[sdiy] Korg Poly61 a peculiar design!!

[Tom Wiltshire]

Have you got a link to some decent schematics for the Poly 61, Karl?

I've got a service manual PDF for it, but it's more-or-less unreadable.

Thanks,
Tom

On 6 Apr 2011, at 01:44, karl dalen wrote:

> VCF:
> 12db LP LM13600 using a MS20 VCA with limiting diodes in resonance
> feedback but before VCA also the feedback not only goes to the input
> of the filter but to first cap like half way S/K and no local OTA
> gain loops but one fixed over both. 
> 
> I have checked trough all other analog Korg products and this VCF
> are unique to the P61.It should have a peculiar Bw hump because of
> this.
> 
> VCA:
> MS20 final VCA. One SSM2056 EG split use VCA and VCF.
> 
> DCO1:
> It's a complete saw VCO core, a regular Op amp integrator/comparator
> design. The ramp reset comparator also feeds another comparator that
> sends this signal back to *both* 1/3 of DCO1 and DCO2 8253 timers gate
> input, then one of the 8253 timers sends a pulse back to the integrator
> NPN ramp reset transistor in the VCO!! Odd!
> 
> Could this perhaps mean that half of timers are actually used to measure
> the frequency of the VCO and if wrong in realtime restart it on a per
> cycle basis?! It could work if the value are compared to a predefined
> LUT. Or is it just a bogus thing to distract anyone viewing the schematic?
> 
> The output of the ramp integrator are sent to a comparator to create
> PW wave.
> 
> DCO2:
> DCO2 are made out of 1/3 8253 driving a 4520 divider who's output's
> are selected by a switches and summed by a weighted resistor net
> acting like a DAC to create saw. A redundant signal goes to divider
> reset to actually turn off the chip..
> 
> Apparently resolution of wave are likely dependent of octave selected
> otherwise only one switch would be needed.   
> 
> VCO ramp integrator note CV are a 7bit R2R generated by the same CPU
> that scan's KBD that then feed a common expo who are muxed out to
> each VCO. (Polysix/Trident) 
> 
> The driver for the timer DCO's are a VC RF multi vibrator who in turn
> are driven by bend and vib voltages, the CLK signal are then driven
> into a prog divider to create semitone offsets for DCO2, finer detuning
> presumably by 8253 timer itself.
> 
> There are no regular muxed DAC for the parameter CV's instead each
> has its own DAC trough a bag of 8255 driving a even larger bag of
> weighted resistor net's.
> 
> DAC for Cf of VCF are only 6 bits. However PWM, VCF, Vibrato modulations
> are full resolution from the analog triangle MG.
> 
> In other words an awful sounding poly but a great, peculiar and funny
> VCO/DCO design. 
> 
> Disclaimer: I might be utterly wrong in all that said here.
> 
> Thanks for your attention....G'day...
> 
> KD
> _______________________________________________
> Synth-diy mailing list
> Synth-diy at dropmix.xs4all.nl
> http://dropmix.xs4all.nl/mailman/listinfo/synth-diy