Korg Poly 800 Oscillators

[>>>marjan<<<]

> It used the MSM5232 chip for tone generation and had one Juno-2 (or maybe
> Juno 106) filter/VCA chip and a nice analog stereo chorus at the output.
> Anyway, since Roland was also using the OKI chip, it must not have been a
> custom-made proprietary Korg part.  

It is NOT custom part, indeed, it's available online from OKI but you
should
be registered or something to get datasheet.

>It's quite possible that they also sold
> the chip to some other manufacturers - possibly some organ makers, most of
> whom were eager to add "synthesizer" features to their organs around that
> time.
> 

Very possible.

> And yes, it is purely clock-derived DCO.  Any analog CV inputs it has are
> used to control its internal volume VCA's, and are not used for pitch
> control.  Much like the often-used 8253 triple timer chip, the Poly 800's
> individual voice pitches are determined by an externally-provided 8-bit
> binary word which is multiplexed to each voice via a 3 or 4-bit word, then
> latched into the the chip at a very high rate.  The master tuning, pitch LFO
> and pitch bender CV's are used to modulate a high-frequency clock, which
> determines the overall pitch range of all the voices (just as it's done on
> the 8253-basd DCO's in the Juno60, Juno106, JX3P, JX8P, GR700, MKS30, JX10,
> MKS70, MKS10, MKS7, Juno1 and 2, Matrix 6, Matrix 1000, and a few others).
> 

Do you have complete schems for JX3P? I have some incomplete from BJ's
old site. I'd like to repair mine and add variable chorus and FM to VCF.
Any ideas how to repair double-taper bender pot as mine is missing
one of those tapers, so it bends just one way?


> The Poly 800 service manual has an internal block diagram of the chip.  Would
> this be at all helpful to you?  It may not show all of the pin I/O's, though.
>  The Roland module's schematic doesn't really say much at all about the chip.
>  I doubt that you would be able to achieve any sort of individual voice PWM
> in there, since the individual voices are not brought to output pins until
> after going through the internal VCA's.

I'd like to see it. It's not included in my set of schems.
I just noticed that there are some unused pins and hoped that something 
more could be squeezed from it.

Now, if anyone could help and show me to some simple increment/decrement
dial schematic, to get rid of those two stupid buttons...


> 
> << Also I have a plan to install sort of dynamic harmonic PWM (based on
> John simonton's simple ADC - but more about it when I test it). >>
> 
> Yes - please explain more!  It sounds interesting.

Well it's just concept to be tested. It could be a module for itself
also.
I'll try to explain: each osc has 4 outputs for harmonics
(2,4,8,16),which
are via cap and res summed to a opamp. Each voice uses that same ouputs
(muxed internally), and also ext signals determine which harmonics are
on and which are off (sort of additive synthesis). So you switch on all
harmonics and they are summed. Between each cap and summing resistor
I'll
put 4066 switch to gnd. Now I get a pot, feed it with DC or LFO from
chorus
(triangle,saw,S/H) and taper to input of Simonton's simple 4 bit ADC.
Depending on input voltage 4 outputs of ADC (bits) are O or 1 and each
that output feeds one of those 4 cmos switches (ie to their control
nodes
to switch them on/off). With some caibrating you could use envelope
outputs to controll too. So As lfo , say triangle, ramps down you slowly
have 16 mixed in,then 8,then 8 and 16, then 4, then 4 and 16, then
4/8/16
and so on which should make those static waves more dynamic sinically,
esp if you play dual mode, mono lines, so you modulate just one osc,
or both but one inverted. If you paly mono mode (full 8 polyphony)
you should switch it to all summing amps are controlled the same.

  

> That's a very nice filter chip.  To my ears, it sounds much better than the
> CEM3320 and all of the other Curtis chips that contained both VCF's and VCA's
> (CEM3372, CEM3379, CEM3396, etc.) - better low end and squishier resonance.

Yes resonance is ear pircing once you trim internal trimpot.
I used to process my M1 thru it (no dynamics, just manual cutoff/res)
to get warmer pads.

> The NJM2069 was also used in the DW6000, DW8000, EX8000, and the old Korg DSM
> 12-bit samplers.  I really love my EX8000!  


Did DSS-1 have it too?


>I wonder if Korg still has those
> chips, and how much they cost?
>

I'll advise you and others to try that. Undicovered secret weapon :)
 
> Please keep us informed about your Poly 800 mods.
> 

I will. It's fuuny how some mods are easy when I look at the schems
and solder it, but when it needs to be written down it becomes very hard
:)


Cheers,


marjan