Jean, etal [ apologies in advance for any mail bogosity - this is sent with a minimal sendmail script to avoid mime attachements. Ive tested it on myself, but u never know.. pls advise if bogosity is subtle - if obvious I hope to know ] This patch corrects a voltage-reference value swap in current code. It now agrees with your previous explanation; the internal voltage ref is higher than the value specd/recommended for an external reference. This is a minimal change, and qualifies strictly as a bug-fix. After .13 is out, Id want to replace/add a module-param to specify the external ref voltage that is actually supplied, though on my board it is the recommended value, so I dont need the param. heres a before and after: voltages drop approx 2%, which is in agreement with calcs: 1-(2966/3025) root at soekris:~# sensors pc87366-isa-6620 Adapter: ISA adapter avi0: +3.01 V (min = +0.00 V, max = +3.01 V) VCORE: +2.00 V (min = +0.00 V, max = +3.01 V) VCC: +4.96 V (min = +0.00 V, max = +6.03 V) VPWR: +13.55 V (min = +5.93 V, max = +28.02 V) +12V: +11.74 V (min = +0.00 V, max = +14.46 V) -12V: -12.97 V (min = -60.61 V, max = -2.76 V) GND: +0.00 V (min = +0.00 V, max = +3.01 V) Vsb: +3.28 V (min = +3.00 V, max = +3.59 V) Vdd: +2.91 V (min = +3.00 V, max = +3.59 V) ALARM Vbat: +3.01 V (min = +2.40 V, max = +3.01 V) AVdd: +3.28 V (min = +3.00 V, max = +3.59 V) Temp: +107 C (low = +0 C, high = +70 C) ALARM Critical: +85 C ALARM root at soekris:~# root at soekris:~# rmmod pc87360; modprobe pc87360; sensors -s ; sensors pc87360: Device 0x09 not activated pc87360 9191-6620: VLM conversion set to 1s period, 160us delay pc87366-isa-6620 Adapter: ISA adapter avi0: +2.95 V (min = +0.00 V, max = +2.95 V) VCORE: +1.95 V (min = +0.00 V, max = +2.95 V) VCC: +4.87 V (min = +0.00 V, max = +5.91 V) VPWR: +12.34 V (min = +6.05 V, max = +27.94 V) +12V: +11.51 V (min = +0.00 V, max = +14.18 V) -12V: -13.89 V (min = -60.61 V, max = -3.89 V) GND: +0.00 V (min = +0.00 V, max = +2.95 V) Vsb: +3.22 V (min = +2.99 V, max = +3.59 V) Vdd: +2.85 V (min = +2.99 V, max = +3.59 V) ALARM Vbat: +2.95 V (min = +2.40 V, max = +2.95 V) AVdd: +3.22 V (min = +2.99 V, max = +3.59 V) Temp: +106 C (low = +0 C, high = +70 C) ALARM Critical: +85 C ALARM Henrik Brix Anderson has kindly tested a version of this patch which added several new module-parameters, and is thus unsuitable for 13 at this late stage. If this doesnt get in (for obvious reasons), Ill repost a version with a mod-param to set external vref. Im copy-pasting his words and signoff, and cc'g him. Brix, when you tested, you evidently measured your vrefext. The code default is 1.211mv, not your measured 1.210mv. Did you use the mod-param to tweak it down 1mv ? Or were the defaults also a 'great' improvment. Can you quantify 'great' ? errs 10mv -> 1mv ?? Did you mess about with the vrefscale setting ? thanks Signed-off-by: Jim Cromie <jcromie at divsol.com> I have tested this patch with a measured vrefext of 1210mV - it greatly improves the precision of my sensor readings compared to the measured voltages. Signed-off-by: Henrik Brix Andersen <brix at gentoo.org> diff -ruNp -X exclude-diffs ../linux-2.6.13-rc5-mm1/drivers/hwmon/pc87360.c vref-int-ext/drivers/hwmon/pc87360.c --- ../linux-2.6.13-rc5-mm1/drivers/hwmon/pc87360.c 2005-08-07 13:17:29.000000000 -0600 +++ vref-int-ext/drivers/hwmon/pc87360.c 2005-08-14 10:29:20.000000000 -0600 @@ -798,7 +798,7 @@ static int pc87360_detect(struct i2c_ada i &= pc87360_read_value(data, LD_TEMP, NO_BANK, PC87365_REG_TEMP_CONFIG); } - data->in_vref = (i&0x02) ? 3025 : 2966; + data->in_vref = (i&0x02) ? 2966 : 3025; /* vref (mV) * 2.54 */ dev_dbg(&new_client->dev, "Using %s reference voltage\n", (i&0x02) ? "external" : "internal");
