Re: PWM Amp Design

"Genome" <ilike_spam@xxxxxxxxxxx> wrote in message
news:V1t6e.6093$C2.5422@xxxxxxxxxxxxxxxxxxxxxxx
>
> "Larry Brasfield" <donotspam_larry_brasfield@xxxxxxxxxxx> wrote in message
> news:Gcd6e.12$F24.425@xxxxxxxxxxxxxxxxxx
>
> [Garbage Spice Simulation Snipped]
>
>> The above simulation should be a convincing demonstration
>> that using the controller to get the PWM filtered output
>> response apparently desired by the OP is feasable and
>> unlikely to present serious problems. Obviously, gains
>> and maybe offsets will need adjustment once the VCVS
>> is replaced by the PWM IC. When current limiting is put
>> into place, some attention to limiting in the controller will
>> be in order. A sensitivity analysis for L and C variation
>> would be smart. It might be a good idea to make sure
>> no limit cycles are possible, using time domain simulation
>> and a range of step inputs.
....
> Let's just see the error report for your circuit,
>
> Questionable use of curly braces in "c1 vx 0 {7.2?/lpfs}"
> error: Error: unknown token in: "7.2[?]/lpfs"
> Questionable use of curly braces in "l2 va vx {22.5?/lpfs}"
> error: Error: unknown token in: "22.5[?]/lpfs"
> Questionable use of curly braces in "b1 va 0 v={{gx}*v(vc)}"
> error: Error: undefined symbol in: "(gx)*[v](vc)"
> Circuit: * C:\news\carlen\larry.asc
>
> Error on line 4 : c1 vx 0 (7.2?/lpfs)
> Unknown parameter "/lpfs"
> Error on line 5 : l2 va vx (22.5?/lpfs)
> Unknown parameter "/lpfs"
>
> Spice doesn't generate these sort of reports for general entertainment. It
> gives you indications about ERRORS in your circuit.

My simulation ran without diagnostics. There were
no '?' characters in the file I pasted into my post.
After filtering your post for useful content, I can
see that some process between my clipboard and
the post retrieved into my newsreader has replaced
certain ASCII characters with '?'. That is something
worth knowing, so, despite your caterwauling post,
I thank you for the tip.

> The analysis might run
> but there is the old adage about 'garbage in, garbage out'.

Such wisdom. Another gem.

> Shall we try again?
>
> Instead of 7.2?/LPFS try 7u2/LPFS
> Instead of 22.5?/LPFS try 22u5/LPFS
> Instead of .param LPFS=.1 try .param LPFS=1
> Instead of V={GX*V(VC)} try V=GX*V(VC)
> L1, that 250? inductor will be interpreted as a 250H inductor. Try the
> proper value of 250u.
>
> You could write V=80*V(VC) for B1 and include a note to the effect that you
> are assuming that the peak to peak amplitude of the modulator triangle wave
> is 2V and you are assuming a bridge supply of 80V. Then you might as well
> just use the proper values for the components.

I could, but that goes beyond the stated purpose
of my post. I could design the whole circuit, but
there would be no purpose to it beyond a silly
attempt to impress you and your ilk.

["... stupid ..." questions cut.]

> Oh, R2 serves no visible purpose.

It does if a behavioral voltage source is the only
consumer of what drove R2.

> So...... Let's just fix and tidy up your circuit.

[Cut circuit which also suffered from posting path
issues, but does simulate with group delay performance
way outside of anything the OP stated or implied.]

> Much prettier.

So you say. To simulate the PWM, I would revert
the VCVS back to a behavioral source so that the
PWM delay characteristic could be added.

> You're doing it again, I've told you about this before. You are plotting the
> closed loop response rather than measuring the loop response. See how I've
> inserted an AC source, VAC, in the loop.

The loop gain is not especially interesting, being readily
determined by inspection. I made no effort to plot it.

> Do an AC analysis, the circuit as given is set up to do that, and plot
> V(VERR)/V(VC). Notice how it crosses over at 400KHz with little to no phase
> margin.

What you call VERR is not what is commonly called the
error. As for doing AC analysis, I don't know quite what
you simulated, but my controller, with loop gain set to bring
the complex poles well into the LHP, has plenty of phase
margin and a well behaved step response.

[More of same cut.]
[More folderol cut.]

--
--Larry Brasfield
email: donotspam_larry_brasfield@xxxxxxxxxxx
Above views may belong only to me.


.