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: *

 Smoke Analysis on subcircuit 

Last post Tue, Jun 18 2013 7:44 AM by alokt. 10 replies.
Started by tt543 11 Jun 2013 12:45 AM. Topic has 10 replies and 1532 views
Page 1 of 1 (11 items)
Sort Posts:
  • Tue, Jun 11 2013 12:45 AM

    • tt543
    • Not Ranked
    • Joined on Tue, Jun 11 2013
    • Hod Hasharon, Israel
    • Posts 11
    • Points 205
    Smoke Analysis on subcircuit Reply

    Hello,

    I've been trying to run smoke analysis on my design for a while. I've managed to run it on active devices (FETs and diodes), but it refuses to run on passive components - subcircuit based .lib files. For example I have an inductor with the following model:

    .subckt ind_reg0 1 2

    L1 1 3 10e-6

    R1 3 2 0.3

    .ENDS

    I also added the following definition to device property file:

    ("IND_REG0"

    (Creator "XXXXX")

    ("device_info"

    ( SYMBOL_NAME "inductor" )

    ( MODEL_TYPE 2 )

    ( PORT_ORDER

    ("1")

    ("2") )

    )

    ("device_max_ops"

    (LI "0.11")

    )

    )

    When I run smoke analysis (after transient) I get the "non-fatal error: Current data not found for Smoke test LI" I've verified that I check "All" in data collection options. I also tried updating node names in pre-smoke section of template property file. But nothing helped.

    Does anyone have an idea how to run smoke analysis on subcircuit based passives?

    Thanks in advance for you help!   

    Filed under:
    • Post Points: 35
  • Tue, Jun 11 2013 1:23 AM

    • oldmouldy
    • Top 10 Contributor
    • Joined on Tue, Jul 15 2008
    • Woking, Surrey
    • Posts 1,454
    • Points 24,710
    Re: Smoke Analysis on subcircuit Reply
    This definitely works with passive devices from the supplied libraries. I have a source, R, L and Variables to set the limits. Variables has an LMAX which is used by a CURRENT property on the L part and this ends up as "LI" for the AA Smoke session. You might want to try this with "standard" parts with a hierarchical block for your part and see what data gets generated.
    • Post Points: 20
  • Tue, Jun 11 2013 2:18 AM

    • tt543
    • Not Ranked
    • Joined on Tue, Jun 11 2013
    • Hod Hasharon, Israel
    • Posts 11
    • Points 205
    Re: Smoke Analysis on subcircuit Reply

    I tried using passives from analog.lib: a source and L and R. I saw that both parts have already some properties deifned as variables: for example inductor has CURRENT property with value set to LMAX. I ran smoke and somehow it already had some ratings specified for the resistor, but not for the inductor. I tried defining variables  with PARAM part, but it didn't change anything.

    How these variables (LMAX,DSMAX,RMAX etc) should be specified to be passed to Smoke analysis as ratings? 

    • Post Points: 20
  • Tue, Jun 11 2013 11:19 AM

    • oldmouldy
    • Top 10 Contributor
    • Joined on Tue, Jul 15 2008
    • Woking, Surrey
    • Posts 1,454
    • Points 24,710
    Re: Smoke Analysis on subcircuit Reply
    Like PARAM is a "special" part for parametric analysis, Variables from the pspice_elem library is used to set the variables for Advanced Analysis, like Smoke. Place a Variables part and the "standard" Variables will be within it.
    • Post Points: 20
  • Tue, Jun 11 2013 11:27 PM

    • tt543
    • Not Ranked
    • Joined on Tue, Jun 11 2013
    • Hod Hasharon, Israel
    • Posts 11
    • Points 205
    Re: Smoke Analysis on subcircuit Reply

    First, thank you for your help.

    I added the Variables part. And yet it doesn't work for inductor. I could get smoke analysis  for a resistor (PDM, TB and RV) and a capacitor (CV and CI), but nothing for an inductor. Also, for some reason the resistor PDM gets 0% derating all the time, and I didn't specify it. 

    • Post Points: 5
  • Sun, Jun 16 2013 7:15 AM

    • tt543
    • Not Ranked
    • Joined on Tue, Jun 11 2013
    • Hod Hasharon, Israel
    • Posts 11
    • Points 205
    Re: Smoke Analysis on subcircuit Reply

    I have also one additional question. How can I use resistors and capacitors having only symbol definition (.olb) in smoke analysis. I know that pspice somehow recognizes such parts (once the value is specified) and can simulate it. These components also have RMAX/CMAX etc attributes assigned automaticaly. But once I run the Smoke analysis, it skips these parts.

    Thanks 

    • Post Points: 5
  • Sun, Jun 16 2013 8:44 PM

    • alokt
    • Top 25 Contributor
    • Joined on Thu, Aug 21 2008
    • Noida, Uttar Pradesh
    • Posts 243
    • Points 3,495
    Re: Smoke Analysis on subcircuit Reply
    Using "L" part from analog library should have worked for smoke analysis. I tried and it did work for me. It should test device for following two tests and show the results

    Current ratingDielectric testCan you share the log file or design files?

    About your question "resistors and capacitors having only symbol definition (.olb) " : Technically NO. It is true that PSpice can simulate these parts based on just olb, however PSpice Advance analysis modules like Smoke need few additional information. These details are not available on symbol instance. I would suggest you should use these symbols from Analog or PSpice_elem library (olb), this way tools can pick the necessary additional information from corresponding library files.About: resistor PDM gets 0% : This may be due to design/resistor parameters. Check for power dissipation in resistor, if this exceed it's rated value (which is by default 250mW) you would get a deration. Second condition under which you may observe this is when device temperature > Tbreak. Tbreak is function of Slope (RSMAX), and MAXTEMP. Refer Smoke chapter (page ~ 175-180) under "PSpice Advanced Analysis User Guide" details and example.
    • Post Points: 35
  • Mon, Jun 17 2013 12:38 AM

    • tt543
    • Not Ranked
    • Joined on Tue, Jun 11 2013
    • Hod Hasharon, Israel
    • Posts 11
    • Points 205
    Re: Smoke Analysis on subcircuit Reply

    What do you mean by "L" part from AWB? Is there AWB library? I couldn't find that one.

    Is there a way to replace all the resistors/capacitors in a schematic to point to another library (not one by one)?

    Thanks

    • Post Points: 20
  • Mon, Jun 17 2013 12:50 AM

    • alokt
    • Top 25 Contributor
    • Joined on Thu, Aug 21 2008
    • Noida, Uttar Pradesh
    • Posts 243
    • Points 3,495
    Re: Smoke Analysis on subcircuit Reply

    It is analog not AWB. You can use "Replace Cache" feature to replace all instances of a part in schematic with another library component. Go to project Manager, Expand Design Cache, Select part which you would like to replace, Do RMB and select "Replace Cache" and provide necessary input in dialog.

    • Post Points: 5
  • Tue, Jun 18 2013 5:27 AM

    • tt543
    • Not Ranked
    • Joined on Tue, Jun 11 2013
    • Hod Hasharon, Israel
    • Posts 11
    • Points 205
    Re: Smoke Analysis on subcircuit Reply

    I tried again the "L" part from analog library and it didn't do anything during the smoke analysis - I mean the summary table didn't show it. But when I looked into the smoke.log file I found the following lines:

    Reference Designator = L1

    The following parameter(s) exceeded the

    Maximum Operating Value(s) specified

    LI ( PEAK AVG RMS )

    Seems that it did run smoke analysis on the inductor, but doesn't display it in the results. Do you know how to make it show the inductor in the summary table?

    • Post Points: 20
  • Tue, Jun 18 2013 7:44 AM

    • alokt
    • Top 25 Contributor
    • Joined on Thu, Aug 21 2008
    • Noida, Uttar Pradesh
    • Posts 243
    • Points 3,495
    Re: Smoke Analysis on subcircuit Reply

    Please check if you have any visibility filters turned on. There are two possible filters 1. Component Filter 2. Parameter Filter.  Do RMB and check the status of these filter options.

    These two may hide a component or a set of specific test for a component from result table. I can not think of any other reason if log file indicate that component is tested.

     

    • Post Points: 5
Page 1 of 1 (11 items)
Sort Posts:
Started by tt543 at 11 Jun 2013 12:45 AM. Topic has 10 replies.