Continue to Site

Welcome to EDAboard.com

Welcome to our site! EDAboard.com is an international Electronics Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.

Register Log in

[SOLVED] ADS simualtion Error problem

Status
Not open for further replies.
tony_lth

tony_lth

Advanced Member level 5
Joined
Jul 28, 2010
Messages
2,087
Helped
377
Reputation
756
Reaction score
368
Trophy points
1,363
Location
Beijing
Activity points
12,629
When I used ADS in simulation, I always can't combat the following ERROR, although I have done as the suggest from ADS. Anyone can help or suggest?

Solver convergence summary:
--------------------------
RHS:
2.221e-10 A at node XN3.DIODE4.a, harmonic index (9)
1.648e-10 A at node XN3.DIODE4.a, harmonic index (8)
1.068e-10 A at node XN3.DIODE4.a, harmonic index (7)

Update:
1.168e-08 V at node XN3._net670, harmonic index (2)
1.168e-08 V at node XN3.DIODE4.a, harmonic index (2)
1.053e-08 V at node XN3._net697, harmonic index (2)

Solution:
10 V at node XN3._net639, harmonic index (0)
10 V at node XN3._net700, harmonic index (0)
0.7419 V at node XN3.DIODE4.a, harmonic index (2)
--------------------------

-------------------------------------------------------------------------------
Simulation finished with errors.

I notice ADS always stop at 10V, but I don't know how to adjust the setting.
 

Attachments

  • ADS_suggest.jpg
    ADS_suggest.jpg
    41.6 KB · Views: 117

Maybe some nets do not have DC path and the harmonic index(0) can not be convergene to a fixed value.
 
Hi, kayaoo,

It surely have DC path. The problem is that I can't optimize properly. Somtime when I changed some setting, the optimization can go several iterations; next time maybe stop at the 1st iteration at another setting. I have tried Harmonic Balance and Envelope simulation for the same project, but the results are almost same. Is there anyway to solve the problem so that the optimization can run smoothly?

---------- Post added at 06:23 ---------- Previous post was at 05:24 ----------

Here is another error stop message, hope someone can help.
 

Attachments

  • ADS Error2.jpg
    ADS Error2.jpg
    54.6 KB · Views: 104

Of course, the Envelope simulation is based on HB. First you can try another solver type to improve convergence rate.
Could you share the project? then I can have a look
 
Hi, jayce, thanks for your reply. Could you give me your email box? I can send the project to you. Because the intellectual property, I can't share it at the forum.

Here is another error message, I have to increase the max iteration to 153600, the optimization become slower and slower.
 

Attachments

  • ADS Error3.jpg
    ADS Error3.jpg
    45.2 KB · Views: 116

Did you try to switch on "direct solver" or "krylov solver" ? I see on the picture you are not running with Krylov solver, sometimes this could help.
 
I had tried various convergence mode and slover, even read the official manual, and try to adjust various parameters.

Here is the simulation log file.
 

Attachments

  • error_Ky_Evenlope.txt
    8.6 KB · Views: 94

If I use Krylov slover, it stop soon, even not begin to solve one iteration. If I select robust matrix re-use, and direct solver, it can go at least 1 iteration.
 

Try Transient Assited HB with activating it clearly.

Show me netlists regarding Analysis statements.
 
Last edited:
Thank you, Pancho,

I opened TAHB, but TAHB closed auto when doing HB optimum.

The following is the netlist file. If you need more information, pls let me know.
 

Attachments

  • ADS_netlist.zip
    3.4 KB · Views: 77

You are invoking optimizer with HB analysis.

tony_lth said:
I opened TAHB, but TAHB closed auto when doing HB optimum.
Click to expand...
Do you mean TAHB does not work when optimizer is invoked ?

You set "FundOversample=8192" in HB setting.
Generally this is extraordinal.

You don't run HB_Noise Analysis, so "OSCwPhNoise" has no meaning.

Try setting "Restart=yes" in HB analysis.

Later I will run your netlist by using ADS2009U1.
 
Last edited:

Do you mean TAHB does not work when optimizer is invoked ?
Click to expand...

Yes.

You set "FundOversample=8192" in HB setting.
Generally this is extraordinally.
Click to expand...
The ADS suggest me increase more and more, and I found it works, although a little slow.

You don't run HB_Noise Analysis, so "OSCwPhNoise" has no meaning.
Click to expand...
I see.

Later I will run your netlist by using ADS2009U1.
Click to expand...
Thank you very much.
 

Last edited:
Can your goal target for optimization, "dBm(OCXO_9_1[::,9]) > 0.0" be truely satisfied with the following variable's ranges ?

Code: 
L1=9.2293 opt { 1 to 22 } \
L2=7.55931 opt { 1 to 51 } \
CM=4.40185 opt { 1 to 220 } \
C2=24.0249 opt { 1 to 470 } \
Cout=94.5039 opt { 0.1 to 100 } \
R=5.6 opt{ 3 to 30 }

Code: 
Optim:Optim1  OptimType="hpVMO" ErrorForm="L2" MaxIters=500 P=2 DesiredError=0.0
You use "Gradient" as optimization method.

Try to use other method such as "Hybrid".
Also reconsider goal target and variable's ranges for optimization.


As another possibilty for resolving your problem, I think you should try the followings.

Currently your simulation flows are like following.
Code: 
[Optimization of {L1,L2,CM,C2,Cout,R} for target dBm(OCXO_9_1[::,9]) ; Optim1 with HB3 setting Parametric Sweep of OCXO_freq ]



Instead try following.
Code: 
[Parametric Sweep of OCXO_freq ; Sweep1]
        [Optimization of {L1,L2,CM,C2,Cout,R} for target dBm(OCXO_9_1[::,9]) ; Optim1 with HB3]

Here you must not set sweeping OCXO_freq in HB3.
Instead sweep OCXO_freq by using "ParamSweep" and control "Optim".

See The Designer's Guide Community Forum - How to simulate this circuit in Spectre???

Later I also try this.
 
Last edited:
Can your goal target for optimization, "dBm(OCXO_9_1[::,9]) > 0.0" be truely satisfied with the following variable's ranges ?
Click to expand...
I just hope the result can reach my goal as close as possible.
Thanks for your reply, the optimum run properly regarding your suggests.
 

Can your goal target for optimization, "dBm(OCXO_9_1[::,9]) > 0.0" be truely satisfied with the following variable's ranges ?

Code: 
R=5.6 opt{ 3 to 30 }
Click to expand...

I checked again, the R range should be 1~5.6

Sorry for bring any inconvienience.
 

Thank all of you, I have got satisfied result.
 

Status
Not open for further replies.

Similar threads

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Back
Top