[lm-sensors] MSI K9N4 SLI-F sensors-detect says use w83627ehf

[lm-sensors] MSI K9N4 SLI-F sensors-detect says use w83627ehf

[David Hubbard]

Hi Curt,

On 3/20/07, Curt Blank <curt at curtronics.com> wrote:
> Motherboard: MSI K9N4 SLI-F
> Chipset: nVidia nForce 500 SLI
> SuSE Linux 10.1
> Kernel: 2.6.16.27-0.9
> lm-sensors: 2.10.0-10 (from SuSE 10.1 distro)
>
> I run sensors-detect that came with the SuSE distribution and it can't
> find any sensors. So then I download and run the latest sensors-detect and
> it tells me to use the w83627ehf driver. But when I try it fails:
>
> # modprobe w83627ehf
> FATAL: Error inserting w83627ehf
> (/lib/modules/2.6.16.27-0.9-smp-bnfsrvr/kernel/drivers/hwmon/w83627ehf.ko):
> No such device
>
> It has a confidence of 9 but it fails. Am I missing something here?
> w83627ehf has been in the kernel since 2.6.13, but do I maybe need a newer
> version then I have?
>
> These are the modules currently installed:
>
> # lsmod
> Module                  Size  Used by
> i2c_dev                12552  0
> hwmon                   5128  0
> eeprom                  9488  0
> i2c_isa                 7552  0
> i2c_nforce2             9344  0
>
> I have the output from the sensors-detect if anyone wants to see it.

Can you post a URL to a datasheet or high-resolution image of your motherboard?

I suspect it is not actually a w83627ehf, 627ehg, or 627dhg chip in
your board. The datasheet or image will help us figure that out.

David

[lm-sensors] MSI K9N4 SLI-F sensors-detect says use w83627ehf

[Curt Blank]

Ug. Yes I can do that and I will. I was just thinking in order to take a 
decent picture I'm going to have to disconnect everything and this server 
has a lot. I'll do that tonight. And thanks for the quick reply.

Just for grins yesterday I tried loading all the Winbond drivers and I got 
a number to load successfully but there were no sensors. Unless having 
multiple loaded causes a conflict. i2c_dev might not have been loaded 
though, if that makes a difference.

-Curt

On Tue, 20 Mar 2007, David Hubbard wrote:

> Hi Curt,
> 
> On 3/20/07, Curt Blank <curt at curtronics.com> wrote:
> > Motherboard: MSI K9N4 SLI-F
> > Chipset: nVidia nForce 500 SLI
> > SuSE Linux 10.1
> > Kernel: 2.6.16.27-0.9
> > lm-sensors: 2.10.0-10 (from SuSE 10.1 distro)
> >
> > I run sensors-detect that came with the SuSE distribution and it can't
> > find any sensors. So then I download and run the latest sensors-detect and
> > it tells me to use the w83627ehf driver. But when I try it fails:
> >
> > # modprobe w83627ehf
> > FATAL: Error inserting w83627ehf
> > (/lib/modules/2.6.16.27-0.9-smp-bnfsrvr/kernel/drivers/hwmon/w83627ehf.ko):
> > No such device
> >
> > It has a confidence of 9 but it fails. Am I missing something here?
> > w83627ehf has been in the kernel since 2.6.13, but do I maybe need a newer
> > version then I have?
> >
> > These are the modules currently installed:
> >
> > # lsmod
> > Module                  Size  Used by
> > i2c_dev                12552  0
> > hwmon                   5128  0
> > eeprom                  9488  0
> > i2c_isa                 7552  0
> > i2c_nforce2             9344  0
> >
> > I have the output from the sensors-detect if anyone wants to see it.
> 
> Can you post a URL to a datasheet or high-resolution image of your
> motherboard?
> 
> I suspect it is not actually a w83627ehf, 627ehg, or 627dhg chip in
> your board. The datasheet or image will help us figure that out.
> 
> David
> 
> 

[lm-sensors] MSI K9N4 SLI-F sensors-detect says use w83627ehf

[David Hubbard]

Hi Curt,

On 3/20/07, Curt Blank <curt at curtronics.com> wrote:
> Ug. Yes I can do that and I will. I was just thinking in order to take a
> decent picture I'm going to have to disconnect everything and this server
> has a lot. I'll do that tonight. And thanks for the quick reply.
>
> Just for grins yesterday I tried loading all the Winbond drivers and I got
> a number to load successfully but there were no sensors. Unless having
> multiple loaded causes a conflict. i2c_dev might not have been loaded
> though, if that makes a difference.
>
> -Curt

You may be able to find a picture of the motherboard online...

You also don't have to take a picture of the motherboard if you can
identify which is the hardware monitoring chip...

For example: you should be able to find the CPU. It will be close to
the Northbridge. There might be a southbridge / ATA controller on the
board. There will be very few other large chips. So just google all
the letters and numbers on each chip and you should be able to tell
what each does.

Post results here. :-)

David

[lm-sensors] MSI K9N4 SLI-F sensors-detect says use w83627ehf

[Jan van Tiggelen]

Curt, David, Jean, hi, (again)

The schematics are in your personal boxes. I will not post them to the list.

> 
> On Wed, 21 Mar 2007, David Hubbard wrote:
> 
> And, when you get the schematic for this MB, I would greatly appreciate it
> if you could tell me what input is what voltage. 
> I'd also lik to see if the Northbridge fan is connected to the w83627dhg. 
> 
> -Curt
> > 
> > If you were getting unknown device ID, we should probably add another
> > device ID to the driver (0xa020 ?). I'll take a look at it when I get 
> > home tonight.
> > 
> > David
> > 
> 
> Hi David, 
> 
> No I think the Device ID issue is due to the mask needing to be 0xFFE0 as 
> Jan pointed out. I'm going to change that and try it.
> 
> -Curt
> 

The chip is indeed a W83627DHG, instead of the W83627EHF. It's a good catch
that you got it working. But, after reviewing the schematics, of which the
W83267EHF is public via:
http://www.winbond-usa.com/products/winbond_products/pdfs/PCIC/W83627EHF_EHG
c.pdf

1 - The W83627DHG features are at some parts very different from the
W83627EHF: It has Intel PECI/SST interface support (why would they use that
on an AMD mainboard?), Dual-current mode temperature sensing, 4-pin FAN
support, alerting outputs and also a 8 VID I/O. 

2 - Derriving from (1) the pin-out purposes are very different in this case.
E.g. pin 119 (CPUFAN) which was used as a CPU-fan connection, the 'new pin'
used for that function is pin 121 and rated at 'Default High' (probably only
at boot time)

3 - From the (internal) BIOS notes I can tell you that the instruction set
in the bios for W627EHF and W627DHG is almost similar, except for the 4-pin
FAN feature. The rest of the instruction set is based on SMARTFAN III code.

Lastly, on a side note, I have to say that I'm not 100% sure what impact
this has on the suggested 'changes' between 627EHF and 627DHG driver. I'll
leave that up to someone else to determine. (David / Jean?)

Kind regards,

Jan van Tiggelen

[lm-sensors] MSI K9N4 SLI-F sensors-detect says use w83627ehf

[Jean Delvare]

Hi Jan,

On Wed, 21 Mar 2007 20:41:36 +0100, Jan van Tiggelen wrote:
> The chip is indeed a W83627DHG, instead of the W83627EHF. It's a good catch
> that you got it working. But, after reviewing the schematics, of which the
> W83267EHF is public via:
> http://www.winbond-usa.com/products/winbond_products/pdfs/PCIC/W83627EHF_EHG
> c.pdf
> 
> 1 - The W83627DHG features are at some parts very different from the
> W83627EHF: It has Intel PECI/SST interface support (why would they use that
> on an AMD mainboard?), Dual-current mode temperature sensing, 4-pin FAN
> support, alerting outputs and also a 8 VID I/O. 

You shouldn't expect the consumer board manufacturers to pick the best
chip for their board, nor to wire it in an optimal way. In general they
care more about low prices (which they achieve with high volumes) than
technical correctness.

> 2 - Derriving from (1) the pin-out purposes are very different in this case.
> E.g. pin 119 (CPUFAN) which was used as a CPU-fan connection, the 'new pin'
> used for that function is pin 121 and rated at 'Default High' (probably only
> at boot time)
> 
> 3 - From the (internal) BIOS notes I can tell you that the instruction set
> in the bios for W627EHF and W627DHG is almost similar, except for the 4-pin
> FAN feature. The rest of the instruction set is based on SMARTFAN III code.
> 
> Lastly, on a side note, I have to say that I'm not 100% sure what impact
> this has on the suggested 'changes' between 627EHF and 627DHG driver. I'll
> leave that up to someone else to determine. (David / Jean?)

On our side, we don't care that much about pins. We care about
registers. If the pinout is different but the registers are the same,
the driver doesn't need any special care. The w83627ehf driver already
differentiates between the W83627EHF and W83627DHG where it needs to. If
you see anything that is not handled properly in that respect on a
real-world motherboard, that's a bug, report it to David.

-- 
Jean Delvare

[lm-sensors] MSI K9N4 SLI-F sensors-detect says use w83627ehf

[Jan van Tiggelen]

Curt, hi,-

Assuming that you've figured out a correct setting for your FAN divisors,
last thing left to do is setting right labels and Vin calculation for the
other channels.

So, based on your feedback of voltage levels, and with the schematics of the
MSI K9N4 SLI-F and the pre-set definitions of the W83627EHF calculations I
could suggest for sensors.conf:

chip "w83627dhg-*"

labels:
label in0 "VCore"
label in1 "+12V"
label in2 "5V"
label in3 "-12V"
label in5 "AVCC"
label in7 "VSB"
label in8 "VBAT"

computations:
compute in1 @*(1+(56/10)), @/(1+(56/10))
compute in2 @*(1+(22/10)), @/(1+(22/10))
compute in3 -@*(1+(56/10)), -@/(1+(56/10))

Let me know if this gives you the correct readings. If it's not, I'll try
and get the values for Rin and Rf.

And for the FAN labels:
Label fan3 "Northbridge Fan"


Honestly speaking: for the voltage computations I lost track a bit, since
627EH is derived from 781D in turn...
Maybe you could post a better update, with the mainboard on hand?

Kind regards,

Jan van Tiggelen

[lm-sensors] MSI K9N4 SLI-F sensors-detect says use w83627ehf

[Curt Blank]

Hi Jan,

Thanks.

Jan van Tiggelen wrote:
> Curt, hi,-
> 
> Assuming that you've figured out a correct setting for your FAN divisors,
> last thing left to do is setting right labels and Vin calculation for the
> other channels.

The divisors were ok out of the box, I was just trying to change 4 & 5 
to see if there where fans connected before I had the schematic and it 
showed there were no fans connected.

> 
> So, based on your feedback of voltage levels, and with the schematics of the
> MSI K9N4 SLI-F and the pre-set definitions of the W83627EHF calculations I
> could suggest for sensors.conf:
> 
> chip "w83627dhg-*"
> 
> labels:
> label in0 "VCore"
> label in1 "+12V"
> label in2 "5V"
> label in3 "-12V"
> label in5 "AVCC"
> label in7 "VSB"
> label in8 "VBAT"
> 
> computations:
> compute in1 @*(1+(56/10)), @/(1+(56/10))
> compute in2 @*(1+(22/10)), @/(1+(22/10))
> compute in3 -@*(1+(56/10)), -@/(1+(56/10))
> 
> Let me know if this gives you the correct readings. If it's not, I'll try
> and get the values for Rin and Rf.

I'll give this a try.

> 
> And for the FAN labels:
> Label fan3 "Northbridge Fan"

I did already. :)

> 
> 
> Honestly speaking: for the voltage computations I lost track a bit, since
> 627EH is derived from 781D in turn...
> Maybe you could post a better update, with the mainboard on hand?

Ok, but what do you mean by a better update? And MB on hand?

Here's the results of plugging in your equations:

w83627dhg-isa-0a10
Adapter: ISA adapter
VCore:     +1.23 V  (min =  +0.00 V, max =  +1.74 V)
+12V:     +12.30 V  (min =  +6.97 V, max =  +5.23 V)
5V:       +10.44 V  (min =  +7.48 V, max = +10.75 V)
-12V:     -21.54 V  (min =  -5.91 V, max = -18.27 V)
in4:       +1.54 V  (min =  +1.23 V, max =  +0.02 V)
AVCC:      +1.46 V  (min =  +0.24 V, max =  +0.78 V)
in6:       +1.61 V  (min =  +0.17 V, max =  +0.53 V)
VSB:       +3.26 V  (min =  +2.18 V, max =  +2.26 V)
VBAT:      +2.99 V  (min =  +2.08 V, max =  +0.86 V)

-Curt

> 
> Kind regards,
> 
> Jan van Tiggelen
> 
> 
> _______________________________________________
> lm-sensors mailing list
> lm-sensors at lm-sensors.org
> http://lists.lm-sensors.org/mailman/listinfo/lm-sensors

[lm-sensors] MSI K9N4 SLI-F sensors-detect says use w83627ehf

[Curt Blank]

Hi Jan,

den_tichel at zonnet.nl wrote:
> Hi Curt,
> 
> Based on the schematics I've sent you, it seems that there is some 
> miscomputation.
> 
>> So, based on your feedback of voltage levels, and with the schematics of the
>> MSI K9N4 SLI-F and the pre-set definitions of the W83627EHF calculations I
>> could suggest for sensors.conf:
>>
>> chip "w83627dhg-*"
>>
>> labels:
>> label in0 "VCore"
>> label in1 "+12V"
>> label in2 "5V"
>> label in3 "-12V"
>> label in5 "AVCC"
>> label in7 "VSB"
>> label in8 "VBAT"
>>
>> computations:
>> compute in1 @*(1+(56/10)), @/(1+(56/10))
>> compute in2 @*(1+(22/10)), @/(1+(22/10))
>> compute in3 -@*(1+(56/10)), -@/(1+(56/10))
>>
>> Let me know if this gives you the correct readings. If
>> it's not, I'll try and get the values for Rin and Rf.
>>
>> Here's the results of plugging in your equations:
>>
>> w83627dhg-isa-0a10
>> Adapter: ISA adapter
>> VCore:     +1.23 V  (min =  +0.00 V, max =  +1.74 V)
>> +12V:     +12.30 V  (min =  +6.97 V, max =  +5.23 V)
>> 5V:       +10.44 V  (min =  +7.48 V, max = +10.75 V)
>> -12V:     -21.54 V  (min =  -5.91 V, max = -18.27 V)
>> in4:       +1.54 V  (min =  +1.23 V, max =  +0.02 V)
>> AVCC:      +1.46 V  (min =  +0.24 V, max =  +0.78 V)
>> in6:       +1.61 V  (min =  +0.17 V, max =  +0.53 V)
>> VSB:       +3.26 V  (min =  +2.18 V, max =  +2.26 V)
>> VBAT:      +2.99 V  (min =  +2.08 V, max =  +0.86 V)
>>
>> -Curt
>>
> 
> So right now I could suggest a different computation for:
> 
> compute in3 -@*(1+(232/10)), -@/(1+(232/10))
> 
> The calculation for the 5V is OK, but then again... in2 is not 5V. So 
> now I'm guessing in6 is 5V, so change your in2 labels and computation 
> to in6.
> 
> As a last challenge:
> How are your system temps displayed? After all, they are connected 
> through thermal resistors on SYS_TMP and CPU_TMP.
> 
> Kind regards,
> 
> Jan van Tiggelen

I see what you mean on the schematic, but one question, that schematic 
is for an EHG, this is a DHG. The only reason I mention that is because 
I put the new computation in and it made it worse:

w83627dhg-isa-0a10
Adapter: ISA adapter
VCore:     +1.23 V  (min =  +0.00 V, max =  +1.74 V)
+12V:     +12.30 V  (min =  +6.97 V, max =  +5.23 V)
in2:       +3.26 V  (min =  +2.34 V, max =  +3.36 V)
-12V:     -78.99 V  (min = -21.68 V, max = -66.99 V)
in4:       +1.54 V  (min =  +1.23 V, max =  +0.02 V)
AVCC:      +1.46 V  (min =  +0.24 V, max =  +0.78 V)
5V:        +5.15 V  (min =  +0.54 V, max =  +1.69 V)
VSB:       +3.26 V  (min =  +2.18 V, max =  +2.26 V)
VBAT:      +2.99 V  (min =  +2.08 V, max =  +0.86 V)
Case Fan: 5113 RPM  (min = 1205 RPM, div = 8)
CPU Fan:  2909 RPM  (min = 1004 RPM, div = 8)
NB Fan :  6887 RPM  (min = 2008 RPM, div = 4)
Sys Temp:    +41?C  (high =   +55?C, hyst =   +40?C)
CPU Temp:  +33.5?C  (high = +75.0?C, hyst = +40.0?C)

And the Sys Temp seems to be fine, but the CPU Temp is reading low when 
sensors says 34?C the BIOS says 38?C.

And this is what k8temp is showing at the same time as above:

k8temp-pci-00c3
Adapter: PCI adapter
Core1 Temp:
              +43?C
Core2 Temp:
              +46?C

I guessed that the temp1 and temp3 were the two cores, if not, what 
would they be? temp2 and temp4 show nothing.

-Curt