OPA735: The error when simulated with PSPICE for TI

Hi Miao,

I would simulate it in TINA-TI. There's a TINA-TI reference design:

https://www.ti.com/product/OPA735#design-tools-simulation

Kai

I have attached the OPA735 Tina-TI simulation model below.

OPAx735 TINA-TI Reference Design.TSC

Hi  kai,

      Thanks for your response!

Hi Marek ,

     Thanks for your support also !

     But my customer want to study and know what cause the pspice model's error.

     Could you kindly help to clarify how to correct the error in the model? Thank you so much!

Best Regards
Miao Bai

Miao,

The issue is that the circuit failed to converge - this means it couldn't find DC operating point.  This kind of issue is typically caused by either simulation tool limitation or op amp model problem.  However, since the identical model works just fine in Tina-TI (see below), it must be PSPICE for TI related issue - I have copied JC to address it.

Hi Miao,

Would you please share the project archive? ([FAQ] PSPICE-FOR-TI: How do I share PSpice for TI projects? - Simulation, hardware & system design tools forum - Simulation, hardware & system design tools - TI E2E support forums)

Thanks,
JC

Hi Marek & JC,

    Thanks for your support!

    I 'll  post the archive here if I get the customer's response.

Best Regards
Miao Bai

Miao,

I've been told by our macro-model developer that your customer used a legacy (old) model in their schematic and this is the reason for non-convergence.

The latest model is converging in PSpice of TI (~39s) - see below.  Thus, the solution is for your customer to use latest model-symbol.

Hi  Marek,

     Thanks for your support!

    My customer have tested the latest model: OPAx735 PSpice Model (Rev. E). It seems it is ok only when V3 is 7.18V , the other voltage is not worked also.

    Could you help to confirm it again? Thank you so much!

Hi  JC,

     Could you help to explain the error meaning further?

     My customer would like to know how to correct not only get a correct model.

     Thank you so much!

Best Regards
Miao Bai
TI Customer Support Center

Miao,

The circuit works fine for 5.9V<V3<8V.  For V3 below 5.9V the output is slammed against ground while above 8V the output is limited by positive rail.

Hi  Marek and JC,

    Good day! 

    Thanks for your response.

   My customer still wonder what is the error meaning?  He wants to know how to correct it, not just get a correct model.

   1)  Could you kindly explain detailed what is the meaning of the  mistakes ?

       What does V[OUT] = 1.712V  \ 1.297V mean?

Hi Miao,

Those simply means that these nodes have exceeded the convergence criteria.

Thanks,
JC

Adding to JC reply, the simulator must first find an initial biasing point (to converge) before ANY analysis can be perform.  All the errors like "these devices failed to converge" simply mean that for whatever reason simulator could not find an initial biasing point - this happens all the time and is caused either by the model or circuit issue or due to simulator's convergence algorithm limitation. However, if your customer used an old symbol with different pin names, etc. than the names used inside the macro-model, the simulator would not be able to reconcile each pin function vis-a-vis macro-model names and thus would issue an error.  All in all, there is no reason for customer to ponder on each errors because they are typically generic in nature.  Instead. they should do what it takes to solve the problem by making the necessary changes to eliminate such errors - like using the latest OPA735 symbol.

Even if the symbol is new, the LIB file might be older one! Therefore, to debug it choose ‘EXPAND’ and ‘LIBRARY’ checkboxes under the simulation profile to tell output which LIB file is exactly used for simulation - see below.

Below please find the zip files to be used for OPA735 simulations.

OPAx735.zip

Hi  Marek ,

    Good day!

    Sorry to bother you again.

   My customer try to correct the error, he find it work also when he changed  "Relative accuracy of V's and I's" from 0.05 to 100, do you think it is right or not?

Best Regards
Miao Bai

Yes, this is correct approach - changing analysis parameter default values is one way to help simulator to find circuit initial biasing point (to converge).

Great!  Good luck.