BIOS fan control settings - what do they really mean?

I have an ECS Z77h2-A3 V1.2 motherboard, and am trying to figure out how to
adjust the fan control settings in the bios to keep the processor temperature
controled as I like.

The setting available a Range My current setting
Smart Fan start PWM value 0-255 50
Smart Fan start PWM TEMP(-) 10-127 25
DeltaT 0-15 2
Smart Fan Slope PWM value 1-15 15
CPU Fan Full Speed Pffset(-) 12 Created by other settings?
Not adjustable

I really can't make sense of what these settings really mean/do. Can anyone help
me make sense of how these are used?

The first 2 seem pretty clear. The DeltaT seems to be the change in temp between
adjustments, and it would seem like the Slope value should be the difference
added to the PWM setting for each DeltaT advance. But, I cannot seem to find
values that allow the fan to run very quietly at idle, and quickly speed up to
full speed as the temp gets to 60C. I can get this to happen using the eSF fan
control utility, but would like to have the bios set it on boot.

Bob F wrote:
I have an ECS Z77h2-A3 V1.2 motherboard, and am trying to figure out how to
adjust the fan control settings in the bios to keep the processor temperature
controled as I like.

The setting available a Range My current setting
Smart Fan start PWM value 0-255 50
Smart Fan start PWM TEMP(-) 10-127 25
DeltaT 0-15 2
Smart Fan Slope PWM value 1-15 15
CPU Fan Full Speed Pffset(-) 12 Created by other settings?
Not adjustable

I really can't make sense of what these settings really mean/do. Can anyone help
me make sense of how these are used?

The first 2 seem pretty clear. The DeltaT seems to be the change in temp between
adjustments, and it would seem like the Slope value should be the difference
added to the PWM setting for each DeltaT advance. But, I cannot seem to find
values that allow the fan to run very quietly at idle, and quickly speed up to
full speed as the temp gets to 60C. I can get this to happen using the eSF fan
control utility, but would like to have the bios set it on boot.

My guess, is the first value "50" needs to be
adjusted downwards. That's the idle value, as
near as I can tell.

If the fan is not being called for, it should
spin at the "start value". A four pin PWM fan
has its own "hardware minimum" value, so in some
cases, dialing the 50 number down, the fan
minimum becomes evident, rather than the
PWM control behavior as such. If you set the
50 number to 0, the fan will likely continue
to spin, as a minimal spin is specified as a
behavior by the Intel PWM fan spec.

| ___________
| /
| /
| /
50| ________/
|
|
|_________________________
25

Page 64 of the manual, happens to have that curve too.
The picture also tells me, the fan response has hysteresis,
to prevent the fan speed from changing, unless the temperature
moves more than a certain amount from where it is currently.
That would prevent "continuous wander" and might lead to a
more "stair-step" behavior.

http://download.ecs.com.tw/dlfileecs...V1.2%20low.pdf

This is ASCII art for a hysteresis curve, and the path
the hardware would take.

---------
/ /
Speed v ^
/ /
---------

Temp

Years ago, I would find the SuperI/O on the board,
read the part number off it, and download the
spec for the chip. As some of these BIOS interfaces
contain the same number of parameters, as are
shown in the "SmartFan" portion of the chip spec.
Now, on a modern system, this could be provided
by something other than the SuperI/O hardware monitor,
in which case no documentation would be available.

http://www.ecs.com.tw/ECSWebSite/ima...herboard_2.jpg

The rectangular ITE chip with "IT87xx" might be the SuperI/O
in this case. The picture isn't in focus all
that well, and is probably 8x lower res than I'd
like :-) I'm just guessing at the part number I'm seeing
on there. I have at least one IT87xx spec on disk
here, but it's not likely to be the one on your
board.

You can get a flavor of the chip spec writing
style, from the sample page. Try PDF page 119
here for an example (119 as measured from the
front of the doc). It shows a register based
temperature control scheme, with hysteresis.
I see it also has an "off" value, and if the
fan is properly designed, it won't actually
turn off. It will spin at "fan_min" instead
of "start_value".

http://www.rom.by/files/IT8718F-S.pdf

As they say with any fan control - "Good luck..." :-)

Paul

Paul wrote:
Bob F wrote:
I have an ECS Z77h2-A3 V1.2 motherboard, and am trying to figure out
how to adjust the fan control settings in the bios to keep the
processor temperature controled as I like.

The setting available a Range My current setting
Smart Fan start PWM value 0-255 50
Smart Fan start PWM TEMP(-) 10-127 25
DeltaT 0-15 2
Smart Fan Slope PWM value 1-15 15
CPU Fan Full Speed Pffset(-) 12 Created by
other settings? Not adjustable

I really can't make sense of what these settings really mean/do. Can
anyone help me make sense of how these are used?

The first 2 seem pretty clear. The DeltaT seems to be the change in
temp between adjustments, and it would seem like the Slope value
should be the difference added to the PWM setting for each DeltaT
advance. But, I cannot seem to find values that allow the fan to run
very quietly at idle, and quickly speed up to full speed as the temp
gets to 60C. I can get this to happen using the eSF fan control
utility, but would like to have the bios set it on boot.

My guess, is the first value "50" needs to be
adjusted downwards. That's the idle value, as
near as I can tell.

If the fan is not being called for, it should
spin at the "start value". A four pin PWM fan
has its own "hardware minimum" value, so in some
cases, dialing the 50 number down, the fan
minimum becomes evident, rather than the
PWM control behavior as such. If you set the
50 number to 0, the fan will likely continue
to spin, as a minimal spin is specified as a
behavior by the Intel PWM fan spec.

| ___________
| /
| /
| /
50| ________/
|
|
|_________________________
25

Page 64 of the manual, happens to have that curve too.
The picture also tells me, the fan response has hysteresis,
to prevent the fan speed from changing, unless the temperature
moves more than a certain amount from where it is currently.
That would prevent "continuous wander" and might lead to a
more "stair-step" behavior.

http://download.ecs.com.tw/dlfileecs...V1.2%20low.pdf

This is ASCII art for a hysteresis curve, and the path
the hardware would take.

---------
/ /
Speed v ^
/ /
---------

Temp

Years ago, I would find the SuperI/O on the board,
read the part number off it, and download the
spec for the chip. As some of these BIOS interfaces
contain the same number of parameters, as are
shown in the "SmartFan" portion of the chip spec.
Now, on a modern system, this could be provided
by something other than the SuperI/O hardware monitor,
in which case no documentation would be available.

http://www.ecs.com.tw/ECSWebSite/ima...herboard_2.jpg

The rectangular ITE chip with "IT87xx" might be the SuperI/O
in this case. The picture isn't in focus all
that well, and is probably 8x lower res than I'd
like :-) I'm just guessing at the part number I'm seeing
on there. I have at least one IT87xx spec on disk
here, but it's not likely to be the one on your
board.

IT8893E is the one I find on the board.



You can get a flavor of the chip spec writing
style, from the sample page. Try PDF page 119
here for an example (119 as measured from the
front of the doc). It shows a register based
temperature control scheme, with hysteresis.
I see it also has an "off" value, and if the
fan is properly designed, it won't actually
turn off. It will spin at "fan_min" instead
of "start_value".

http://www.rom.by/files/IT8718F-S.pdf

As they say with any fan control - "Good luck..." :-)

The value of 50 seems to produce an initial value around 1070 rpm. Starting the
torture test on prime95 increases the rpm to 1085 with the values I posted.
Actually, after awhile it got up to
1100 rpm with the cores at 57,66,63 and 62 C. That's with the slope set at the
maximum of 15. I guess I need to try small #'s for slope.

I changed the slope to 3 from 15 with DeltaT staying at 2. Virtually no change.
Changed the Start PWM value from 50 to 40 - no change.
This is wierd. The eSF program clearly works with the CPU fan connector, but the
BIOS fan control settings seem to change nothing.


I wish I could see what settings the eSF program produces.

Bob F wrote:
Paul wrote:
Bob F wrote:
I have an ECS Z77h2-A3 V1.2 motherboard, and am trying to figure out
how to adjust the fan control settings in the bios to keep the
processor temperature controled as I like.

The setting available a Range My current setting
Smart Fan start PWM value 0-255 50
Smart Fan start PWM TEMP(-) 10-127 25
DeltaT 0-15 2
Smart Fan Slope PWM value 1-15 15 PWM value / unit
CPU Fan Full Speed Pffset(-) 12 Created by
other settings? Not adjustable

I really can't make sense of what these settings really mean/do. Can
anyone help me make sense of how these are used?

The first 2 seem pretty clear. The DeltaT seems to be the change in
temp between adjustments, and it would seem like the Slope value
should be the difference added to the PWM setting for each DeltaT
advance. But, I cannot seem to find values that allow the fan to run
very quietly at idle, and quickly speed up to full speed as the temp
gets to 60C. I can get this to happen using the eSF fan control
utility, but would like to have the bios set it on boot.

My guess, is the first value "50" needs to be
adjusted downwards. That's the idle value, as
near as I can tell.

If the fan is not being called for, it should
spin at the "start value". A four pin PWM fan
has its own "hardware minimum" value, so in some
cases, dialing the 50 number down, the fan
minimum becomes evident, rather than the
PWM control behavior as such. If you set the
50 number to 0, the fan will likely continue
to spin, as a minimal spin is specified as a
behavior by the Intel PWM fan spec.

| ___________
| /
| /
| /
50| ________/
|
|
|_________________________
25

Page 64 of the manual, happens to have that curve too.
The picture also tells me, the fan response has hysteresis,
to prevent the fan speed from changing, unless the temperature
moves more than a certain amount from where it is currently.
That would prevent "continuous wander" and might lead to a
more "stair-step" behavior.

http://download.ecs.com.tw/dlfileecs...V1.2%20low.pdf

This is ASCII art for a hysteresis curve, and the path
the hardware would take.

---------
/ /
Speed v ^
/ /
---------

Temp

Years ago, I would find the SuperI/O on the board,
read the part number off it, and download the
spec for the chip. As some of these BIOS interfaces
contain the same number of parameters, as are
shown in the "SmartFan" portion of the chip spec.
Now, on a modern system, this could be provided
by something other than the SuperI/O hardware monitor,
in which case no documentation would be available.

http://www.ecs.com.tw/ECSWebSite/ima...herboard_2.jpg

The rectangular ITE chip with "IT87xx" might be the SuperI/O
in this case. The picture isn't in focus all
that well, and is probably 8x lower res than I'd
like :-) I'm just guessing at the part number I'm seeing
on there. I have at least one IT87xx spec on disk
here, but it's not likely to be the one on your
board.

IT8893E is the one I find on the board.



You can get a flavor of the chip spec writing
style, from the sample page. Try PDF page 119
here for an example (119 as measured from the
front of the doc). It shows a register based
temperature control scheme, with hysteresis.
I see it also has an "off" value, and if the
fan is properly designed, it won't actually
turn off. It will spin at "fan_min" instead
of "start_value".

http://www.rom.by/files/IT8718F-S.pdf

As they say with any fan control - "Good luck..." :-)

The value of 50 seems to produce an initial value around 1070 rpm.
Starting the torture test on prime95 increases the rpm to 1085 with
the values I posted. Actually, after awhile it got up to
1100 rpm with the cores at 57,66,63 and 62 C. That's with the slope
set at the maximum of 15. I guess I need to try small #'s for slope.

I changed the slope to 3 from 15 with DeltaT staying at 2. Virtually
no change. Changed the Start PWM value from 50 to 40 - no change.
This is wierd. The eSF program clearly works with the CPU fan
connector, but the BIOS fan control settings seem to change nothing.

Actually, If I set the BIOS to NORMAL setting, the fan runs at 2100 rpm with
little change from idle to torture test. That's settings 180, 30, 3, 10 , 23.




I wish I could see what settings the eSF program produces.

Bob F wrote:
Paul wrote:
Bob F wrote:
I have an ECS Z77h2-A3 V1.2 motherboard, and am trying to figure out
how to adjust the fan control settings in the bios to keep the
processor temperature controled as I like.

The setting available a Range My current setting
Smart Fan start PWM value 0-255 50
Smart Fan start PWM TEMP(-) 10-127 25
DeltaT 0-15 2
Smart Fan Slope PWM value 1-15 15
CPU Fan Full Speed Pffset(-) 12 Created by
other settings? Not adjustable

I really can't make sense of what these settings really mean/do. Can
anyone help me make sense of how these are used?

The first 2 seem pretty clear. The DeltaT seems to be the change in
temp between adjustments, and it would seem like the Slope value
should be the difference added to the PWM setting for each DeltaT
advance. But, I cannot seem to find values that allow the fan to run
very quietly at idle, and quickly speed up to full speed as the temp
gets to 60C. I can get this to happen using the eSF fan control
utility, but would like to have the bios set it on boot.
My guess, is the first value "50" needs to be
adjusted downwards. That's the idle value, as
near as I can tell.

If the fan is not being called for, it should
spin at the "start value". A four pin PWM fan
has its own "hardware minimum" value, so in some
cases, dialing the 50 number down, the fan
minimum becomes evident, rather than the
PWM control behavior as such. If you set the
50 number to 0, the fan will likely continue
to spin, as a minimal spin is specified as a
behavior by the Intel PWM fan spec.

| ___________
| /
| /
| /
50| ________/
|
|
|_________________________
25

Page 64 of the manual, happens to have that curve too.
The picture also tells me, the fan response has hysteresis,
to prevent the fan speed from changing, unless the temperature
moves more than a certain amount from where it is currently.
That would prevent "continuous wander" and might lead to a
more "stair-step" behavior.

http://download.ecs.com.tw/dlfileecs...V1.2%20low.pdf

This is ASCII art for a hysteresis curve, and the path
the hardware would take.

---------
/ /
Speed v ^
/ /
---------

Temp

Years ago, I would find the SuperI/O on the board,
read the part number off it, and download the
spec for the chip. As some of these BIOS interfaces
contain the same number of parameters, as are
shown in the "SmartFan" portion of the chip spec.
Now, on a modern system, this could be provided
by something other than the SuperI/O hardware monitor,
in which case no documentation would be available.

http://www.ecs.com.tw/ECSWebSite/ima...herboard_2.jpg

The rectangular ITE chip with "IT87xx" might be the SuperI/O
in this case. The picture isn't in focus all
that well, and is probably 8x lower res than I'd
like :-) I'm just guessing at the part number I'm seeing
on there. I have at least one IT87xx spec on disk
here, but it's not likely to be the one on your
board.

IT8893E is the one I find on the board.


You can get a flavor of the chip spec writing
style, from the sample page. Try PDF page 119
here for an example (119 as measured from the
front of the doc). It shows a register based
temperature control scheme, with hysteresis.
I see it also has an "off" value, and if the
fan is properly designed, it won't actually
turn off. It will spin at "fan_min" instead
of "start_value".

http://www.rom.by/files/IT8718F-S.pdf

As they say with any fan control - "Good luck..." :-)

The value of 50 seems to produce an initial value around 1070 rpm. Starting the
torture test on prime95 increases the rpm to 1085 with the values I posted.
Actually, after awhile it got up to
1100 rpm with the cores at 57,66,63 and 62 C. That's with the slope set at the
maximum of 15. I guess I need to try small #'s for slope.

I changed the slope to 3 from 15 with DeltaT staying at 2. Virtually no change.
Changed the Start PWM value from 50 to 40 - no change.
This is wierd. The eSF program clearly works with the CPU fan connector, but the
BIOS fan control settings seem to change nothing.


I wish I could see what settings the eSF program produces.



Speedfan

http://www.almico.com/speedfan451.exe

Install and run. Wait until probe finished.

Configure button (main window)
Advanced tab

In the "Chip" pulldown, you'll have to spot a matching
IT8893E bus entry, to match what the probe detected.

Now, mine doesn't show fan controls at all. But maybe if you
use it after eSF, you'll get some idea. I have some PWM generic
entries that are "OFF".

That's the only tool I know of.

My guess is, something on the system changes the BIOS settings
after the BIOS has handed off. It could be that eSF has installed
a Startup task to do something.

*******

And while I have a copy of Speedfan installed, *nothing* messes
with my fan speeds here. All fans run at full speed. I have a
Zalman resistor device to trim the front fan down a bit.

Even the video card fan was disconnected from the on-card control
and runs from a fixed speed source. I carefully tested the video card
at full load, and adjusted the static speed for adequate cooling.
So my video card is not thermally regulated. This prevents the NVidia
GPU fan from wailing at 100% when Linux is running. Linux and
Windows make exactly the same noise level here. My video card has
a relatively modest power level, which is why I can do this mod.
A card with 200W+ power level, you're best advised to use the
provided cooling solution without modification.

Paul

Paul wrote:
Bob F wrote:
Paul wrote:
Bob F wrote:
I have an ECS Z77h2-A3 V1.2 motherboard, and am trying to figure
out how to adjust the fan control settings in the bios to keep the
processor temperature controled as I like.

The setting available a Range My current setting
Smart Fan start PWM value 0-255 50
Smart Fan start PWM TEMP(-) 10-127 25
DeltaT 0-15 2
Smart Fan Slope PWM value 1-15 15
CPU Fan Full Speed Pffset(-) 12 Created by
other settings? Not adjustable

I really can't make sense of what these settings really mean/do.
Can anyone help me make sense of how these are used?

The first 2 seem pretty clear. The DeltaT seems to be the change in
temp between adjustments, and it would seem like the Slope value
should be the difference added to the PWM setting for each DeltaT
advance. But, I cannot seem to find values that allow the fan to
run very quietly at idle, and quickly speed up to full speed as
the temp gets to 60C. I can get this to happen using the eSF fan
control utility, but would like to have the bios set it on boot.
My guess, is the first value "50" needs to be
adjusted downwards. That's the idle value, as
near as I can tell.

If the fan is not being called for, it should
spin at the "start value". A four pin PWM fan
has its own "hardware minimum" value, so in some
cases, dialing the 50 number down, the fan
minimum becomes evident, rather than the
PWM control behavior as such. If you set the
50 number to 0, the fan will likely continue
to spin, as a minimal spin is specified as a
behavior by the Intel PWM fan spec.

| ___________
| /
| /
| /
50| ________/
|
|
|_________________________
25

Page 64 of the manual, happens to have that curve too.
The picture also tells me, the fan response has hysteresis,
to prevent the fan speed from changing, unless the temperature
moves more than a certain amount from where it is currently.
That would prevent "continuous wander" and might lead to a
more "stair-step" behavior.

http://download.ecs.com.tw/dlfileecs...V1.2%20low.pdf

This is ASCII art for a hysteresis curve, and the path
the hardware would take.

---------
/ /
Speed v ^
/ /
---------

Temp

Years ago, I would find the SuperI/O on the board,
read the part number off it, and download the
spec for the chip. As some of these BIOS interfaces
contain the same number of parameters, as are
shown in the "SmartFan" portion of the chip spec.
Now, on a modern system, this could be provided
by something other than the SuperI/O hardware monitor,
in which case no documentation would be available.

http://www.ecs.com.tw/ECSWebSite/ima...herboard_2.jpg

The rectangular ITE chip with "IT87xx" might be the SuperI/O
in this case. The picture isn't in focus all
that well, and is probably 8x lower res than I'd
like :-) I'm just guessing at the part number I'm seeing
on there. I have at least one IT87xx spec on disk
here, but it's not likely to be the one on your
board.

IT8893E is the one I find on the board.


You can get a flavor of the chip spec writing
style, from the sample page. Try PDF page 119
here for an example (119 as measured from the
front of the doc). It shows a register based
temperature control scheme, with hysteresis.
I see it also has an "off" value, and if the
fan is properly designed, it won't actually
turn off. It will spin at "fan_min" instead
of "start_value".

http://www.rom.by/files/IT8718F-S.pdf

As they say with any fan control - "Good luck..." :-)

The value of 50 seems to produce an initial value around 1070 rpm.
Starting the torture test on prime95 increases the rpm to 1085 with
the values I posted. Actually, after awhile it got up to
1100 rpm with the cores at 57,66,63 and 62 C. That's with the slope
set at the maximum of 15. I guess I need to try small #'s for slope.

I changed the slope to 3 from 15 with DeltaT staying at 2. Virtually
no change. Changed the Start PWM value from 50 to 40 - no change.
This is wierd. The eSF program clearly works with the CPU fan
connector, but the BIOS fan control settings seem to change nothing.


I wish I could see what settings the eSF program produces.



Speedfan

http://www.almico.com/speedfan451.exe

Install and run. Wait until probe finished.

Configure button (main window)
Advanced tab

In the "Chip" pulldown, you'll have to spot a matching
IT8893E bus entry, to match what the probe detected.

Speedfan shows the device IT8728F. WTF? I guess the chip I found was something
else?



Now, mine doesn't show fan controls at all. But maybe if you
use it after eSF, you'll get some idea. I have some PWM generic
entries that are "OFF".

Nothing Speedfan shows seems to relate to the Bios Setting choices.



That's the only tool I know of.

My guess is, something on the system changes the BIOS settings
after the BIOS has handed off. It could be that eSF has installed
a Startup task to do something.

eSF clearly only works when run or when enabled to run on boot. Otherwise, the
BIOS settings determine the fan operation, but not in an understandable way.


*******

And while I have a copy of Speedfan installed, *nothing* messes
with my fan speeds here. All fans run at full speed. I have a
Zalman resistor device to trim the front fan down a bit.

Even the video card fan was disconnected from the on-card control
and runs from a fixed speed source. I carefully tested the video card
at full load, and adjusted the static speed for adequate cooling.
So my video card is not thermally regulated. This prevents the NVidia
GPU fan from wailing at 100% when Linux is running. Linux and
Windows make exactly the same noise level here. My video card has
a relatively modest power level, which is why I can do this mod.
A card with 200W+ power level, you're best advised to use the
provided cooling solution without modification.

Paul

Bob F wrote:
Paul wrote:
Bob F wrote:
Paul wrote:
Bob F wrote:
I have an ECS Z77h2-A3 V1.2 motherboard, and am trying to figure
out how to adjust the fan control settings in the bios to keep the
processor temperature controled as I like.

The setting available a Range My current setting
Smart Fan start PWM value 0-255 50
Smart Fan start PWM TEMP(-) 10-127 25
DeltaT 0-15 2
Smart Fan Slope PWM value 1-15 15
CPU Fan Full Speed Pffset(-) 12 Created by
other settings? Not adjustable

I really can't make sense of what these settings really mean/do.
Can anyone help me make sense of how these are used?

The first 2 seem pretty clear. The DeltaT seems to be the change in
temp between adjustments, and it would seem like the Slope value
should be the difference added to the PWM setting for each DeltaT
advance. But, I cannot seem to find values that allow the fan to
run very quietly at idle, and quickly speed up to full speed as
the temp gets to 60C. I can get this to happen using the eSF fan
control utility, but would like to have the bios set it on boot.
My guess, is the first value "50" needs to be
adjusted downwards. That's the idle value, as
near as I can tell.

If the fan is not being called for, it should
spin at the "start value". A four pin PWM fan
has its own "hardware minimum" value, so in some
cases, dialing the 50 number down, the fan
minimum becomes evident, rather than the
PWM control behavior as such. If you set the
50 number to 0, the fan will likely continue
to spin, as a minimal spin is specified as a
behavior by the Intel PWM fan spec.

| ___________
| /
| /
| /
50| ________/
|
|
|_________________________
25

Page 64 of the manual, happens to have that curve too.
The picture also tells me, the fan response has hysteresis,
to prevent the fan speed from changing, unless the temperature
moves more than a certain amount from where it is currently.
That would prevent "continuous wander" and might lead to a
more "stair-step" behavior.

http://download.ecs.com.tw/dlfileecs...V1.2%20low.pdf

This is ASCII art for a hysteresis curve, and the path
the hardware would take.

---------
/ /
Speed v ^
/ /
---------

Temp

Years ago, I would find the SuperI/O on the board,
read the part number off it, and download the
spec for the chip. As some of these BIOS interfaces
contain the same number of parameters, as are
shown in the "SmartFan" portion of the chip spec.
Now, on a modern system, this could be provided
by something other than the SuperI/O hardware monitor,
in which case no documentation would be available.

http://www.ecs.com.tw/ECSWebSite/ima...herboard_2.jpg

The rectangular ITE chip with "IT87xx" might be the SuperI/O
in this case. The picture isn't in focus all
that well, and is probably 8x lower res than I'd
like :-) I'm just guessing at the part number I'm seeing
on there. I have at least one IT87xx spec on disk
here, but it's not likely to be the one on your
board.
IT8893E is the one I find on the board.


You can get a flavor of the chip spec writing
style, from the sample page. Try PDF page 119
here for an example (119 as measured from the
front of the doc). It shows a register based
temperature control scheme, with hysteresis.
I see it also has an "off" value, and if the
fan is properly designed, it won't actually
turn off. It will spin at "fan_min" instead
of "start_value".

http://www.rom.by/files/IT8718F-S.pdf

As they say with any fan control - "Good luck..." :-)
The value of 50 seems to produce an initial value around 1070 rpm.
Starting the torture test on prime95 increases the rpm to 1085 with
the values I posted. Actually, after awhile it got up to
1100 rpm with the cores at 57,66,63 and 62 C. That's with the slope
set at the maximum of 15. I guess I need to try small #'s for slope.

I changed the slope to 3 from 15 with DeltaT staying at 2. Virtually
no change. Changed the Start PWM value from 50 to 40 - no change.
This is wierd. The eSF program clearly works with the CPU fan
connector, but the BIOS fan control settings seem to change nothing.


I wish I could see what settings the eSF program produces.


Speedfan

http://www.almico.com/speedfan451.exe

Install and run. Wait until probe finished.

Configure button (main window)
Advanced tab

In the "Chip" pulldown, you'll have to spot a matching
IT8893E bus entry, to match what the probe detected.

Speedfan shows the device IT8728F. WTF? I guess the chip I found was something
else?


Now, mine doesn't show fan controls at all. But maybe if you
use it after eSF, you'll get some idea. I have some PWM generic
entries that are "OFF".

Nothing Speedfan shows seems to relate to the Bios Setting choices.


That's the only tool I know of.

My guess is, something on the system changes the BIOS settings
after the BIOS has handed off. It could be that eSF has installed
a Startup task to do something.

eSF clearly only works when run or when enabled to run on boot. Otherwise, the
BIOS settings determine the fan operation, but not in an understandable way.


The IT8893E is the PCI Express to PCI bridge. Intel has
removed PCI bus from the Southbridge. Motherboard manufacturers
spend extra money, adding an x1 bridge to make a PCI bus.
For the convenience of their customers (people with PCI
sound cards they would like to reuse). Intel considers
the PCI bus to be "Legacy", just like a PS/2 connector.
Which is why they removed it.

http://www.ite.com.tw/en/product/category?cid=1

But ITE is losing it, in terms of datasheets.
You have to search other sites, to find out about the
products. The iteusa.com web site was shut down.
Only head office ite.com.tw remains.

*******

The IT8728F part number is part of IT87xx series, so
that makes sense. And some other site may have a copy
of the PDF. This is the first site that came up in
a search.

"IT8728F Preliminary Specification V0.4.2"

http://www.datasheetspdf.com/datasheet/IT8728F.html

(This site probably doesn't allow direct linking, but
the browser records this link for the 2,219,569 byte file.)

http://www.datasheetspdf.com/datashe....php?id=770616

PDF page 112 has the same diagram as I showed in the previous example.

*******

I don't know of a practical way to dump the registers. With
a copy of "giveio", you can punch through the barrier to
hardware access. But you'd still have to find the recipe
to reach the LPC bus, do a probe, find the chip, then
dump the registers. (Even though Speedfan might list
the chip as being on SMBUS, it's actually on LPC. And
this charade has existed for years - Linux started this
tradition of mis-naming low speed I/O busses. You will
find "LPC" sprinkled though the above PDF file. LPC
is a nibble wide bus that runs much faster than an
SMBUS, and also doesn't corrupt data when multiple
requests are in flight.)

I seem to remember dumping fan divisors in the past, using
some tool. (The idea was, to make fan speed measurement
work at low RPMs - some motherboards stop counting if
the fan is below 1800 RPMS.) I can't remember if that
was the (discontinued) MBM5 or it was Speedfan.

Sorry I can't shine any more light on the mystery.

Paul