[sdiy] tempco puzzle

[Seb Francis]

What would you say the target "smallest noticeable difference" would be for a perceptive human?  Paul Perry suggested 1dB, but does this also depend on the actual volume?

I did the following calculations ...

Assume an input voltage of 1V for the VCA (the control port follows a control law of ~ -30mV/dB, so this gives a gain of ~ -33dB).  For every 10C of temperature change, the control law changes by 3.5% which corresponds to 1dB at this attenuation.  At lower attenuations, the temperature becomes more significant, but then perception also becomes less.

The effect of temperature is probably more of a problem for > unity gain (e.g. with an input voltage of -1V, gain of +33dB), because then the sound is louder and smaller changes in volume can be perceived - I guess?

I'd like to base my decision on some maths, but as I said above I'm rather in the dark about what target "smallest noticeable difference" to aim for.

Seb



Chris Crosskey wrote:

> No you don't....not for a mixer or a VCA....humans can't hear that sor of
> variation in loudness.....it's not worth the hassle...if it was pitch it
> would be worth doing....
>
> chrisc
>
> > -----Original Message-----
> > From: Seb Francis [mailto:seb at is-uk.com]
> > Sent: 01 October 2002 03:38
> > To: synth diy
> > Subject: [sdiy] tempco puzzle
> >
> >
> > Hi,
> >
> > I'm building a Mixer/VCA with an SSM2018 I got lying around.
> > It has a CV tempco of -3500ppm/C, so I need to build the CV
> > input stage with a 3500ppm/C tempco.
> >
> > The required gain of the CV input stage is -0.25, and I have
> > 1K 3900ppm/C tempcos.
> >
> > The question is .. how to do it with *minimum* extra components?
> >
> > The only way I can work out uses 3 inverting apamps:
> > CVs => Inverting unity-gain summer
> > => Inverting unity-gain
> > => Inverting 0.25 gain with the 1K tempco + 150R 100ppm/C
> > resistor in the feedback loop
> >
> > I've been racking my brains for hours trying to work out a
> > more elegant way to do this, but have got nowehere (maybe I'm
> > trying too hard over 3 opamps, but it's nice to find the
> > neatest solution ;-)
> >
> > Single opamp + potential divider with tempco doesn't give a
> > near linear coefficient because the gain is not small enough.
> >
> > There must be a neat way to do this!
> >
> > Seb
> >
> >
>
> This e-mail, and any attachment, is confidential. If you have received it in
> error, please delete it from your system, do not use or disclose the
> information in any way, and notify me immediately.