Heads up on ASUS motherboards

[Todd]

Hi All,

I am a system builder and I wanted to give my colleagues a heads
up on ASUS motherboards. Seems they work fine when they work, but
good luck with any customer service items. In other words you
are self insuring if you use their stuff.

For instance, I got a motherboard with a single bent CPU pin.
The pin was bent at a 45 degree angle. I missed it when I
when to install the CPU.

By the way, I do not bend CPU pins. I line up the CPU with
the tabs and gently set it down as if I was handling
nitroglycerin. And, when folks bend pins by improperly
inserting CPUs, they bend a mess of pins, not just one.
And how in the world did it get bent at 45 degrees?

Since ASUS refused to accept bent CPU pins, my distributor
would not take the board back. I called ASUS and they confirmed
that bent CPU pins are not covered. My salesman at my distributor
said that every salesman that works there has also noticed this
ASUS specific problem. He also said that getting any kind of
customer service out of ASUS is a nightmare. He pointed me
to Intel motherboards. I prefer Supermicro.

So, avoid ASUS. Cost me $200.00 to figure it out, the hard way.

I hope this saves someone else from making my mistake.
-T

 

[Paul]

Hi All,

I am a system builder and I wanted to give my colleagues a heads
up on ASUS motherboards. Seems they work fine when they work, but
good luck with any customer service items. In other words you
are self insuring if you use their stuff.

For instance, I got a motherboard with a single bent CPU pin.
The pin was bent at a 45 degree angle. I missed it when I
when to install the CPU.

By the way, I do not bend CPU pins. I line up the CPU with
the tabs and gently set it down as if I was handling
nitroglycerin. And, when folks bend pins by improperly
inserting CPUs, they bend a mess of pins, not just one.
And how in the world did it get bent at 45 degrees?

Since ASUS refused to accept bent CPU pins, my distributor
would not take the board back. I called ASUS and they confirmed
that bent CPU pins are not covered. My salesman at my distributor
said that every salesman that works there has also noticed this
ASUS specific problem. He also said that getting any kind of
customer service out of ASUS is a nightmare. He pointed me
to Intel motherboards. I prefer Supermicro.

So, avoid ASUS. Cost me $200.00 to figure it out, the hard way.

I hope this saves someone else from making my mistake.
-T

This can happen if:

1) Your seller "re-boxes" returned products, without inspecting them.
Motherboards are not sealed on purpose, to facilitate this practice
(so there is no way to know if the thing is "factory fresh"). The ESD
bags on the ones I've bought, were never "sealed". No sticker. In fact
on one motherboard I bought at a local store, they removed it from the
bag in the shop, to prove it had the revision number I was looking for.

2) Each motherboard is given a functional test at the factory. A processor
is inserted. While a VCC or GND pin could easily be bent, and not
affect system operation, I'd think that there would be a quick visual
inspection by the tester, when they place the PNP cap into the LGA socket.
The PNP cap is there to prevent damage. But the cap is not a "brick wall",
and is there partially to prevent snagging or dirt getting into
the socket area.

3) If the box is crushed in shipping (i.e. UPS), then shipping damage
should be taken up with the shipper. Occasionally, a PNP cap is
compromised by crushing damage.

The most likely explanation, is your board was returned by another
customer, not properly inspected, and offered for sale again.
If the bag on the motherboard had a seal that needed to be broken,
this would be easier to detect. If you had an issue of this nature,
"not new product", you'd take that up with the seller. Use a
"credit card dispute" if you cannot get satisfaction.

Depending on the socket, up to 2/3rds of all pins are for power.
It's because the processor draws up to 100 amps at low voltage, and
the bond wires or lead frame can only handle a fraction of an amp per
contact. Thus a lot of contacts are needed. With luck, you bend or
break a pin which is expendable, so it doesn't need to be fixed.
If instead of 200 pins to carry VCC, there were only 199 pins,
the processor is none the wiser. If you can find a pinout diagram for
the socket, you can figure this out for yourself. (I needed this
info, when I "insulated" a pin on purpose on my other LGA775 motherboard,
so I could overclock it by changing the BSEL logic signal level. My
other motherboard, doesn't have good code to program the clock generator,
so a "BSEL mod" is how you overclock it. The clock jumps by 33% when
I ground that signal )

If the pin (spring) in the LGA socket, is a bus signal, then snapping
that off is game over.

Paul

 

[BillW50]

It's because the processor draws up to 100 amps at low voltage, and
the bond wires or lead frame can only handle a fraction of an amp per
contact...

A hundred amps? Tell me more about these processors. I want to learn
more. Even if they operate at 2 volts, that is 200 watts per processor.

 

[Paul]

A hundred amps? Tell me more about these processors. I want to learn
more. Even if they operate at 2 volts, that is 200 watts per processor.

Well, head back to the Prescott era. Try page 17 here. (Click the download
link on the upper right.)

http://www.intel.com/content/www/us...down-11-0-processor-power-delivery-guide.html

At one time, the P4 processors were in the S478 socket. Intel looked
at the power projections, and decided the sky was the limit. And
thus, the LGA775 socket was born. The main difference between
775 and 478, was extra VCC and GND pins for power. There are
practically no new logic signals (maybe two or three). Granted,
you could say the land grid array contacts weren't rated for the
same current as a ZIF pin and contact could carry. But I like
to think they added all those pins, so they could continue to
make "barn-burner" processors.

In any case, that would probably be close to the peak of the
Intel power thing. I'm not going to go looking for datasheets
now, to find a V versus I curve for every family. I don't even
know what VRD or VRM spec they're up to now.

The purpose of explaining this, is to point out that a lot of
the pins can be power and ground. And because they're in parallel,
you can snap those off and still use the motherboard. (Just make
sure a power doesn't bend over and touch a ground.) I always
got a chuckle when someone would report they busted
pins off, and they were pleasantly surprised their processor
still worked. A certain ratio of power/ground to signal pins,
was always required to handle things like simultaneous switching
noise. But in the case of processors, the number of pins can
also be justified by the large core power component.

If you get a processor datasheet, look at all the single point
V and I specifications, the load line (dynamic V versus I),
and you multiply the numbers together, you'll hardly ever
get a match to the TDP. The datasheet is remarkably inconsistent
in that respect, and defies that kind of analysis. But this hasn't
stopped Intel from writing their datasheets that way. I think if
you accused them of something, they'd simply point you to the separate
document that covers thermal issues. They'd tell you you're not
supposed to be multiplying numbers off the datasheet together in
an attempt to arrive at a TDP. And if you insist on doing it that
way, you'll only end up concluding the core is running at around
1 volt or so.

Paul

 

[BillW50]

In Paul typed:

Well, head back to the Prescott era. Try page 17 here. (Click the
download link on the upper right.)

http://www.intel.com/content/www/us...down-11-0-processor-power-delivery-guide.html

At one time, the P4 processors were in the S478 socket. Intel looked
at the power projections, and decided the sky was the limit. And
thus, the LGA775 socket was born. The main difference between
775 and 478, was extra VCC and GND pins for power. There are
practically no new logic signals (maybe two or three). Granted,
you could say the land grid array contacts weren't rated for the
same current as a ZIF pin and contact could carry. But I like
to think they added all those pins, so they could continue to
make "barn-burner" processors.

In any case, that would probably be close to the peak of the
Intel power thing. I'm not going to go looking for datasheets
now, to find a V versus I curve for every family. I don't even
know what VRD or VRM spec they're up to now.

The purpose of explaining this, is to point out that a lot of
the pins can be power and ground. And because they're in parallel,
you can snap those off and still use the motherboard. (Just make
sure a power doesn't bend over and touch a ground.) I always
got a chuckle when someone would report they busted
pins off, and they were pleasantly surprised their processor
still worked. A certain ratio of power/ground to signal pins,
was always required to handle things like simultaneous switching
noise. But in the case of processors, the number of pins can
also be justified by the large core power component.

If you get a processor datasheet, look at all the single point
V and I specifications, the load line (dynamic V versus I),
and you multiply the numbers together, you'll hardly ever
get a match to the TDP. The datasheet is remarkably inconsistent
in that respect, and defies that kind of analysis. But this hasn't
stopped Intel from writing their datasheets that way. I think if
you accused them of something, they'd simply point you to the separate
document that covers thermal issues. They'd tell you you're not
supposed to be multiplying numbers off the datasheet together in
an attempt to arrive at a TDP. And if you insist on doing it that
way, you'll only end up concluding the core is running at around
1 volt or so.

Whenever I am looking for answers, it is you Paul who keeps coming up
with them. Remember the show MacGyver? I think you have him beat. ;-)

 

[Todd]

This can happen if:

1) Your seller "re-boxes" returned products, without inspecting them.
Motherboards are not sealed on purpose, to facilitate this practice
(so there is no way to know if the thing is "factory fresh"). The ESD
bags on the ones I've bought, were never "sealed". No sticker. In fact
on one motherboard I bought at a local store, they removed it from the
bag in the shop, to prove it had the revision number I was looking for.

No sign of them doing that.

2) Each motherboard is given a functional test at the factory. A processor
is inserted. While a VCC or GND pin could easily be bent, and not
affect system operation, I'd think that there would be a quick visual
inspection by the tester, when they place the PNP cap into the LGA socket.
The PNP cap is there to prevent damage. But the cap is not a "brick wall",
and is there partially to prevent snagging or dirt getting into
the socket area.

Never though to look. Never sold ASUS before and they are the ones
with the bent CPU pin problem. Even if I found one, ASUS would
not back me up.

3) If the box is crushed in shipping (i.e. UPS), then shipping damage
should be taken up with the shipper. Occasionally, a PNP cap is
compromised by crushing damage.

No sign of that.

The most likely explanation, is your board was returned by another
customer, not properly inspected, and offered for sale again.
If the bag on the motherboard had a seal that needed to be broken,
this would be easier to detect. If you had an issue of this nature,
"not new product", you'd take that up with the seller. Use a
"credit card dispute" if you cannot get satisfaction.

Depending on the socket, up to 2/3rds of all pins are for power.
It's because the processor draws up to 100 amps at low voltage, and
the bond wires or lead frame can only handle a fraction of an amp per
contact. Thus a lot of contacts are needed. With luck, you bend or
break a pin which is expendable, so it doesn't need to be fixed.
If instead of 200 pins to carry VCC, there were only 199 pins,
the processor is none the wiser. If you can find a pinout diagram for
the socket, you can figure this out for yourself. (I needed this
info, when I "insulated" a pin on purpose on my other LGA775 motherboard,
so I could overclock it by changing the BSEL logic signal level. My
other motherboard, doesn't have good code to program the clock generator,
so a "BSEL mod" is how you overclock it. The clock jumps by 33% when
I ground that signal )

If the pin (spring) in the LGA socket, is a bus signal, then snapping
that off is game over.

Paul

Socket 1155. The symptom was no video. If I removed the memory,
I got beep codes, which means the processor was working. So,
I thing the MB is back anyway, pin or no pin.

Moral of the story, if you sell ASUS motherboards, ASUS will
not be there to back you up. You self insure.

 

[Todd]

I thing the MB is back anyway, pin or no pin.

Should have been:

I think the MB is bad anyway, pin or no pin.

Stinkin' typos

 

[Paul]

Should have been:

I think the MB is bad anyway, pin or no pin.

Stinkin' typos

The initial warranty is provided by the seller. (Typically 30 days
with respect to proper order fulfillment.)

Except in cases where it explicitly states (on the seller's site)
that they will refer all issues to Asus.

I agree it's difficult to get through to Asus. The thing is,
there's more than one way to do it. Take a look at the Newegg.com
reviews for a motherboard, and when you see cases where Asus has
responded to a complaint right in the review section, there'll be
an email address to contact.

This is another path (different than the last time I looked).

http://service.asus.com/ProductCategory.aspx?pid=6

Bare minimum though, to get action, you'd need to fill out a
report. And we all know how well the web works, and what happens
to a user's carefully filled out report. Even when you have a
reference number, and contact them to expedite, you may still be
stonewalled by "we don't have a report by that number". Just try
all the interfaces to Asus that you can, to increase the odds
of success.

If it was me, and I was within the 30 days, I think I'd still try
the seller, even if the page said all warranty claims go through
Asus. If you told a convincing story (shipping damage), perhaps
they'd handle it.

And you also have "credit card dispute" to get satisfaction. One
time, it took me roughly a week of phoning people, until I got
satisfaction. A company not involved in the transaction "made
things right" and a package showed up at my door unannounced
(because they'd had so many reports of customers being screwed
like I was). I threatened the parties to the transaction with
non-payment, but didn't have to follow through when my
missing item showed up all on its own. So keep at it.

Paul

 

[Paul in Houston TX]

From: "Todd" <[email protected]>

Never had any problems with ASUS motherboards. I consider your
inconsistent and the CPU is not a part of the motherboard as it is not
soldered in, you have to install it.

Modern MB's have springy pins on the MB socket.
The cpu itself does not have pins.
The slightest touch sideways on the cpu socket pins
and the pins bend or snap off.
Don't ask how I know

 

[Todd]

The initial warranty is provided by the seller. (Typically 30 days
with respect to proper order fulfillment.)

Except in cases where it explicitly states (on the seller's site)
that they will refer all issues to Asus.

I agree it's difficult to get through to Asus. The thing is,
there's more than one way to do it. Take a look at the Newegg.com
reviews for a motherboard, and when you see cases where Asus has
responded to a complaint right in the review section, there'll be
an email address to contact.

This is another path (different than the last time I looked).

http://service.asus.com/ProductCategory.aspx?pid=6

Bare minimum though, to get action, you'd need to fill out a
report. And we all know how well the web works, and what happens
to a user's carefully filled out report. Even when you have a
reference number, and contact them to expedite, you may still be
stonewalled by "we don't have a report by that number". Just try
all the interfaces to Asus that you can, to increase the odds
of success.

If it was me, and I was within the 30 days, I think I'd still try
the seller, even if the page said all warranty claims go through
Asus. If you told a convincing story (shipping damage), perhaps
they'd handle it.

And you also have "credit card dispute" to get satisfaction. One
time, it took me roughly a week of phoning people, until I got
satisfaction. A company not involved in the transaction "made
things right" and a package showed up at my door unannounced
(because they'd had so many reports of customers being screwed
like I was). I threatened the parties to the transaction with
non-payment, but didn't have to follow through when my
missing item showed up all on its own. So keep at it.

Paul

As a reseller, I purchase through distributors. In other words,
I buy wholesale. So, none of the above will help me. And, believe
me, I let ASUS know that I will never sell their motherboards again.
They don't care. ASUS also told me that bent CPU pins are never
covered by warranty. And since ASUS is the only one with a bent CPU
socket in pin problem ...

I suppose I could do the credit card thing, but my reseller would drop
me and that would be a very stupid move on my part. Relationships
outlive transactions. Besides, my salesman is feeling guilty,
and I kind of like having him in the position. He is a great guy.
Love guilt discounts!

Here is an idea, ASUS should be honest and sell their motherboards
"as in". Maybe cut $20.00 from the price.

Moral of the story, if you sell ASUS motherboards, ASUS will
not be there to back you up. You self insure.

-T

I wonder how many typos I made this time. Hay! I went
to publik skool

 

[Todd]

Modern MB's have springy pins on the MB socket.
The cpu itself does not have pins.
The slightest touch sideways on the cpu socket pins
and the pins bend or snap off.
Don't ask how I know

From my original post:

By the way, I do not bend CPU pins. I line up the CPU with
the tabs and gently set it down as if I was handling
nitroglycerin. And, when folks bend pins by improperly
inserting CPUs, they bend a mess of pins, not just one.
And how in the world did it get bent at 45 degrees?

Seriously, I handle them like nitroglycerin. You can not
believe how precise and gentle I am.

I have been doing this for 17 years. Never, never bent a
CPU (socket) pin. And my salesman confirmed that this is
a ASUS unique problem. This is why he wants to switch me
to Intel motherboards. (I will stick with Supermicro. They
have World class support.)

A web search found a lot of ASUS CPU socket bent pins too.

-T

 

[Paul in Houston TX]

From my original post:


Seriously, I handle them like nitroglycerin. You can not
believe how precise and gentle I am.

I have been doing this for 17 years. Never, never bent a
CPU (socket) pin. And my salesman confirmed that this is
a ASUS unique problem. This is why he wants to switch me
to Intel motherboards. (I will stick with Supermicro. They
have World class support.)

A web search found a lot of ASUS CPU socket bent pins too.

-T

Mine that had the bent pin was an Asus socket 2011.
I don't know if I bent it or it was like that.
The pin that was ruined controlled one of the ram sticks.
I forgot particulars. I got a Gigabyte X79-UD5 to replace it
and am happy.

 

[Nil]

I am a system builder and I wanted to give my colleagues a heads
up on ASUS motherboards. Seems they work fine when they work, but
good luck with any customer service items. In other words you
are self insuring if you use their stuff.

I've been hearing similar stories for years, though not always in that
much detail.

I have to say that I've used lots of ASUS MBs, and they have been the
most trouble-free of any for me. The worry about what would happen when
I *DID* have a problem has always been at the back of my mind, but it
hasn't happened yet, knock on wood.

 

[David H. Lipman]

From my original post:


Seriously, I handle them like nitroglycerin. You can not
believe how precise and gentle I am.

I have been doing this for 17 years. Never, never bent a
CPU (socket) pin. And my salesman confirmed that this is
a ASUS unique problem. This is why he wants to switch me
to Intel motherboards. (I will stick with Supermicro. They
have World class support.)

A web search found a lot of ASUS CPU socket bent pins too.

I'm sure there are idiots who do not know how to properly handle chips and motherboards
for *many* motherboard manufacturers. But they are not posting this all over Unix and
Windows news groups like you have.

Only 17 years ?

Damn dude, I was replacing Intel 8088/8086 CPUs with NEC V20/V30 chips before ASUSTek was
created and and installing 80386sx modules in replacing 80286 chips about the time they
started up. Back in the 8080/88/86 days we had to manulally install RAM chips in groups of
9 in banks. Never bent a pin, never destroyed a CPU. I'm writing this reply on a ASUSTek
home built system now.

 

[Todd]

Mine that had the bent pin was an Asus socket 2011.
I don't know if I bent it or it was like that.
The pin that was ruined controlled one of the ram sticks.
I forgot particulars. I got a Gigabyte X79-UD5 to replace it
and am happy.

I have never, never had this problem with any other motherboard.

 

[Todd]

I'm sure there are idiots who do not know how to properly handle chips and motherboards
for *many* motherboard manufacturers. But they are not posting this all over Unix and
Windows news groups like you have.

Only 17 years ?

Not to give away my age, but I have been involved in
electronic repair for over 40 years. 17 specifically as
a computer repair/consultant.

Damn dude, I was replacing Intel 8088/8086 CPUs with NEC V20/V30 chips before ASUSTek was
created and and installing 80386sx modules in replacing 80286 chips about the time they
started up. Back in the 8080/88/86 days we had to manulally install RAM chips in groups of
9 in banks. Never bent a pin, never destroyed a CPU. I'm writing this reply on a ASUSTek
home built system now.

Oh when you get an ASUS board up and running they
run damn fine. That was not my point. ASUS has
defective CPU sockets that cause the pins to bend over.
And ASUS will not backup you up. This problem is
only with ASUS motherboards.

Moral of the story, if you sell ASUS motherboards, ASUS will
not be there to back you up. You self insure.

-T

 

[Todd]

I've been hearing similar stories for years, though not always in that
much detail.

I have to say that I've used lots of ASUS MBs, and they have been the
most trouble-free of any for me. The worry about what would happen when
I *DID* have a problem has always been at the back of my mind, but it
hasn't happened yet, knock on wood.

Once they are up and running they run damn fine.

 

[BillW50]

Damn dude, I was replacing Intel 8088/8086 CPUs with NEC V20/V30 chips...

Can I drop a 8086 CPU in a system designed for a NEC V20? I would think
I might have a problem since from what I remember is a V20 is closer to
a 80186 and the BIOS might have a problem with a 8086, you think?

Speaking about the 80186, I can't think of a single computer that ever
used it. Can anybody think of one?

 

[BillW50]

I never went the reverse route. The NEC V20/V30 were faster than their
Intel counterparts. Presumably is the support chips were expecting the
expanded capabilitties of the NEC chips then it would fail.

What are the Intel counterparts to NEC V20/V30? I said the V20 was
compatible with the 80186, but I think I was mistaken. As I think it was
compatible to the 80188 instead, I just can't remember. And what was the
V30 compatible with?

I remember that the V20 and the 80186 or was it a 80188 was almost like
a 286, but missing a few features. So was the V30 fully compatible with
the 286? If not, than what was it compatible with?

And what happened with NEC CPUs? I recall the V20/V30 very well. But did
anything else come out from them?

Another big player back then was MOS Technology. Although it had zero
compatibility with either. Although they may have out sold both Intel
and together NEC back then.

 

[Todd]

Can I drop a 8086 CPU in a system designed for a NEC V20? I would think
I might have a problem since from what I remember is a V20 is closer to
a 80186 and the BIOS might have a problem with a 8086, you think?

Speaking about the 80186, I can't think of a single computer that ever
used it. Can anybody think of one?

My memory, which stinks, is that we went from 8086 to 80286.
Somewhere I think I remember someone designing an a
custom (not a computer) board about an 80186.