I've been using Texas Instruments "TINA" simulator for a while now and it has been helpful. I must admit it has saved me a re-spin of circuit board design or two.
Assuming I have described the real world conditions accurately, is it always right?
I have made mistakes like not taking into account the signal source impedance and such but so far it seems to be always right once I've described the real conditions correctly.
Specifically I'm wondering if its' FFT simulation can be counted on. It outputs a THD, total harmonic distortion, figure.
I don't care to own a stand alone spectrum analyzer insturment- big $.
Maybe I should be considering a software FFT for my laptop or something to verify the real circuit matches simulation. Of course then my signal generator would need to put out a prefect sign wave - I don't think it does. Perhaps one can synthesize this pretty good with a computer/software too
What has been your experience?
Assuming I have described the real world conditions accurately, is it always right?
I have made mistakes like not taking into account the signal source impedance and such but so far it seems to be always right once I've described the real conditions correctly.
Specifically I'm wondering if its' FFT simulation can be counted on. It outputs a THD, total harmonic distortion, figure.
I don't care to own a stand alone spectrum analyzer insturment- big $.
Maybe I should be considering a software FFT for my laptop or something to verify the real circuit matches simulation. Of course then my signal generator would need to put out a prefect sign wave - I don't think it does. Perhaps one can synthesize this pretty good with a computer/software too
What has been your experience?
The sims are only ever as good as the models, including your models of the parasitic values in the layout, and some of the supplied models are rather lacking in terms of modelling things like CMRR.
Simulation is useful, but I would never consider it definitive for anything much, ya got to measure your real device to verify that the simulation is correct on this occasion.
A good soundcard and software can take you a long, long way in audio frequency measurement, not a good as a Prism or AP, but more then sufficient for many things.
Regards, Dan.
Simulation is useful, but I would never consider it definitive for anything much, ya got to measure your real device to verify that the simulation is correct on this occasion.
A good soundcard and software can take you a long, long way in audio frequency measurement, not a good as a Prism or AP, but more then sufficient for many things.
Regards, Dan.
It's just SPICE, and as such is best used as a design and development aid, to be confirmed by careful bench testing. It's not always right.
The device models are approximate and the pcb layout is critical, for example.
This is what I was going to say.
Simulation assumes "perfect" grounds and zero parasitics. Poor grounding can introduce distortion and even instability.
You can add some inductance and resistance to your simulation to see how it affects your circuit.
A sim THD calculation is normally small-signal, assuming that the signal variations in current and voltage in the circuit are very small. So this always is much more optimistic, except maybe if you also measure with very small signals.
OTOH 'real' distortion measurements are more useful using transient analysis and spectrum analysis. Using an FFT analyzer you can set the signal levels as you like and the sim will use that to calculate the spectral lines which gives you distortion info at any signal level you want.
Jan
THD and FFT are result of the same transient analysis and this is not small-signal analysis.
Not in my sim, there the THD (which is a graph of harmonic distortion over frequency) is definitely small-signal, with a unity input signal.
Maybe this differs from sim to sim. Mine is the Proteus suite from Labcenter in the UK.
Jan
I use LTspice, and it does not show a graph of harmonic distortion over frequency. There some possibility to do that, but from THD generated with transient analysis, so it will be signal level dependent.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.