This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

LM63: Tachometer Mode

Part Number: LM63

Dear Sir,

From LM63 datasheet show "Note: If the PWM Clock is 360 kHz, mode 00 is used regardless of the setting of these two bits." it's mean register "4A" bit 3 set to "0" the Tachometer Mode can't use "10" and "11" mode? What is the working principle of Smart-tach mode? If my fan PWM frequency is 25kHz, how can I do the Lookup table to control the fan speed? 

Best regards

Sam

  • Hi Sam,

    Below is my feedback to your inquires.

    From LM63 datasheet show "Note: If the PWM Clock is 360 kHz, mode 00 is used regardless of the setting of these two bits." it's mean register "4A" bit 3 set to "0" the Tachometer Mode can't use "10" and "11" mode?

    You are correct, when using 360kHz mode the Tachometer mode will function as mode 00.

    What is the working principle of Smart-tach mode?

    Smart-tach measurement accommodates for the corrupted tachometer pulses when using switching transistor drive.

    If my fan PWM frequency is 25kHz, how can I do the Lookup table to control the fan speed?

    Please see the application note on page 29 of the datasheet. With the internal clock set to 360 kHz, setting registers 4D to 7 the PWM frequency should be 25.71 kHz.


    Thanks,

  • Thanks for replying.

    2 more questions for you.

    1) what does it means about "Tach_Count_Multiple" and "Edges Used" on register 0x46 on pp25 of datasheet? how to apply these parameters to RPM calculation?

    2) what does "Step resolution %" means on pp29 of datasheet 

    3)  What’s the definition and usage of the “Edges Used” item of register 0x4B bit [1..0]?

    4)  There are two Remote Temperature OFFSET registers 0x11 (MSB) and 0x12 (LSB). Does it mean that we could set the offset values (added or subtracted) to compensate for the different values measured by other sensors to generate the final Remote

    Diode Temperature reading in registers 0x01 and 0x10?

  • Hi Peter,

    Below is my feedback.

    1) what does it means about "Tach_Count_Multiple" and "Edges Used" on register 0x46 on pp25 of datasheet? how to apply these parameters to RPM calculation?

    Please see "COMPUTING RPM OF THE FAN FROM THE TACH COUNT" on page 30 on how to calculate RPM.


    2) what does "Step resolution %" means on pp29 of datasheet

    The step resolution % is the Duty Cycle resolution percentage that is represented in register 0x4C.
    If you are using 25.71kHz setting your Duty cycle resolution will be 7.14%. Therefore values in register 0x4C will be the following
    0x01: ~7.14%
    0x02: ~14.28%
    0x03: ~21.42%
    ....
    0x07: ~50%
    ...
    0x0E: ~100%

    3) What’s the definition and usage of the “Edges Used” item of register 0x4B bit [1..0]?

    Do you mean in registers 0x46? I do not see this "Edges Used" for these bits. If so please see answer to question 1.


    4) There are two Remote Temperature OFFSET registers 0x11 (MSB) and 0x12 (LSB). Does it mean that we could set the offset values (added or subtracted) to compensate for the different values measured by other sensors to generate the final Remote

    Yes, negative offsets will be in 2's compliments form.

    Diode Temperature reading in registers 0x01 and 0x10?

    Yes

    Thanks,

  • Hi David,

        The "Edges Used" bits are defined in register 0x46 on page 25.

        Since these two bits are read only, does it mean that the "Edges Used" and "Tach_Count_Multiple" fields are automatically set by the LM63 chip?

        Besides, the example of computing RPM on page 30 uses the value of register 0x46 as the LSB in calculating the Total_Tach_Count. Bit 0 and bit 1 aren't masked off in the calculation, and it seems that the "Edges Used" and "Tach_Count_Multiple" bits are used in the RPM calculation.

        Thanks.

             Todd

  • Hi Todd,

    I am still working on getting this information for you.  Please give me a few more days so I can give you a detailed answer.

    Thanks,

  • Hi Todd,

    You are correct; these bits are used to represent the RPM and to indicate how the LM63 calculates the RPM. The "Edges Used" bit indicates how many edges the LM63 detected and used when calculating the RPM. The "Tach Count Multiple" bit indicates the multiplier the LM63 used to calculate the RPM.

    The Tach mode set in 0x4A <1:0> determines when the Tach monitor will detect the edges of the tach input.

    Mode 00: The Tach monitor is on regardless of its corresponding PWM.

    Mode 01: The Tach monitor is on only when its corresponding PWM is high.

    Modes 10 and 11: The Tach monitor is on only when its corresponding PWM is high. The PWM high time will be extended up to 50ms to allow enough time to calculate the fan speed.

     

    For example, if the LM63 only detects 2 edges the LM63 multiplies this time period by a factor of 4. This results in the Bits [1:0] of Register 46hex to be set to 01 [Edges Used = 2, Tach_Count_Multiple = 4].

    If only 3 edges are detected, then the LM63 multiplies this time period by a factor of 2. This results in the Bits [1:0] of Register 46hex to be set to 10 [Edges Used = 3, Tach_Count_Multiple = 2].

    If 5 edges are detected, then the LM63 multiplies this time period by a factor of 1. This results in the Bits [1:0] of Register 46hex to be set to 11 [Edges Used = 5, Tach_Count_Multiple = 1].

     

    Please see the below diagrams for further explanation

  • Hi David,

        Thanks for your information. However, the description of the register 0x4A <1:0> on page 21 is somewhat different from what you mentioned. The datasheet doesn't mention that modes 01, 10, and 11 are used only when its corresponding PWM value is high. Besides, what does the "Smart-tach mode" mean for mode 10 or 11? What does "Use with direct PWM drive of fan power" mean for mode 10 or 11?

        Regarding to the "Edges Used" and "Tach_Count_Multiple" information, it seems your diagrams are targeting for capturing pulse edges in the time interval of two pulses. Is this rule applied for fans with 2 pulses/rev? Or it is applied to all fans? Since we set PWM_Clock_Select to 360 kHz, the Tach_Count_Multiple is fixed to 1 according to the datasheet. Does it mean bit 0 of register 0x46 will always be fixed to 1? Or we need to forcefully set this bit to 1 RPM calculation? By the way, how to interpret the "Edges Used" bit when setting PWM_Clock_Select to 1?

        Thanks.

           Todd

  • Hi Todd,

    Besides, what does the "Smart-tach mode" mean for mode 10 or 11?

    Please see above reply on Smart-tach mode.

    "Smart-tach measurement accommodates for the corrupted tachometer pulses when using switching transistor drive"

    What does "Use with direct PWM drive of fan power" mean for mode 10 or 11?

    Direct PWM drive of fan power is the use case where the PWM directly involves turning the fan's power supply on and off typically by controlling a MOSFET. Please see Figure 3 for an example.

    Is this rule applied for fans with 2 pulses/rev? Or it is applied to all fans?

    This applies to all fans. Please see "COMPUTING RPM OF THE FAN FROM THE TACH COUNT" section of the datasheet to determine RPM if not using a 2-pulse/rev fan.

    Since we set PWM_Clock_Select to 360 kHz, the Tach_Count_Multiple is fixed to 1 according to the datasheet. Does it mean bit 0 of register 0x46 will always be fixed to 1? Or we need to forcefully set this bit to 1 RPM calculation?

    This will be fixed to 1 when PWM_Clock_Select is set to 360 kHz.

    By the way, how to interpret the "Edges Used" bit when setting PWM_Clock_Select to 1?

    No change in interpretation. When PWM_Clock_Select is set to 360 kHz the number edges used will be 5.


    Thanks,

  • Hi David,

        We configure the 8-step Lookup Table with four temperature groups as the following:

    0x50: 0x4B,      0x51: pwm1

    0x52: 0x4B,      0x53: pwm1

    0x54: 0x55,      0x55: pwm2

    0x56: 0x55       0x57: pwm2

    0x58: 0x5A      0x59: pwm3

    0x5A: 0x5A      0x5B: pwm3

    0x5C: 0x5F      0x5D: pwm4

    0x5E: 0x5F       0x5F: pwm4

        By running the thermal simulation, the fan speed was changed when reaching the configured temperature. However, we found that the fan speed would not slow down when the temperature was decreased. For example, the fan speed was set by pwm3 when the temperature detected by the remote diode was over 90 degree. However, the fan speed was kept at the one set by pwm3 when the temperature was decreased to 75 degree.

        What should be the correct behavior of the LM63 when the monitored temperature was decreased?

        Thanks.

                         Todd

  • Hi Todd,

    This is not the expected behavior of the device.  Device should change PWM with change of temperature.  What are your PWM settings (0x4A-0x5F)?  Also what is your Tcrit and Remote alerts values set to?

    Thanks,

  • Hi David,

        Here is the register dump of 0x4A-0x5F.

           0   1   2   3   4   5   6   7   8   9   A   B   C   D   E   F

    40:   .   .    .    .    .    .    .    .    .     .   00 0F 00 07 00 04
    50: 50 02 50 02 55 03 55 03 5A 04 5A 04 5F 05 5F 05

        Besides, the Remote Diode T_CRIT Limit (0x19) is set to 0x7D. Since the ALERT/Tach Select (bit 2) of register 0x03 is set to 1 to select the Tachometer Input mode, we don't change the default value of Remote Alerts.

        Thanks.

            Todd

  • Hi Todd,

    With those exact settings I see the device change PWM output with the temperature.  Please be sure that the device is reading the same temperature that you are setting the remote to.  What remote device are you using?

    Thanks,

  • Hi David,

        The remote device being monitored is a network SOC supporting the WiFi function. We had generated the WiFi traffic to increase the CPU utilization of the SOC, and we did see the fan speed increased as the temperature reading back from register 0x01 (MSB) and 0x10(LSB) increased. However, the PWM setting didn't switch the lower one when the SOC was cooled down by the fan.

        What's the design logic of the LM63 to switch back to previous steps when the monitored temperature is decreased? Any idea on how to debug this issue?

        Thanks.

                  Todd

  • Hi David,

        It's our bad for not doing the test completely. After retesting today, we found that the fan speed did change when the temperature was decreased. When the temperature was decreased by three to five degrees from the previous stage, the fan would be controlled by lower PWM value.

        Thanks.

               Todd

  • Hi Todd,

    Glad to hear you have resolved the issue.  Yes, the PWM value will only change after dropping below the "Lookup Table Temperature Entry" minus the "Lookup Table Hysteresis" set in register 0x4F.  

    Thanks,