[sdiy] Four new lectures up

[Antti Huovilainen]

On Thu, 20 Apr 2006, Aaron Lanterman wrote:

> 4/12 Digital emulation of a SSM2040/Prophet 5-style filter, including Antti 
> style tanh nonlinearities - I did an OTA instead of a ladder stage since I 
> think the OTA version is easier to understand.

Too bad SSM2040 is not quite OTA (as JH pointed a year or two ago) :).
In SSM2040 you have both hard clipping when the output capacitor falls 
below diff pair ground level and asymmetric diff pair distortion due to 
Early effect and cheapo current mirror (interestingly, later CEM filters 
are really close to OTAs in behaviour). I really need to get around doing 
a proper nonlinear model for SSM2040 someday soon.

> ...I fixed it in next lecture (after I discovered "impulse invariant"

You should use the "Compromise" method Tim Stilson & Julius Smith 
presented way back in 96. Basically it uses a [1/1.3 0.3/1.3] FIR section 
before each one-pole section to get the loop phase shift & freq response 
close to needed.

Antti

"No boom today. Boom tomorrow. There's always a boom tomorrow"
   -- Lt. Cmdr. Ivanova