic can be, with b(tr?)illions made and used.

Add the MILSPEC tag JANTX, that's 100% testing (and a huge markup on cost).

I don't remember how many were in the lot, but over 100 "reels" of them
shut down one plant I worked at years ago.

Every one was backwards-banded. Still good, but 100% bad if already
stuffed in thousands of boards for almost everything in your product line.



--
"Shit this is it, all the pieces do fit.
We're like that crazy old man jumping
out of the alleyway with a baseball bat,
saying, "Remember me motherfucker?"
Jim “Dandy” Mangrum

 

[Flasherly]

Think of a 1N4148 silicon diode. That's about as simple as anything
electronic can be, with b(tr?)illions made and used.

Add the MILSPEC tag JANTX, that's 100% testing (and a huge markup on cost).

Appears Space and Military are within the same guidelines. TX
signifying a screen designate above [or mid-QC] for "military
application only," with "visual" being the entry below singular Space
(NASA/DOD) -J -S designates, at a subset semiconductors such and if
then respectively of the highest JAN rating for being "off this
world," as best we know it.
Interesting stuff, DOD's MIL/STD pub. specs. Class A and B tests also
appear encompassing JAN specs overall -- 1000-5000 series at the top 5-
qualifiers for 5004 and -5 assignments to Screening, which can run a
gamut from multiple latitudes across visual methodologies, similarly,
burn-ins, to include temperature, acceleration, radiographic,
radiation, and general particle impact -- to hazard the B designate
conation code being of less stringent procedures. Bond pull and fine/
coarse seals, I haven't a clue as what those two signify to a discrete
or die assembly component. . .microscopy at robotic assembly
confidence levels, as would the welder's work be scrutinized with x-
ray, say, at a nuclear power plant?

 

[larry moe 'n curly]

You'd need pretty cheap "hired help" to make that pay off

Once the card was repaired, would it fail again ? Was it
actually a design defect in the National chip ? You would
think a "current policeman" design would be pretty thoroughly tested
(as the function is so simple).

IOgear fixed my NEC cards because the only USB cards they still had in
stock were based on an ALi USB chip with no driver for Windows 98,
only 98SE, and IOgear had claimed Win98 compatibility for the
product. None of the repaired cards failed, although I didn't try to
zap any of them, and I did replace the bulk filter capacitors with
tantalums. I'm currently running one of those cards on an ECS Nforce6
motherboard that blew out 2 of its USB ports.