MS20 ESP Simulation (WOW)

[Magnus Danielson]

From: Harry Bissell <harrybissell at prodigy.net>
Subject: Re: MS20 ESP Simulation (WOW)
Date: Wed, 19 Jan 2000 21:05:34 -0500

Hi Harry!

> I've simulated part of the MS-20 ESP (TomG / Korg) and got a big surprise.
> I'm only looking at the part I didn't understand (re: the converter). The
> first two comparators form a zero crossing detector which fires a positive
> pulse to the converter.

Hmm... do you have a Spice file to share?

> At low frequencies (below about 1.5-2K) the converter froms a one shot
> multivibrator which gives a pulse every input cycle (as I expected...)
> 
> Above that, a strange (and wonderful) thing happens. The pulses begin to
> overlap, and the off time (zero) becomes fixed. So you get one, two, three,
> four cycles of positive level and then a fixed zero...

Aha, so pulse frequency converts into PWM.

> Not at all what I expected to see. I don't understand yet how it is
> happening... but it seems to work. This extends the range of the scheme
> above where the "tachometer" circuit should fail.
> 
> Above 5K the pulses go to a permanent "high" level and the output voltage
> can't rise any further... But that should have happened two octaves ago...

Ah, so you gain two octaves, that is usefull to know!

> I don't know why the output filter is linked to the "low cut" input filter.
> Somebody help me out ???

You tune your averaging time constant to match the range, basically it will
trim the low end properties so that the behaviour becomes related to the lower
end. As soon as I saw this I went "Ah, naturally - I would also do that" since
it was such an "obvious" solution (which is not obvious by the way).

Do you see?

> BTW if you got your ears on, JH.... isn't Korg engineering wonderful...
> I wish they hurry up and put the "Empire" out of business !!!

Hehehe ;)

All this talk soon forces me to fire up my MS-20 and actually play and measure
the ESP.

I never used the ESP for much, but when I do it has worked well for me.
An lovely idea that I think should be reused and possibly brougth into
perferction.

No one seems to have mentioned the use of PLLs for frequency tracking.
Also, you could use some quick method to bootstrap other methods into the
right range at least.

Cheers,
Magnus