ACPI power supply relation?

  • Thread starter Thread starter Steve Miller
  • Start date Start date
Steve said:
I got the adapter and put it in; checked the voltages; still the
same. :(

Just to refresh our memories, when is it that the fan stops spinning ?

S1 Standby Fan continues to spin
S3 Suspend to RAM Fan stops spinning, +5VSB powers RAM
S4 Hibernate to disk Fan stops spinning, power supply state is immaterial

Recovery from any of the above, fan should spin. For thermally
controlled fans, the fan spins faster as the processor (or
the surrounding air) gets hotter.

If there is a problem with the motherboard header behavior,
powering the fan, you need an adapter that runs off a Molex.
The adapter must have a "yellow wire" with the RPM signal,
and on the end of that, a standard fan connector, which goes
to the motherboard. Some motherboards will shut off, if they
fail to see an RPM signal delivered to the three or four pin
CPU fan header.

The problem with fan adapters, is not all of them have the
necessary wiring. This one for example, does solve the problem
of getting +12V from a Molex connector, to a three or four pin fan.
But there is no provision for the RPM signal, which would be a
separate cable with the yellow wire on it. So this one would
keep the fan happy, but unless a yellow (RPM) wire runs from
the small connector in this picture, to the RPM pin on the
motherboard CPU fan header, the BIOS may choose to shut down
(or complain). You could try one of these and see what
happens.

http://www.startech.com/item/CPUFAN...Internal-Power-LP4-CPU-Fan-Adapter-Cable.aspx

The last "good" cable of the type I describe, may have come
with a ThermalTake heatsink of some sort. If I needed one now,
I have TX3 male and female plugs here, so I can build my own.
And the place I bought those, doesn't carry them any more,
so whatever I've got left, is all there is.

Try entering 37-603-5 and 37-6203-5 here, as these are the
three pin female and three pin male connectors I've got, for
building fan cables. They don't care enough about the products,
to show pictures of each type. They have two datasheets, showing
connectors of various widths. In any case, these would be
examples of what you'd need, to build your own fan cabling.
You start with a Molex Y cable (Radio Shack used to carry those),
for the Molex part, and then you'd use connectors like these,
to cable up the fan end. My local supplier doesn't carry these
particular bags any more. There were five connectors per bag,
including the necessary crimp pins. One is crimp type, the
other is solder type (you use whatever you can find...).

http://www.mode-elec.com/

http://www.mode-elec.com/pdf/connectors/37-6XX.pdf
http://www.mode-elec.com/pdf/connectors/37-62XX.pdf

Paul
 
Just to refresh our memories, when is it that the fan stops spinning ?

S1 Standby                Fan continues to spin
S3 Suspend to RAM         Fan stops spinning, +5VSB powers RAM
S4 Hibernate to disk      Fan stops spinning, power supply state isimmaterial

Recovery from any of the above, fan should spin. For thermally
controlled fans, the fan spins faster as the processor (or
the surrounding air) gets hotter.

If there is a problem with the motherboard header behavior,
powering the fan, you need an adapter that runs off a Molex.
The adapter must have a "yellow wire" with the RPM signal,
and on the end of that, a standard fan connector, which goes
to the motherboard. Some motherboards will shut off, if they
fail to see an RPM signal delivered to the three or four pin
CPU fan header.

The problem with fan adapters, is not all of them have the
necessary wiring. This one for example, does solve the problem
of getting +12V from a Molex connector, to a three or four pin fan.
But there is no provision for the RPM signal, which would be a
separate cable with the yellow wire on it. So this one would
keep the fan happy, but unless a yellow (RPM) wire runs from
the small connector in this picture, to the RPM pin on the
motherboard CPU fan header, the BIOS may choose to shut down
(or complain). You could try one of these and see what
happens.

http://www.startech.com/item/CPUFANADAPT-6-TX3-to-Large-Internal-Powe...

The last "good" cable of the type I describe, may have come
with a ThermalTake heatsink of some sort. If I needed one now,
I have TX3 male and female plugs here, so I can build my own.
And the place I bought those, doesn't carry them any more,
so whatever I've got left, is all there is.

Try entering 37-603-5 and 37-6203-5 here, as these are the
three pin female and three pin male connectors I've got, for
building fan cables. They don't care enough about the products,
to show pictures of each type. They have two datasheets, showing
connectors of various widths. In any case, these would be
examples of what you'd need, to build your own fan cabling.
You start with a Molex Y cable (Radio Shack used to carry those),
for the Molex part, and then you'd use connectors like these,
to cable up the fan end. My local supplier doesn't carry these
particular bags any more. There were five connectors per bag,
including the necessary crimp pins. One is crimp type, the
other is solder type (you use whatever you can find...).

http://www.mode-elec.com/

http://www.mode-elec.com/pdf/connec.../www.mode-elec.com/pdf/connectors/37-62XX.pdf

    Paul


Hello Paul,

I'm glad you got back to me. I'm not exactly sure what to think now.
You already
knew it's not being the missing voltage which I had to try to make
sure, firstly.

I found an blog meanwhile http://blog.stuart.shelton.me/archives/124#ACPI
that let
me assume to fix this I have to flash the BIOS, which I'm not going to
do.

Quote:

What is unfortunate, then, is that Jetway have broken ACPI DSDT
tables, built
with Microsoft's non standards-compliant ASL compiler, installed with
their BIOS.

Could it be in the programming? And, was there an other workaround for
Windows XP other than to flash the BIOS?

Steve
 
Steve said:
Hello Paul,

I'm glad you got back to me. I'm not exactly sure what to think now.
You already
knew it's not being the missing voltage which I had to try to make
sure, firstly.

I found an blog meanwhile http://blog.stuart.shelton.me/archives/124#ACPI
that let
me assume to fix this I have to flash the BIOS, which I'm not going to
do.

Quote:

What is unfortunate, then, is that Jetway have broken ACPI DSDT
tables, built
with Microsoft's non standards-compliant ASL compiler, installed with
their BIOS.

Could it be in the programming? And, was there an other workaround for
Windows XP other than to flash the BIOS?

Steve

Can we go back over the symptoms again ? <--------------

ACPI DSDT is part of the BIOS. I tried looking at that ASL stuff
before, but didn't understand it. In that blog, Stuart fixed his,
by overriding the tables offered by the BIOS, for his Linux OS.
(In other words, he inserted a patched ACPI DSDT table into
his OS, taking the place of the table that would normally be
fetched from the BIOS during boot.)

Is that a reasonable way to go about things, hacking each OS
to try to correct a BIOS problem ? Does a Windows OS even
allow that ?

So please repeat what your symptoms are. If the problem is
limited to a fan that refuses to run, you can bypass that
with a separate power cable. But if the problems are more
substantial, then perhaps working around them isn't practical.

If worse came to worse, you could always look for another
780G motherboard. The idea being, that you could drop a
new board into place, without too much change to your
current OS. As long as you didn't screw up the BIOS
settings too much, before your first boot :-)

I cannot even get to the stupid jetway.com.tw site right
now. The server seems to be refusing connections. So I
can't even review the materials they offer for the PA78M4-H.

Paul
 
Can we go back over the symptoms again ?  <--------------

ACPI DSDT is part of the BIOS. I tried looking at that ASL stuff
before, but didn't understand it. In that blog, Stuart fixed his,
by overriding the tables offered by the BIOS, for his Linux OS.
(In other words, he inserted a patched ACPI DSDT table into
his OS, taking the place of the table that would normally be
fetched from the BIOS during boot.)

Is that a reasonable way to go about things, hacking each OS
to try to correct a BIOS problem ? Does a Windows OS even
allow that ?

So please repeat what your symptoms are. If the problem is
limited to a fan that refuses to run, you can bypass that
with a separate power cable. But if the problems are more
substantial, then perhaps working around them isn't practical.

If worse came to worse, you could always look for another
780G motherboard. The idea being, that you could drop a
new board into place, without too much change to your
current OS. As long as you didn't screw up the BIOS
settings too much, before your first boot :-)

I cannot even get to the stupid jetway.com.tw site right
now. The server seems to be refusing connections. So I
can't even review the materials they offer for the PA78M4-H.

    Paul

Well, I considered an other motherboard seriously, even before I
purchased the Jetway. But, I think you'll agree, Jetway was a cool
name for an mainboard manufacturer.

I have learned, meanwhile, any nation or region got it's, how to say,
specialties. If you're talking about the financial craft you're
talking about a Swiss bank account probably, if talking about shows
it's the Dr. Martens brand perhaps, and so forth.

Well, I just do see things like that, by the time being. Therefore I
still think its ok to give them a try. The board was feeling very well
in my box, and I think I spent the money rightly.

The symptom now was the fan stops spinning after resume from
hibernation. The monitor doesn't go off as well. after the set time.

I got an other fan which was connected to the power supply directly,
Perhaps that's the way to go? This one was always working on the
previous PC silently. If not I will go for an other heat sink.

The other idea was, Linux gives you these opportunities on many sides,
as far as I know.

It pretty cool as for the time being I think. The ACPI specification
was an Intel invention, which might be the actual cause it's not
working, for several reasons. Or the reason Jetway was not complying
to it, however.

As for what I think, no one so far has understood, computers are not
made of hardware and software, actually but electricity as well.
Therefore, the network a single computer yet was, was not understood
in time. Jetway seems not to be making this mistake.

Steve

As well, I got an error message after trying to resume from
hibernation saying the computer was not fully ACPI compliant. I should
update the hard BIOS. :)
 
Steve said:
Well, I considered an other motherboard seriously, even before I
purchased the Jetway. But, I think you'll agree, Jetway was a cool
name for an mainboard manufacturer.

I have learned, meanwhile, any nation or region got it's, how to say,
specialties. If you're talking about the financial craft you're
talking about a Swiss bank account probably, if talking about shows
it's the Dr. Martens brand perhaps, and so forth.

Well, I just do see things like that, by the time being. Therefore I
still think its ok to give them a try. The board was feeling very well
in my box, and I think I spent the money rightly.

The symptom now was the fan stops spinning after resume from
hibernation. The monitor doesn't go off as well. after the set time.

I got an other fan which was connected to the power supply directly,
Perhaps that's the way to go? This one was always working on the
previous PC silently. If not I will go for an other heat sink.

The other idea was, Linux gives you these opportunities on many sides,
as far as I know.

It pretty cool as for the time being I think. The ACPI specification
was an Intel invention, which might be the actual cause it's not
working, for several reasons. Or the reason Jetway was not complying
to it, however.

As for what I think, no one so far has understood, computers are not
made of hardware and software, actually but electricity as well.
Therefore, the network a single computer yet was, was not understood
in time. Jetway seems not to be making this mistake.

Steve

As well, I got an error message after trying to resume from
hibernation saying the computer was not fully ACPI compliant. I should
update the hard BIOS. :)

OK, so for the moment, we don't understand why the fan won't spin, after
the system recovers from hibernation. It implies the fan header
is controllable. But it doesn't suggest why that control
doesn't function. (My initial reaction would be, is the fan
an ACPI item ?).

"I got an error message after trying to resume from
hibernation saying the computer was not fully ACPI compliant."

That is a more interesting message. If that were truly the case, then
you wouldn't have been able to do an ACPI installation of WinXP. If
you go to Device Manager, and examine the "Computer" entry, it would
have ACPI in the name, if an ACPI HAL was installed. I had a computer
once, that wasn't ACPI compatible, and it wouldn't install ACPI.
I ended up with a "Standard PC" as the entry for "Computer".

So maybe what the error message is saying, is the machine is
partially compliant.

At this point, you know there is some kind of ACPI problem,
but your symptoms might not be exactly like that blog.

Are you able to get to jetway.com.tw right now. I still
cannot get there, "connection refused". I'd at least
want to read any warnings they have, about the version
of flasher to use for flashing the BIOS. And look for
release notes, to see if the update has addressed your
problem at all.

You'd think a number of people would be pissed, if
ACPI related features are glitchy. And a lot of
people would freak out, if the fan didn't spin on
recovery. You would think that would merit some response
from them. Most motherboard companies seem to try to
align the contents of their manual (their "contract" with
the customer), to how the hardware behaves. So they should
continue BIOS development, until all the features mentioned
in the manual, have been implemented. (My Asrock board
being an exception I guess.)

Another way to solve this, would be to look for an alternate
source of BIOS. But that isn't exactly foolproof either,
as you don't know what you're getting, when you try that.
There is at least one company, still selling BIOS.

Paul
 
OK, so for the moment, we don't understand why the fan won't spin, after
the system recovers from hibernation. It implies the fan header
is controllable. But it doesn't suggest why that control
doesn't function. (My initial reaction would be, is the fan
an ACPI item ?).

"I got an error message after trying to resume from
  hibernation saying the computer was not fully ACPI compliant."

That is a more interesting message. If that were truly the case, then
you wouldn't have been able to do an ACPI installation of WinXP. If
you go to Device Manager, and examine the "Computer" entry, it would
have ACPI in the name, if an ACPI HAL was installed. I had a computer
once, that wasn't ACPI compatible, and it wouldn't install ACPI.
I ended up with a "Standard PC" as the entry for "Computer".

So maybe what the error message is saying, is the machine is
partially compliant.

At this point, you know there is some kind of ACPI problem,
but your symptoms might not be exactly like that blog.

Are you able to get to jetway.com.tw right now. I still
cannot get there, "connection refused". I'd at least
want to read any warnings they have, about the version
of flasher to use for flashing the BIOS. And look for
release notes, to see if the update has addressed your
problem at all.

You'd think a number of people would be pissed, if
ACPI related features are glitchy. And a lot of
people would freak out, if the fan didn't spin on
recovery. You would think that would merit some response
from them. Most motherboard companies seem to try to
align the contents of their manual (their "contract" with
the customer), to how the hardware behaves. So they should
continue BIOS development, until all the features mentioned
in the manual, have been implemented. (My Asrock board
being an exception I guess.)

Another way to solve this, would be to look for an alternate
source of BIOS. But that isn't exactly foolproof either,
as you don't know what you're getting, when you try that.
There is at least one company, still selling BIOS.

    Paul

Well Paul, I do think to flash the BIOS might be such a good idea.

I read if something goes wrong the mainboard renders useless.
My current situation was, I only have got this one computer and have
no clue on what to do in case of flashing the BIOS would fail!

Steve
 
Steve said:
Well Paul, I do think to flash the BIOS might be such a good idea.

I read if something goes wrong the mainboard renders useless.
My current situation was, I only have got this one computer and have
no clue on what to do in case of flashing the BIOS would fail!

Steve

The key to reducing the risk, is if the BIOS chip is socketed.
But I did see the word "SPI" at some point, in looking at your
board. That means the BIOS could be an 8 pin chip which is
soldered to the board. And if it is soldered, it won't be
easy to recover from a bad BIOS flash.

And if you only have the one computer, that also weighs against
flashing the BIOS.

Some motherboard makers are pretty good about this. If the
board is bricked, and has SPI flash, instead of socketed PLCC
flash, they'll reflash it for you. So some of them cover
this under the warranty, and you'd pay shipping perhaps.
But Jetway is a smaller company, so I don't know what
their policy would be.

I still cannot get to the Jetway site in Taiwan. I hope I'm
not doing something stupid.

Paul
 
This address http://www.jetway.com.tw/jw/ was working for me. But, I don't have a floppy drive
anyway.

If I'm gonna be kind with the computer it allows me to resume from hibernate once. Without the
fan getting on strike, but only one time . I'm just trying if it was the same in BIOS S1 or S3.

I think I remember, when I flash the BIOS the smart fan configuration would be gona which was
unacceptable. It's simply too noisy without.

Steve
 
Back
Top