Home > Community > Forums > RF Design > PSS analysis - oscillator

Email

* Required Fields

Recipients email * (separate multiple addresses with commas)

Your name *

Your email *

Message *

Contact Us

* Required Fields
First Name *

Last Name *

Email *

Company / Institution *

Comments: *

 PSS analysis - oscillator 

Last post Mon, Mar 19 2012 2:34 PM by Tawna. 3 replies.
Started by JuliaP 14 Mar 2012 03:44 AM. Topic has 3 replies and 4812 views
Page 1 of 1 (4 items)
Sort Posts:
  • Wed, Mar 14 2012 3:44 AM

    • JuliaP
    • Not Ranked
    • Joined on Wed, Nov 30 2011
    • Posts 2
    • Points 40
    PSS analysis - oscillator Reply

     

    Dear all,

    I am simulating an oscillator (RLC lumped element connected to a sustaining amplifier) in SpectreRF 5.10.41.121508. The oscillation frequency is about 0.85 MHz. There is some initial current condition (50nA) for the inductance L.

    The transient analysis runs well and shows the oscillations building up (100mVpp range).

    However, when I run a PSS analysis with either traponly or gear2only integration method, I encounter start-up problems or wrong estimation of the frequency (for instance 6MHz instead of 0.85 MHz). I choose for the simulation parameters the HARD-I case suggested in the Oscillator Noise Analysis in SpectreRF.

    I know that the tstab, method and the steadyratio are very important to get the right convergence of PSS but so far it is not working out.

    Below is the output log of the simulations.

    Could you please give me some suggestions for the proper start-up of the oscillations in PSS analysis?

    Kind regards

    Julia

    ***********************************************

    Transient Analysis `tran': time = (0 s -> 5 us)

    ***********************************************

    Notice from spectre during IC analysis, during transient analysis `tran'.

    L0: Initial condition computed for node L0:1 is in error by 50 nA (100 %).

    To reduce error in computed initial conditions, decrease `rforce'.

    However, setting rforce too small may result in convergence

    difficulties or in the matrix becoming singular.

    Important parameter values:

    start = 0 s

    outputstart = 0 s

    stop = 5 us

    step = 5 ns

    maxstep = 100 ns

    ic = all

    skipdc = no

    reltol = 10e-06

    abstol(I) = 1 pA

    abstol(V) = 30 nV

    temp = 27 C

    tnom = 27 C

    tempeffects = all

    errpreset = moderate

    method = traponly

    lteratio = 3.5

    relref = sigglobal

    cmin = 0 F

    gmin = 1 pS

    maxrsd = 0 Ohm

    mos_method = s

    mos_vres = 50 mV

    tran: time = 126.6 ns (2.53 %), step = 13.16 ns (263 m%)

    tran: time = 376.7 ns (7.53 %), step = 10.06 ns (201 m%)

    tran: time = 642.4 ns (12.8 %), step = 28.11 ns (562 m%)

    tran: time = 879.9 ns (17.6 %), step = 9.945 ns (199 m%)

    tran: time = 1.127 us (22.5 %), step = 13.19 ns (264 m%)

    tran: time = 1.377 us (27.5 %), step = 10.69 ns (214 m%)

    tran: time = 1.631 us (32.6 %), step = 10.87 ns (217 m%)

    tran: time = 1.885 us (37.7 %), step = 13.1 ns (262 m%)

    tran: time = 2.126 us (42.5 %), step = 9.989 ns (200 m%)

    tran: time = 2.392 us (47.8 %), step = 24.88 ns (498 m%)

    tran: time = 2.635 us (52.7 %), step = 9.975 ns (199 m%)

    tran: time = 2.878 us (57.6 %), step = 12.63 ns (253 m%)

    tran: time = 3.131 us (62.6 %), step = 10.87 ns (217 m%)

    tran: time = 3.385 us (67.7 %), step = 10.7 ns (214 m%)

    tran: time = 3.637 us (72.7 %), step = 14.54 ns (291 m%)

    tran: time = 3.881 us (77.6 %), step = 9.952 ns (199 m%)

    tran: time = 4.131 us (82.6 %), step = 18.58 ns (372 m%)

    tran: time = 4.38 us (87.6 %), step = 10.07 ns (201 m%)

    tran: time = 4.63 us (92.6 %), step = 12.18 ns (244 m%)

    tran: time = 4.884 us (97.7 %), step = 11.09 ns (222 m%)

    Number of accepted tran steps = 421.

    Initial condition solution time = 10 ms.

    Intrinsic tran analysis time = 80 ms.

    Total time required for tran analysis `tran' was 90 ms.

    finalTimeOP: writing operating point information to rawfile.

    ******************

    DC Analysis `dcOp'

    ******************

    Important parameter values:

    reltol = 10e-06

    abstol(I) = 1 pA

    abstol(V) = 30 nV

    temp = 27 C

    tnom = 27 C

    tempeffects = all

    gmin = 1 pS

    maxrsd = 0 Ohm

    mos_method = s

    mos_vres = 50 mV

    Convergence achieved in 4 iterations.

    Total time required for dc analysis `dcOp' was 0 s.

    dcOpInfo: writing operating point information to rawfile.

    ************************************************************

    Periodic Steady-State Analysis `pss': guessed fund = 850 kHz

    ************************************************************

    =================================

    `pss': time = (0 s -> 10.8824 us)

    =================================

    Important parameter values in tstab integration:

    start = 0 s

    outputstart = 0 s

    stop = 10.8824 us

    period = 1.17647 us

    step = 6.17647 ns

    maxstep = 10 ns

    ic = dc

    skipdc = no

    reltol = 10e-06

    abstol(I) = 1 pA

    abstol(V) = 30 nV

    temp = 27 C

    tnom = 27 C

    tempeffects = all

    method = traponly

    lteratio = 3.5

    relref = sigglobal

    cmin = 0 F

    gmin = 1 pS

    maxrsd = 0 Ohm

    mos_method = s

    mos_vres = 50 mV

    pss: time = 276.2 ns (2.54 %), step = 10 ns (91.9 m%)

    pss: time = 816.2 ns (7.5 %), step = 10 ns (91.9 m%)

    pss: time = 1.366 us (12.6 %), step = 10 ns (91.9 m%)

    pss: time = 1.906 us (17.5 %), step = 10 ns (91.9 m%)

    pss: time = 2.456 us (22.6 %), step = 10 ns (91.9 m%)

    pss: time = 2.996 us (27.5 %), step = 10 ns (91.9 m%)

    pss: time = 3.546 us (32.6 %), step = 10 ns (91.9 m%)

    pss: time = 4.086 us (37.5 %), step = 10 ns (91.9 m%)

    pss: time = 4.626 us (42.5 %), step = 10 ns (91.9 m%)

    pss: time = 5.176 us (47.6 %), step = 10 ns (91.9 m%)

    pss: time = 5.716 us (52.5 %), step = 10 ns (91.9 m%)

    pss: time = 6.261 us (57.5 %), step = 4.706 ns (43.2 m%)

    pss: time = 6.802 us (62.5 %), step = 4.706 ns (43.2 m%)

    pss: time = 7.348 us (67.5 %), step = 4.706 ns (43.2 m%)

    pss: time = 7.894 us (72.5 %), step = 4.706 ns (43.2 m%)

    pss: time = 8.435 us (77.5 %), step = 4.706 ns (43.2 m%)

    pss: time = 8.981 us (82.5 %), step = 4.706 ns (43.2 m%)

    pss: time = 9.522 us (87.5 %), step = 4.706 ns (43.2 m%)

    pss: time = 10.07 us (92.5 %), step = 4.706 ns (43.2 m%)

    pss: time = 10.61 us (97.5 %), step = 4.706 ns (43.2 m%)

    Error found by spectre at time = 10.8824 us during periodic steady state

    analysis `pss'.

    V(VOUT) is too small to reliably detect the period of the oscillator. This

    may be because nodes with insignificant signal levels were chosen, or

    it may be because the oscillator was never properly started.

    Analysis `pss' terminated prematurely due to error.

    modelParameter: writing model parameter values to rawfile.

    element: writing instance parameter values to rawfile.

    outputParameter: writing output parameter values to rawfile.

    designParamVals: writing netlist parameters to rawfile.

    primitives: writing primitives to rawfile.

    subckts: writing subcircuits to rawfile.

     

    Filed under:
    • Post Points: 35
  • Wed, Mar 14 2012 9:26 AM

    • Tawna
    • Top 25 Contributor
    • Joined on Thu, Jul 10 2008
    • Snohomish, WA
    • Posts 209
    • Points 5,565
    RE: PSS analysis - oscillator Reply
    Hi Julia,
     
    First, you are using an extremely old version of spectre (5.1.41.xxx).   The current version of spectre is 11.1 and is
    not in the IC hierarchy.  Spectre now resides in the MMSIM hierarchy.
     
    Second, please look at Cadence Online Support Solutions:  (the first one is the most important)
     

    11534912 (COS) Recommended setup for running PSS on Voltage Controlled Crystal Oscillators (VCXO)
    11514284 (COS) FAQ: SpectreRF Harmonic Balance engine convergence tips for driven and autonomous circuits

    Please note that you must be a Cadence Customer on Maintenance in order to access these Solutions.
    If you are a University student, you will need to consult with your Cadence contact person at your University to
    get instructions to log in.
     
    tstab is important when starting oscillators.  I rarely need to touch steadyratio.  Also, I recommend using the Harmonic Balance rather
    than Shooting Newton engine.
     
    Best regards,
     
    Tawna
    Best regards, Tawna Wilsey Staff Support AE, Global Customer Support Cadence Design Systems, Inc.
    • Post Points: 20
  • Mon, Mar 19 2012 2:47 AM

    • JuliaP
    • Not Ranked
    • Joined on Wed, Nov 30 2011
    • Posts 2
    • Points 40
    RE: PSS analysis - oscillator Reply
    Hello Tawna,

    Thank you for your feedback and recommendations. I am going to use HB from now on.

    Best regards

    Julia
    • Post Points: 5
  • Mon, Mar 19 2012 2:34 PM

    • Tawna
    • Top 25 Contributor
    • Joined on Thu, Jul 10 2008
    • Snohomish, WA
    • Posts 209
    • Points 5,565
    RE: PSS analysis - oscillator Reply
    Hi Julia,
     
    You are quite welcome!  Have fun simulating!
     
    best regards,
     
    Tawna
     
    P.S.  Note that if you are simulating a ring oscillator, you would want to use shooting pss, not hb (harmonic balance).
    Best regards, Tawna Wilsey Staff Support AE, Global Customer Support Cadence Design Systems, Inc.
    • Post Points: 5
Page 1 of 1 (4 items)
Sort Posts:
Started by JuliaP at 14 Mar 2012 03:44 AM. Topic has 3 replies.