On Sat, 31 Mar 2007 02:18:09 +0100, "john jardine"
wrote:


"John Larkin" wrote in message
.. .
On Fri, 30 Mar 2007 16:05:14 -0700, "Joel Kolstad"
wrote:

"Fred Bartoli"
fred._canxxxel_this_bartoli@RemoveThatAlso_free. fr_AndThisToo
wrote in message ...
A small cap (1n or so) across the 1K feedback resistor will be welcomed
too.

Is that meant to surpress any spikes when the '4051 switches (i.e,. if it
"breaks" before it "makes")?

It's kind of a clever circuit, IMO. What do the circuit design gurus
think of
it?

I'm waiting to hear someone suggest you could do this with an 8 pin PIC
or
AVR... 4 lines to the '4051, and another one to receive frequency
commands
over, e.g., RS-232. :-)


Just connect four pic pins to an output node with 1k, 2k, 4k, and 8k
resistors, and you're done.

John


Seems about 30% THD?.

Is that what you get with 16 amplitude levels spaced, presumably, at
16 even time intervals? I'd guess somewhat better than 30%; a square
wave is only 43. Possibly you could make a better sine wave if the
resistors weren't binary weighted. More math than I care to sign up
for!

We just did a homebrew 8-channel DDS waveform generator, with 16-bit
math and 14-bit dacs, 2048 point sine lookup table. Sinewaves look
pretty good.

John