Why TINA-TI shows THD=0.035% for Voltage Generator with Ri=0 ?

Pavel,

I am a modeling engineer in the WEBENCH modeling team.

What options did you choose for the TINA Fourier Series Analysis?
We have found the THD and individual harmonics can have a strong dependence not only on the circuit &  transient parameters, but also the Fourier Analysis parameters.
We have found this to be true for other Spice simulators as well.

I tried reproducing your results with  just a 1Vpeak, 10kHz sinusoidal source driving a 1k resistor.
The transient sim used the parameters you gave in your message.

TINA's Fourier Analysis gave anywhere from 0.06% THD to 4.7e-8% depending on the number of harmonics and the number of samples chosen in the Fourier Analysis options.

For example, if 128 samples with the default number of 10 harmonics gave THD = 0.011%.
If the number of harmonics increased to 30, the THD doubled. Increasing the number of harmonics to 40 caused the THD to increase to 0.06%.

If the number of samples was increased to the max (65536), the THD for 10 harmonics dropped to 4.65e-8%.
Increasng the number of harmonics to 30 gave THD = 7.2e-8% and 40 harmonics caused the THD to drastically increasae to 0.066%; same as the case of 128 samples..

I hope this helps.
If you don't get the results you are looking for, Is it possible to share your schematic or the TINA .TSC file itself?

Regards,

John

Pavel,

We usually do not try to capture nonlinear distortion behavior with our op amp models.
That being said, it can be useful to compare a model’s HD behavior to what is shown in the device data sheet.
This approach can be a useful tool to understand high level behavior & trends before any hardware is built or  evaluated.

 I downloaded your circuit file and tried some experiments with Fourier Analysis and the associated Transient sims.

 Experiment #1: Original circuit and sim parameters
With the default simulation analysis parameters the reported THD was 1%.
One reason to not trust this result that the Fourier coefficient for the fundamental waveform (k=1) is 11.26m, rather than the expected value equal to the signal’s rms value: 0.707*8Vp = 5.67Vrms.   We have found this kind of strange behavior is due to simulation parameters that are too relaxed. Default sim parameters are usually balanced trade-offs between speed and accuracy. It is common practice to tailor  sim parameters to met the requirements for accuracy & sim time.

 Experiment #2: Modify sim analysis parameters (DC)
The first step was to tighten the sim analysis parameters associated with DC convergence.
The following changes were made:

1. DC absolute current error was changed from a default value of 1n to 10p.
2. DC absolute voltage error was changed from a default value of 1u to 100n.
3. DC relative error was changed from a default value of 1m to 1u.

These changes resulted in a  reported output THD = 0.00754%.

 Experiment #3: Modify TRAN & FOUR analysis parameters for output THD
Digging into some of the details of the Fourier Analysis option via the Help menu found this feature uses only a single cycle of the fundamental signal to calculate the Fourier series coefficients. So the time duration for the Fourier Analysis begins at the Start Time and ends after a single cycle of the fundamental frequency. Any subsequent data is ignored. With this in mind, the 10ms start time for the Transient sim and the Fourier analysis was left unchanged, while the Stop time was reduced from 100ms to 11ms.
This decreased the sim time, but did not cnage the results relative to the 100ms Stop time.

In addition, the max transient step size was changed from the default value of 10G to 1u.
The reuced step size would result in a long sim time with the 100ms Stop time, but with a 11ms Stop time, the sim takes only a few seconds.

The reported output THD for the tightened DC &Tran sim parameters was 1.67e-6%.

 Experiment #4: THD of signal source
The next step was to look at the THD of the signal source using the DC & TRAN sim parameters described above.
If the input circuit was disconnected from the rest of the amplifer, and VG1 was loaded only by R7 and C8, the reported THD was 3.96e-9%.
If VG1, R7 and C8 were connected in the normal way, the THD increased to 1.77e-6%.

Snapshots of the TINA windows for the experiments above have been pasted into the attached circuit file.
The descriptions reference your schematic designations, but your schematic has been deleted.

 So it looks like a few of the many sim analysis parameters are really the key to some consistent results.
It may be necessary to modify other parameters, but without more information its hard to say anything definite.
One approach is to try tightening things further; if the results don’t change then you probably have optimized your sims and any additional changes won’t be reflected in the results.

I hope this helps.

Please let me know if you have any questions.

Best Regards,

John