| 39 | | Fan readings can be divided by a programmable divider (1, 2, 4 or 8) to give |
| 40 | | the readings more range or accuracy. Not all RPM values can accurately be |
| 41 | | represented, so some rounding is done. With a divider of 2, the lowest |
| 42 | | representable value is around 950 RPM. See doc/fan-divisors for details. |
| | 50 | A different alarm is triggered if the fan speed is too low to be measured. |
| | 51 | It is suggested to increase the fan divider (which is really a clock |
| | 52 | divider) in this case. |
| 47 | | PWM values are from 0 to 255. |
| | 57 | slowest accuracy highest |
| | 58 | measurable around 3000 accurate |
| | 59 | divider speed (RPM) RPM (RPM) speed (RPM) |
| | 60 | 1 1882 18 6928 |
| | 61 | 2 941 37 4898 |
| | 62 | 4 470 74 3464 |
| | 63 | 8 235 150 2449 |
| | 64 | |
| | 65 | For the curious, here is how the values above were computed: |
| | 66 | * slowest measurable speed: clock/(255*divider) |
| | 67 | * accuracy around 3000 RPM: 3000^2/clock |
| | 68 | * highest accurate speed: sqrt(clock*100) |
| | 69 | The clock speed for the PC87360 family is 480 kHz. I arbitrarily chose 100 |
| | 70 | RPM as the lowest acceptable accuracy. |
| | 71 | |
| | 72 | Note that not all RPM values can be represented. This is not only true for |
| | 73 | the measured speeds, but also for the programmable low limits, so don't be |
| | 74 | surprised if you try to set, say, fan1_min to 2900 and it finaly reads |
| | 75 | 2909. Likewise, if you try to set fan1_min to 0, it'll be set to the |
| | 76 | lowest representable RPM value (depending on the divider). |
| | 77 | |
| | 78 | |
| | 79 | Fan Control |
| | 80 | ----------- |
| | 81 | |
| | 82 | PWM (pulse width modulation) values range from 0 to 255, with 0 meaning |
| | 83 | that the fan is stopped, and 255 meaning that the fan goes at full speed. |
| | 84 | |
| | 85 | Be extremely careful when changing PWM values. Low PWM values, even |
| | 86 | non-zero, can stop the fan, which may cause irreversible damage to your |
| | 87 | hardware if temperature increases too much. When changing PWM values, go |
| | 88 | step by step and keep an eye on temperatures. |
| | 89 | |
| | 90 | One user reported problems with PWM. Changing PWM values would break fan |
| | 91 | speed readings. No explanation nor fix could be found. |
| | 92 | |
| | 93 | |
| | 94 | Temperature Monitoring |
| | 95 | ---------------------- |
| | 104 | The PC87366 has three additional temperature channels, based on |
| | 105 | thermistors (as opposed to thermal diodes for the first temperature |
| | 106 | channels). For technical reasons, these channels are held by the VLM |
| | 107 | (voltage level monitor) logical device, not the TMS (temperature |
| | 108 | measurement) one. The formula to convert the measured voltage into a |
| | 109 | temperature isn't known, so the driver doesn't handle these temperature |
| | 110 | channels at all yet. |
| | 111 | |
| | 112 | |
| | 113 | Voltage Monitoring |
| | 114 | ------------------ |
| | 115 | |
| 57 | | external. Some of them are divided by two internally, you will have to |
| 58 | | compensate in sensors.conf. Each voltage measured has associated low and |
| 59 | | high limit, each of which triggers an alarm when crossed. |
| | 117 | external. Some of them (in7:Vsb, in8:Vdd and in10:AVdd) are divided by two |
| | 118 | internally, you will have to compensate in sensors.conf. Others (in0 to in6) |
| | 119 | are likely to be divided externally. The meaning of each of these inputs as |
| | 120 | well as the values of the resistors used for division is left to the |
| | 121 | motherboard manufacturers, so you will have to document yourself and edit |
| | 122 | sensors.conf accordingly. National Semiconductor has a document with |
| | 123 | recommended resistor values for some voltages, but this still leaves much |
| | 124 | room for per motherboard specificities, unfortunately. |
| | 125 | |
| | 126 | Each voltage measured has associated low and high limit, each of which |
| | 127 | triggers an alarm when crossed. |
| | 128 | |
| | 129 | |
| | 130 | General Remarks |
| | 131 | --------------- |
| 62 | | is read at least once. This means that the cause for the alarm may |
| 63 | | already have disappeared! Note that in the current implementation, all |
| 64 | | hardware registers are read whenever any data is read (unless it is less |
| 65 | | than 1.5 seconds since the last update). This means that you can easily |
| 66 | | miss once-only alarms. |
| | 134 | is read at least once. This means that the cause for the alarm may already |
| | 135 | have disappeared! Note that all hardware registers are read whenever any |
| | 136 | data is read (unless it is less than 2 seconds since the last update, in |
| | 137 | which case cached values are returned instead). As a consequence, when |
| | 138 | a once-only alarm triggers, it may take 2 seconds for it to show, and 2 |
| | 139 | more seconds for it to disappear. |
| | 140 | |
| | 141 | Monitoring of in9 isn't enabled at lower init levels (<3) because that |
| | 142 | channel measures the battery voltage (Vbat). It is a known fact that |
| | 143 | repeatedly sampling the battery voltage reduces its lifetime. National |
| | 144 | Semiconductor smartly designed their chipset so that in9 is sampled only |
| | 145 | once every 1024 sampling cycles (that is every 34 minutes at the default |
| | 146 | sampling rate), so the effect is attenuated, but still present. |
