SV: Re: SV: Re: [sdiy] CEM3328 seems to only have buffers at the end

[karl dalen]

--- Antti Huovilainen <ajhuovil at cc.hut.fi> skrev:

> On Wed, 22 Feb 2006, karl dalen wrote:
> 
> > That picture dosent tell me anything!
> 
> The picture should tell you that Moog core clips at roughly 1/4 the signal 
> level compared to OTA filter (due to OTA nonlinearities cancelling 
> better).

With core you mean the first differential input pair?

> Practical meaning is that if you set up Moog & OTA filter so that core 
> sees same signal level, the self oscillation amplitude is 4x for OTA _and 
> therefore any clipper in feedback loop affects OTA more than Moog when 
> other things are equivalent_.

Obviously, but i dont se the point, they are off two different designs!
Its like comparing apples and oranges.
> 
> > Here is something for you to do.
> > Draw up the ladder stage in the simu machine,
> > add some bias and "unbalance the ladder"
> > what do you get.
> 
> Somewhat more distortion and more even order harmonics. The general signal 
> level isn't really affected (tested with 10mV mismatch). In OTA filter any 
> DC offset only affects the DC offset of the output signal for constant 
> cutoff.
> 
> Now I challenge you to do the same tests without simulation. Oh, and 
> you're not allowed to spend more than 15 minutes on it ;) (do you now see 
> the point of simulation?)

No problem, the core are allready set up so it takes about 15 sec to do it!
Also i dont have the hassle with incomplete models or inaccurate models! ;-)
Alltought i do simulate in LT spice and im puzzled with all the strange 
errors i get!! :-) 
 
> > Antti, what is it you are trying to prove?
> 
> Prove?
> That Moog ladder distortion is not due to the input differential pair 
> alone (or even mostly - all stages are more or less equivalent).

OK! But that's allready known!
Who had said that it was like that?

KD