Mono/Poly VCO heater (was: beating a dead DAC )

[Russ L]

JH,

Thanks for the in-depth report. In my Mono/Poly, there is a small
probably 1/8W or 1/4W resistor on top of each SSM2033 with
some thermal compound. I haven't opened the synth again, but
I read through the datasheet, and there is a little footnote about
temperature drift being 3300ppm without that part of the loop.
I couldn't find a tempco in the parts list, but they do specify
a thermistor for the filter compensation. If I open it again I'll
measure the resistor on the 2033. My schematic is more-or-less
unreadable. I don't remember any external high-wattage resistors
that Osamu mentioned. At least not in my Mo/Po.

Russ L
askeyman at earthlink.net

jhaible wrote:

> Hi again,
>
> I've been digging deeper into the schematics now, and this is how I see it:
>
> The Mono/Poly has two temperature regulation loops.
>
> For the record: the SSM2033 has an on chip temerature sensor,
> an on-chip heater, and the necessary control amplifier, and it's
> intended to be a self contained thermostat circuit. This means
> drawing a rather high current, of course.
>
> Now the Mono/Poly circuit does limit this current to a smaller
> value by inserting a 150R resistor into the heater's power supply
> connection. An external circuit watches the voltage drop across
> this 150R resistor and controls an external heating circuit, including
> the power resistor that is located next to the 2033.
>
> Think of it as a cascaded servo loop. When the chip is too cold,
> the internal heater is activated. But this internal heater is artificially
> starved, and can only draw so much current. As soon as the internal
> heater hits its (reduced) limits, an external "booster" is turned on.
>
> This has the clear advantage that the chip suffers much less stress
> from high currents. It still has the advantage of a fast internal
> servo loop for small errors. Only fast or large deviations from the
> reference temperature are handled by the external heater.
> In both cases, the internal sensor controls the loop.
>
> I cannot tell where they set the threshold between the two loops.
> Maybe the external resistor handles only the rough stuff (like
> heating up on power-on) - maybe the external resistor handles
> all but the tiniest deviations.
>
> Apart form the advantage of increasing the chip's life, there
> might or might not be an advantage of better regulation, too.
> Remember the thread about the disadvantage of the heater
> method vs. tempco method ? Non-symmetric heat up and
> cool down time constants ? This might be improoved here,
> as part of the thermal capacity (the power resistor) is
> outside the isolation of the chip package.
>
> All right, part of this last one might be speculation. But
> I have a feeling that I've learned a lot today. I think it's a brilliant
> solution, and once more a Korg circuit has caught my highest
> admiration !
> And thanks to Osamu for pointing this out!
>
> JH.
>
> PS.: I just noticed that they also replaced the internal CV summing
> opamp with an external JFET type. Both modifications must have
> been designed after the main circuitry (external pcbs).
>
> > I checked the article again.  The designer said:
> >
> > "In Mono/Poly, temparature is directry controlled by heaters
> > (resisters).  Surface temparature is about 65 degrees.  Other makers
> > were using thermisters but it is very difficult to stabilize.  Direct
> > temparature control is much easier and more stable, even though its
> > primitive.  Stabilization took 3 or 4 minutes."
> >
> > I've never looked at Mono/Poly schematics.  But, I also have a real
> > Mono/Poly.  It has big resisters on/above VCO chips.  (Maybe the
> > resisters are temparature sensors.)
> >
> > - -- ---- ----------- --- --- - ----- -- --- --- - - ------ - - - --
> >  Osamu HOSHUYAMA          Digital Signal Processing Technology Group
> >                     C&C Media Research Laboratories, NEC Corporation
> >                Email: houshu at ccm.CL.nec.co.jp,  Fax: +81-44-856-2232
> > --------------------------------------------------------------------
> >
> >