[sdiy] CEM3374 vco temp sensor 3394

[Karl Dalen]

The OBI Xpander manual says aproxemately:

:Because the DAC in the expander are a multiplying the
:voltage are directly propotional to the reference.
:When the DAC loads a voltage to the S/H to a VCO the
:temp sensor of that particularely VCO will be used 
:as the reference. Therefore if the internal temperature
:raises its frequency controll voltage will raise propor-
:tionally and keeping the VCO frequency stable.

:The temperature stability of the VCO,s frequency resulting
:of this method are far better than the one achieved with
:the CEM3340... snip...snip, because scaling are performed
:by the multiplying DAC instead of the analog multiplyer
:in the 3340 VCO.

(note each temp sensor from each voice are multiplexed by a 4051.
(the 4051 output goes later to the ref of DAC.)

>Martin  wrote:
>I wonder how this would work when multiple 3374s are controlled by the
>same DAC. Did they rely on similar temperatures for all chips? (I.e. DAC
>reference just from one 3374) Or did they multiplex back the references?

>jh wrote:
>I guess multiplexing is the only thing that makes sense. And it would be
>quite easy - if you multiplex the DAC output, you can as well multiplex
>the DAC reference input.

>Scott wrote:
>The sensor voltage mentioned most certainly will be a band 
>gap generated voltage.This can be scaled internally to +3300ppm by a
>simple linear amplifier and does not require a multiplier.

Now, interesting things creeps up when reading the CEM3394 paper, it says:

:The control scale of the VCO is temperature compensated with an internal
:+3300ppm tempco generator and multiplier. Thus, as chip temperature
:changes, the control voltage applied to the exponential generator
:changes proportional to temperature, effectively canceling the -3300ppm
:tempco of the control scale.

It seams the 3394 VCO contain the sensor of the 3374 and the
multiplier of the 3340? 

Jim's solution are novel but uses a ittle to many components
for my taste. Another interesting issue wich perhaps not have
been thought of:

:The relative amplitudes of the three waveforms have been set as
:follows to give approximately equal loudness The triangle is 27%
:larger than the sawtooth,which is about 27% larger than the pulse wave.

Usually waveforms are scaled equal!
quite clever dude that Mr Curtis!

KD 

- James Patchell <patchell at cox.net> skrev: > If I was doing it, I would
multiplex all of the band gap references so
> that 
> each dac was using it's own reference...although, that would add a bit
> more 
> complexity to the whole thing...

Höstrusk och grå moln - köp en resa till solen på Yahoo! Resor på adressen http://se.docs.yahoo.com/travel/index.html