Michael Black wrote:

"John Walton" ) writes:

A Square Wave is all the harmonics of the fundamental, so a 1.9kHz square
wave will have all the harmonics which will complicate the measurement --

If you take a square wave and filter it you can get a relatively low
distortion fundamental -- it's much easier than you might think -- use an
NE556 dual timer to generate 1.9kHz and 1.3kHz square waves, combine with
10k resistors into an opamp buffer, filter with an MF-10 (now upgraded to
LMF100) switched capacitor low pass filter -- voila.


And why not start with a sinewave oscillator in the first place? That's
the way it used to be done, two transistors each making up phase shift
oscillators.

A square wave oscillator followed by a filter is, in some ways, easier
to implement than a phase-shift oscillator -- particularly if you want
well controlled frequency and amplitude without having to use
sophisticated AGC circuits.

A sqarewave won't complicate measurements, it will downright give
different results. The whole point of two-tone testing is that the
first tone causes a "carrier" out of the SSB transmitter (the single
tone translates to an RF frequency), and the second tone adds modulation.
If the tones aren't pure sinewave, the output of the transmitter will
be radically different.

That's probably why the guy doesn't advocate using it: he's recommending
a square wave that's heavily low-pass filtered to get rid of the
harmonics. If done right this will result in a clean stable sine wave
at the desired frequency.

--

Tim Wescott
Wescott Design Services
http://www.wescottdesign.com

I can get down to 0.01% THD with 4th order active filtering of a square
wave -- as measured on an HP339 THD analyzer. For the fellow who had asked
the original question I suggested the switched cap filter -- but there is
always going to be some noise clocking through. Using the interactive
filter software on Analog Devices or Texas Instruments website is pretty
easy, but necessitates more parts.

Sorry not to have included "odd harmonics" but I guess you know what I
meant.


"Tim Wescott" wrote in message
...
Michael Black wrote:

"John Walton" ) writes:

A Square Wave is all the harmonics of the fundamental, so a 1.9kHz
square
wave will have all the harmonics which will complicate the
measurement --

If you take a square wave and filter it you can get a relatively low
distortion fundamental -- it's much easier than you might think -- use
an
NE556 dual timer to generate 1.9kHz and 1.3kHz square waves, combine
with
10k resistors into an opamp buffer, filter with an MF-10 (now upgraded
to
LMF100) switched capacitor low pass filter -- voila.


And why not start with a sinewave oscillator in the first place? That's
the way it used to be done, two transistors each making up phase shift
oscillators.

A square wave oscillator followed by a filter is, in some ways, easier
to implement than a phase-shift oscillator -- particularly if you want
well controlled frequency and amplitude without having to use
sophisticated AGC circuits.

A sqarewave won't complicate measurements, it will downright give
different results. The whole point of two-tone testing is that the
first tone causes a "carrier" out of the SSB transmitter (the single
tone translates to an RF frequency), and the second tone adds
modulation.
If the tones aren't pure sinewave, the output of the transmitter will
be radically different.

That's probably why the guy doesn't advocate using it: he's recommending
a square wave that's heavily low-pass filtered to get rid of the
harmonics. If done right this will result in a clean stable sine wave
at the desired frequency.

--

Tim Wescott
Wescott Design Services
http://www.wescottdesign.com