uA726 replacement

[Harry Bissell]

To the best of my knowledge the CEM 3340 is NOT a "heated" chip design... It has
excellent matching and compensation onboard so it works as well as a chip / tempco
design. There is a temperature compensation Output that can be used as a reference
for the internal expo converter - they cancel out. Check out the data sheet for
apps. The Prophet V handled the temp problem by "auto calibration" but you can use
the temp output from the chip as the DAC reference and get the benefit of
"temperature compensation" also

There are several "TomG" heated chips... One is just a heater... pass current and it
gets hot. Not regulated. I'm going to try a head to head with "that" design vs. a
regulated (feedback) heated chip.

The unregulated heated chip will get serewed up at high expo currents (self heating)
the regulated one will also, but would eventually correct... I don't know if this
corrects "fast" or "slow"

I'll let you know the shootout results after Aug 20 or so I've been deported to the
"Chicago Gulag" for next week.  :^) Harry

PS. Watch out for those cracks about Tubes or I'll sic my buddy Paul P. on you....

"Paul R. Higgins" wrote:

> I would be very surprised indeed if the "Tom G. style" heated 3046 chip wasn't
> as stable or even more stable than tempco resistor stabilized designs.  There's
> a discussion of exponential converter temperature dependence in Hal Chamberlin's
> "Musical Applications of Microprocessors" (as well as many other sources, I'm
> sure) that gives an equation which I can't remember offhand.  He points out that
> you can compensate for the temperature dependent factor in the equation by two
> means.  The first is to use a tempco resistor, which unfortunately is only an
> approximation (only valid at room temperature?).  The second approach is to
> actually heat the transistor(s) above ambient temperature.  However, to do this
> most effectively, you need all the expo converter transistors on the same
> substrate, so that they all track together.  Hence the popularity of the 3046
> and other transistor arrays.  By holding the temperature dependent factor
> constant, it effectively drops out of the equation.
>
> The only drawback to the heated chip approach is that additional current is
> required for the heater transistor.  I don't consider this a big price to pay
> for the stability it affords.
>
> In fact, the much venerated CEM 3340 VCO chip used this heated substrate trick
> with great success.  Warm-up time was < 30 sec, much better than (dare I say it)
> tubes.  In fact, Craig Anderton used the 3340 in his "Hyperflange and Chorus"
> design to get an exponential triangle LFO that was stable over a very wide
> frequency range (0.001 Hz to low audio frequencies).  It was surely a case of
> over-engineering, but it worked really well.
>
> PRH
>
> > I'll have to try this with a temperature stabilized 3046 vs the "hot" chip.
> > I'm still not convinced that just making it "hot" will really make the tempco
> > problem go away. I've still got my prototype... I'll measure the temperature
> > of the regulated chip, adjust the "hot" chip to the same temp... and let them
> > go head to head....  :^) Harry
> >
> > (PS to Tomg... I haven't tried that regulated 3046 with the external tranny
> > (prodigy was it???)  That looks good too.)
> >
> > Harry (holding his vote for "the way" until I see all the data) Bissell
>
>
> > tomg wrote:
> >
> > > >
> > > > The question is: Is there a LM3046/MAT-02 thermostating
> > > > scheme "blessed" by this group, and can you get it
> > > > to run using a single supply op-amp?
> > >
> > > My standard scheme of collector to +V, base to gnd, and emitter
> > > thru resistors to -V can be modified.
> > >
> > > Collector to +V (say 12V) and emitter 470 ohms from gnd. Then take 2
> > > resistors (say 10k) and split the supply( +6V ?).  Apply this to the base.
> > > Heats right up.
> > >
> > > -tg
>
> Paul Higgins
> email: higg0008 at tc.umn.edu
> University College, University of Minnesota