URI:
   DIR Return Create A Forum - Home
       ---------------------------------------------------------
       Micro-Cap EDA Users
  HTML https://mc12.createaforum.com
       ---------------------------------------------------------
       *****************************************************
   DIR Return to: Micro-Cap 12 Discussion
       *****************************************************
       #Post#: 102--------------------------------------------------
       Re: 10 MHz Crystal Model
   DIR By: MC12
       Date: March 20, 2025, 10:32 am
       ---------------------------------------------------------
       Can't you extract those values from the SPICE model for the
       crystal?
       #Post#: 103--------------------------------------------------
       Re: 10 MHz Crystal Model
   DIR By: Robert
       Date: March 20, 2025, 12:07 pm
       ---------------------------------------------------------
       I wouldn't know how to do that.
       What I found is in the attachment. - Incomplete data.
       I will try to determin the equvivalent component values by doing
       AC analysis at different suitable frequency ranges.
       Will let you know if I find out something useful.
       #Post#: 104--------------------------------------------------
       Re: 10 MHz Crystal Model
   DIR By: Robert
       Date: March 20, 2025, 3:39 pm
       ---------------------------------------------------------
       Sweeping the crystal at far below the resonant frequencies and
       making use of the macro data I finally arrived at the equivalent
       circuit parameters. Due to inaccuracies in the calculus, the
       equivalent circuit resonances are about 3 kHz lower compared to
       X1.
       Please have a look at the two screenshots. - What do you think?
       #Post#: 105--------------------------------------------------
       Re: 10 MHz Crystal Model
   DIR By: Robert
       Date: March 20, 2025, 3:41 pm
       ---------------------------------------------------------
       Sorry, missed oiut on the schematic.
       #Post#: 106--------------------------------------------------
       Re: 10 MHz Crystal Model
   DIR By: MC12
       Date: March 21, 2025, 2:53 am
       ---------------------------------------------------------
       I would say that's damn good!  You've done well there.
       The next thing is to use your equivalent circuit in place of X1
       and X2 in your original experiment, and see whether the
       simulation run with stepped resistor value produces the same
       weird results.
       #Post#: 107--------------------------------------------------
       Re: 10 MHz Crystal Model
   DIR By: Robert
       Date: March 21, 2025, 4:15 pm
       ---------------------------------------------------------
       The bigger R3 the the less pronounced becfomes the series
       resonant peaking.
       Obvious, since the constant voltage source is masked with a
       constant current source (sort of).
       #Post#: 108--------------------------------------------------
       Re: 10 MHz Crystal Model
   DIR By: MC12
       Date: March 22, 2025, 4:08 am
       ---------------------------------------------------------
       Things are beginning to make a glimmer of sense.
       Assuming the annotations on your simulation dB/f mean values of
       R4, you've gone 50 to 500 step 150.
       In the dB/f from Post 5, the simulation parameters quoted on the
       schematic are "50 to 3.55k step 1k" which implies 50, 1.05k,
       2.05k, 3.05k.
       If we ignore the 50R step, the 1k, 2k, and 3k runs would be
       consistent with the green plots.  The 50R plot should be hidden
       behind the red line.
       What I don't understand is why there are four green lines in the
       "a few k" region of the graph (Post 5).
       It seems to me the simulation parameters must be not what I
       would have expected, and those four lines are runs with R4 =
       1.05k, 2.05k, 3.05k, and 4.05k.  I'm willing to bet that if you
       plugged those values into your new simulation you would get
       plots which mimic Post 5 (albeit shifted in frequency).  If
       correct, the question is why your step formula produced those
       simulation run values.
       You could verify by using the same step formula in a simple
       potential divider network.  It might also be worth running
       simulations on the original circuit with specific values of R4
       to see whether setting R4 to 1k, 2k, 3k, 4k produces the same
       lines.
       #Post#: 109--------------------------------------------------
       Re: 10 MHz Crystal Model
   DIR By: Robert
       Date: March 22, 2025, 10:14 am
       ---------------------------------------------------------
       I chose the values 50, 200, 350 and 500 for the sake of
       demonstrating that the higher the termination resistor, the
       lesser the series resonant peaking. For much higher values (in
       the oder of kOhms) the peaking is masked by the high resistance
       in series with the crystal.
       The response plots for X1 and the equivalent circuit
       respectively are qualitatively identical with the ca. 3 kHz
       shift on the frequency axis.
       I think we can't blame MC12 for any issues regarding the crystal
       simulation. - Do you agree? Thanks for dealing with my posts.
       #Post#: 110--------------------------------------------------
       Re: 10 MHz Crystal Model
   DIR By: MC12
       Date: March 22, 2025, 2:16 pm
       ---------------------------------------------------------
       --- Quote from: Robert link ---
       >
       > I chose the values 50, 200, 350 and 500 for the sake of
       demonstrating that the higher the termination resistor, the
       lesser the series resonant peaking. For much higher values (in
       the oder of kOhms) the peaking is masked by the high resistance
       in series with the crystal.
       > The response plots for X1 and the equivalent circuit
       respectively are qualitatively identical with the ca. 3 kHz
       shift on the frequency axis.
       > I think we can't blame MC12 for any issues regarding the
       crystal simulation. - Do you agree?
       >
       --- End Quote ---
       Unless I misunderstood your original post, the "much higher
       values" used in your original simulation are to blame for the
       stated problem (apparent lack of resonances) – am I correct?
       I'm not sure what other "issues" remain, except...
       I still don't understand why "50 to 3.55k step 1k" produces four
       traces at values >>50R.
       --- Quote from: Robert link ---
       >
       > Thanks for dealing with my posts.
       >
       --- End Quote ---
       We're all learning.  All I did was help you find an answer for
       yourself.
       #Post#: 111--------------------------------------------------
       Re: 10 MHz Crystal Model
   DIR By: Robert
       Date: March 23, 2025, 9:34 am
       ---------------------------------------------------------
       I don't understand: "I still don't understand why "50 to 3.55k
       step 1k" produces four traces at values >>50R."
       Stepping: Start = 50; End = 3.05k; Step = 1k, i.e. 50; 1.05k;
       2.05k; 3.05k. - My mistake if I said 3.55k.
       Correction: The calculus for C2 in the schematic has a printing
       error.
       It should be: C2 = 1/(2*Pi*1e7*2.27e6) = 7.01 F = 7 fF
       *****************************************************
       Page 2 of 3
   DIR Previous Page
   DIR Next Page