Tri 2 Sin

[Martin Czech]

> >
> ....well I am pretty happy with the CA3080 ckt from electronotes via 
> Barry Klein. And I suspect that if you look inside the CA3080 you 
> find (suprise) the ckt typology you mention.. perhaps you could
> check next time you are coling a wafer.. ;-)
> 

No.

I know that inside the 3080 there is a diff amp input stage that can be
overdriven, that's not the point of my Email, but maybe my poor english
is too confusing, or I was confused because I was so happy to find that
article ;->

The point is that you can use more complicated arrangements of diff
amps with 4 or  6 or 8 etc. differential transistors in order to get a
much better sine approximation, 2 transistors yield 3% error (best
case) and for example 4 transistors are said to give only 0.02% error
(also best case). And (adding now) the article contains some
information about temperature compensation. Have you ever testet the
above mentioned 3080 converter at different module case temperatures ?
Say Tj=10C (winter situation) and Tj=50C (summer). If the differential
stage is heated up, it should appear that it is more overdriven, ie.
if you adjust the triangle amplitude at Tj=30C for minimum distortion
it may well be that at Tj=50C the sine output has a lot more
distortion. 

I've just made two (ideal) simulations which are plotted in the gifs
attached. In sinesine.gif you can see how bad the approximation with a
simple diff. amp. really is. As you can see the input differential
voltage is about 75mV which is about the optimum. It seems that a
higher input voltage range would give a better response, but in this
case the error arround 0 Volt will be much worse.  Only the positive
input voltage range is shown (symmetry) and for different
temperatures.

sine_err.gif shows the relative error in %. The circuit was adapted in
a way that at Tj=30C the error is minimized, other temperatures give
much more error (the whole simulation setup is a bit suboptimal, so the
error is about 5% and not 3%, but this makes no big difference).  The
temperature error could be compensated by reducing the input voltage
via a tempco network or by using some temperature coefficient for the
diff. amp. current source.

Of course this is all theory. In practice the question would be:  

1) Is the simple diff. pair approximation bad enough, so that
temperature changes are bareley audible ?

Paul, could you please do me a favour and simply listen to your ota
sine shaper, at room temperature and heated up (with a hair dryer or
soldering iron) ? Any difference ?

2) If the quad transistor diff. amp. gives much better distortion, will
the temperature effect be more audible ?

3) Is a picewise linear approximation better? There is a circuit in
Tietze/Schenk that is temperature compensated and needs no (or not so
good) matching of the diodes, compared to the Formant lfo sine
shapers.

And last but not least: is any loadspeaker capable to reproduce a sine
wave as such? Not mine! From this point of view it is a good quiestion
how much distortion of a "sine" wave can be tolerated.


m.c.

------------------------------------------------------------------------
m.c. has made it finally:  3 CDs out now; 72 min. minimum; "1"
(1994-1995),"2" (95-96),"three" (96-97); experimental stuff; mostly
Eimert/Stockhausen style; but also modern popular style











-------------- next part --------------
A non-text attachment was scrubbed...
Name: sinesine.gif
Type: image/gif
Size: 7728 bytes
Desc: sinesine.gif
URL: <http://synth-diy.org/pipermail/synth-diy/attachments/19980609/929ab204/attachment.gif>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: sine_err.gif
Type: image/gif
Size: 6224 bytes
Desc: sine_err.gif
URL: <http://synth-diy.org/pipermail/synth-diy/attachments/19980609/929ab204/attachment-0001.gif>