Clickless SW slew volume control Specification details

We are analyzing the below module for FadeIn & FadeOut implementation

           Volume Controls -> Adjustable Gain -> Clickless SW Slew -> RC Optimize

   

1) Ramp up & Ramp down duration is different. Ramp up duration seems to be less when comparing with Ramp down. Trust both should be identical

2) Cannot map Relationship among  'SW slew Rate ' & 'Slew Time'. While setting Slew rate as 8 , slew duration of '18 msec' is observed. Please confirm whether this is expected.

Please share the specification document for the module., to verify the above points.

Regards,

Kalaiselvi

  • 0
    •  Super User 
    on Mar 4, 2021 2:29 AM 1 month ago

         Hello Kalaiselvi,

         The test circuit shown below allowed me to measure the RC Time Constant for various slew settings:

    I used a change between -5 and 0 dB because this is the maximum change available in one flick of the GUI knob.  This corresponds to  0.56  --  1.0 on the graph.  Going up, one RC time constant is when the level rises from 0.56 to [ 0.56 + 0.44 * (1 - e^-1) ] = 0.84.  This occurs in a time of about 0.32 seconds.  Thus the RC time constant for a slew setting of 14 is 320 mS.   A similar test on the way down also yielded a TC of 320 mS.

         At a slew setting of 7, the rise and fall are too fast for the Real-Time Display so I sent the signal via DAC to a scope.  The scope trace rises and falls by 0.54 V at each transition.  63% of this change happens at one time constant.

    At a slew setting of 7, the rising time constant is 2.5 mS.

    And the measured downward TC is 2.48 mS.

         By increasing the slew # from 7 to 14 (a difference of 7), the time constant increases from 2.5 to 320 mS, a factor of 128 or 2^7.  This suggests the following empirical formula for the time constant:

         TC = 10 * [ 2 ^ (n - 9) ]    mS      where n is the slew number and TC is the RC time constant.

    This is different from the External Slew RC, where it takes a difference of 4 to double the time constant.

         Example:  A slew setting of n = 8 provides a TC of 5 mS.    The 10-90% rise time = TC * 2.2 = 11 mS. 

    Your measurement of 18 mS is about 3.6 time constants, a reasonable time for a "full" slew.

         Best regards,

         Bob

  • Hello Bob,

         Thanks for sharing the analysis details. Please find the table below for measurement details in our side.

    SW Slew Rate

    Expected Time Constant (ms)

    For 10-90% rise time = (TC * 2.2)

       Fade out processing time Measured in Real Time(ms)

    7

    2.5

    5.5

    12.5

    8

    5

    11

    25

    9

    10

    22

    50

    10

    20

    44

    100

    11

    40

    88

    200

    12

    80

    176

    400

    We have few queries below. Please support.

    >> With the above measurements, observed value in real time is 5 times when compared with Expected time constant.

    >> What is meant by full slew?  How to confirm the expected time constants for 'full Slew'. Any specification document available for this?

    >>  We expect both  Fade In &  Fade out duration are identical with this module. Please confirm. In our measurement Fade In duration is 21ms & Fade out duration is 25 ms for slew rate '8'

    >> If the requirement for time constant duration is 500ms? How can we handle this slew rate. Please support.

    Regards,

    Kaaiselvi

  • 0
    •  Analog Employees 
    on Mar 8, 2021 3:05 PM 1 month ago in reply to Ks929

    I usually recommend using hardware slew when possible. Take a look at the PDF in this thread for some details on the slew rates.

    Ken

  • 0
    •  Super User 
    on Mar 9, 2021 1:42 AM 1 month ago in reply to Ks929

         Hello Kaaiselvi,

         The RC slews are named after the RC lowpass filter -- a series resistor and a shunt capacitor.  Thus they follow the familiar RC response described by these equations:

         Slewing up:   v = Vmax ( 1 - e^(-t / T))  where v is the output,  Vmax is the fixed maximum output,  t is elapsed time and T is the time constant.  In the analog RC circuit, T = RC  (ohms x farads).

         Slewing down:  v = Vmax (e^(-t / T))

         Looking at the scope waveforms in my previous post, notice that as time advances, the slewing becomes slower as it approaches its goal.  Thus, where to consider it finished is practically a matter of personal taste.  Mathematically, the slew never reaches its final value.  We can calculate how far it gets by using the e^x function on a calculator:

    elapsed time      % toward goal

         T x 1                    63 %

         T x 2                    86 %

         T x 3                    95 %

         T x 4                    98 %

         T x 5                    99 %

         Evidently your real-time measurements allowed for a nearly complete slew of 99 %.  If this is your desired definition of a complete slew, then go ahead and multiply the time constant by five.  Others might choose a mere  95% completion which occurs after only three time constants.  Notice I had placed "Full" in quotes.

         Fade In and Fade Out are identical in that they follow the RC equations equally.  Both need five time constants to reach 99 % completion.

         We can solve the equation  T = 10 * [ 2 ^ (n - 9) ]    backwards to determine what slew number will provide the desired time constant.  Given the time constant T in mS, the slew number is:

         n = 9 + [ ln (T/10) ]  / ln2   For example, T = 500 mS, then

         n = 9 + [ ln (500/10)] / ln2  =  9 + [ ln (50)] / 0.693  =  9 + [ 3.912 ] / 0.693  =  9 + 5.65  =  14.65

        Since the slew block only accepts whole numbers, use 15 which provides an actual T of 640 mS.  The real-time slew (for 99% completion) would be five times higher, or 3.2 S.

         You can get four times better resolution to choose slew times by using the HW Slew block.  This thread provides a table of slew times.  In the same thread you'll find a link to a Python script which  has contributed for calculating the time constants.

         Best regards,

         Bob

  • 0
    •  Super User 
    on Mar 9, 2021 1:44 AM 1 month ago in reply to Ken.M

         Hello Ken,

         I have to agree with you there.  With the Sigma 300 / 350 chips having a hardware slew engine, we might as well take advantage of it.

         Best regards,

         Bob