Triangle to sinus for high frequencies - how?

M

Mat Nieuwenhoven

Guest
Hi,

For radio amateur porposes, I'm building a 0.1 - 30 MHz sweeping oscillator
(not quite in one sweep) with the 74LS629. I'd like to take off the triangle
waveform which exists over the timing capacitor with a video amp like the
NE5539 and turn it into a reasonable sinus. I've tried the trick with two
antiparallel diodes (1N4148) fed via a series resistor by the triangle, but
that works only to about 1MHz, presumable because of the capacitance of the
diodes.
What is a good method of doing this? Any ideas are appreciated.

Mat Nieuwenhoven
 
Mat Nieuwenhoven wrote:
Hi,

For radio amateur porposes, I'm building a 0.1 - 30 MHz sweeping oscillator
(not quite in one sweep) with the 74LS629. I'd like to take off the triangle
waveform which exists over the timing capacitor with a video amp like the
NE5539 and turn it into a reasonable sinus. I've tried the trick with two
antiparallel diodes (1N4148) fed via a series resistor by the triangle, but
that works only to about 1MHz, presumable because of the capacitance of the
diodes.
What is a good method of doing this? Any ideas are appreciated.
Click to expand...
The usual approach is a diode network with 4 to 5 diodes for each
polarity. Yes, the driving OpAmp has to have the bandwidth.

alternatively :
I strongly recommend to use a DDS from Analog Devices. Accurate
to the miliHertz or even better.

Rene
--
Ing.Buero R.Tschaggelar - http://www.ibrtses.com
& commercial newsgroups - http://www.talkto.net
 
On a sunny day (Mon, 29 Mar 2004 21:28:56 +0100 (CET)) it happened "Mat
Nieuwenhoven" <mnieuw@dontincludethis.zap.a2000.nl> wrote in
<zavrhjqbagvapyhqrguvfmncnay.hvdaw83.pminews@news.text.chello.nl>:

Hi,

For radio amateur porposes, I'm building a 0.1 - 30 MHz sweeping oscillator
(not quite in one sweep) with the 74LS629. I'd like to take off the triangle
waveform which exists over the timing capacitor with a video amp like the
NE5539 and turn it into a reasonable sinus. I've tried the trick with two
antiparallel diodes (1N4148) fed via a series resistor by the triangle, but
that works only to about 1MHz, presumable because of the capacitance of the
diodes.
What is a good method of doing this? Any ideas are appreciated.

Mat Nieuwenhoven


It is difficult, you would need a tunable filter.
Click to expand...
The onluy way *I* know how to make a sine sweep .1-30MHz is the old fashioned
one.
Use one osc at say 100 MHz, and sweep an other one from 100.1 to 130 MHz.
Feed both into a dual gate MOSFET mixer followed by a 40 or so MHzlowpass.
You can sweep a 100 MHz osc with a varicap really fast.
Any other way that only requires 3 transistors and one varicap I'd like to
hear about it :)
JP
 
"Mat Nieuwenhoven" <mnieuw@dontincludethis.zap.a2000.nl> wrote in
news:zavrhjqbagvapyhqrguvfmncnay.hvdaw83.pminews@news.text.chello.nl:

Hi,

For radio amateur porposes, I'm building a 0.1 - 30 MHz sweeping
oscillator (not quite in one sweep) with the 74LS629. I'd like to take
off the triangle waveform which exists over the timing capacitor with
a video amp like the NE5539 and turn it into a reasonable sinus. I've
tried the trick with two antiparallel diodes (1N4148) fed via a series
resistor by the triangle, but that works only to about 1MHz,
presumable because of the capacitance of the diodes.
What is a good method of doing this? Any ideas are appreciated.

Mat Nieuwenhoven
Click to expand...
Tektronix,in the TSG6 sweep generator module of the 1410 NTSC TV signal
generator,used a Tek-made custom IC to shape triangle to sine waves at
frequencies up to 20 Mhz.It used a resistor-diode network on a thick-film
substrate,a large hybrid IC. I do not recall the distortion specs.

--
Jim Yanik
jyanik-at-kua.net
 
Mat Nieuwenhoven wrote...
For radio amateur porposes, I'm building a 0.1 - 30 MHz sweeping oscillator
(not quite in one sweep) with the 74LS629. I'd like to take off the triangle
waveform which exists over the timing capacitor with a video amp like the
NE5539 and turn it into a reasonable sinus. I've tried the trick with two
antiparallel diodes (1N4148) fed via a series resistor by the triangle, but
that works only to about 1MHz, presumable because of the capacitance of the
diodes. What is a good method of doing this? Any ideas are appreciated.
Click to expand...
The simple answer is that this project can give you a sinus headache.

Seriously, though, several posters have described how to convert a
triangle waveform into a sine wave. But be careful, the LS629 doesn't
create a ready-to-use triangle. The LS629 oscillator is an emitter-
coupled multivibrator, and looking at either of its terminals shows a
triangle on top of a squarewave. The triangle you're seeking is across
the two capacitor terminals, and requires a differential amplifier to
retrieve it. The '5539 is a high-gain video opamp, and not directly a
differential amplifier. An additional problem is that you'll need a
high common-mode rejection to 30MHz. That's tough, although wideband
parts like the MAX435 and MAX436 may have a chance. A better solution
may be to use a serious wideband VCO oscillator chip like the MAX038.

Thanks,
- Win

whill_at_picovolt-dot-com
 
On Tue, 30 Mar 2004 08:14:15 -0800, mike <spamme0@juno.com> wrote:

[snip]
The slickest fix I've seen is what HP did in some of their function
generators. You put a resistor in series with the timing cap.
Comparator senses the top of the resistor, but the output still comes
off the cap. As the current increases, the resistor adds more volts to
the sensed triangle, effectively reducing the amplitude of the real
triangle. Tweek the resistor for a first order correction. Helps a LOT.

[snip]
Click to expand...

I discussed this technique here several years ago. It not only fixes
the delay overshoot problem but also linearizes the control curve when
used for a VCO.

...Jim Thompson
--
| James E.Thompson, P.E. | mens |
| Analog Innovations, Inc. | et |
| Analog/Mixed-Signal ASIC's and Discrete Systems | manus |
| Phoenix, Arizona Voice:(480)460-2350 | |
| E-mail Address at Website Fax:(480)460-2142 | Brass Rat |
| http://www.analog-innovations.com | 1962 |

I love to cook with wine. Sometimes I even put it in the food.
 
Mat Nieuwenhoven wrote:

On Mon, 29 Mar 2004 21:28:56 +0100 (CET), Mat Nieuwenhoven wrote:


Hi,

For radio amateur porposes, I'm building a 0.1 - 30 MHz sweeping oscillator
(not quite in one sweep) with the 74LS629. I'd like to take off the triangle
waveform which exists over the timing capacitor with a video amp like the
NE5539 and turn it into a reasonable sinus. I've tried the trick with two
antiparallel diodes (1N4148) fed via a series resistor by the triangle, but
that works only to about 1MHz, presumable because of the capacitance of the
diodes.
What is a good method of doing this? Any ideas are appreciated.


Thank you for all your responses. It seems that the diode/resistor thingy is
the most used solution, and I guess all those designeds know a lot more than
I do. I don't need a perfect sine, though. I'm currently using the TTL
blockwave out of the 629, and driving a filter or other tunable element via a
small C or not so small R, using a log detector you get quite usable results.
So basically anything that looks more like a sine than a block or triangle
would be fine. I get the feeling I'd be spending too much effort for a small
result, I might go with the mixer method instead. The disadvantage of that is
that it is difficult to get a fairly linear sweep.

I wonder what happens if I put the triangle into a log amp (I guess I'd need
to handle positive and negative side seperately and add them again. Surely
this would result in a more rounded-off triangle? Has anyone tried this
technique?
Click to expand...

Unless a pure analog approach can be justified, I'd settle for a DDS.
It does whatever sweep at almost whatever speed under processor control.
A flagship, the AD9854 has quadrature outputs and beside a linear sweep
capability can be updated every 10ns while producing up to 150MHz.
Spectral purity plus the achievable phasenoise is hard to beat.
The simpler and slower DDS cost from 15$ up.

Rene
--
Ing.Buero R.Tschaggelar - http://www.ibrtses.com
& commercial newsgroups - http://www.talkto.net
 
On Mon, 29 Mar 2004 21:28:56 +0100 (CET), Mat Nieuwenhoven wrote:

Hi,

For radio amateur porposes, I'm building a 0.1 - 30 MHz sweeping oscillator
(not quite in one sweep) with the 74LS629. I'd like to take off the triangle
waveform which exists over the timing capacitor with a video amp like the
NE5539 and turn it into a reasonable sinus. I've tried the trick with two
antiparallel diodes (1N4148) fed via a series resistor by the triangle, but
that works only to about 1MHz, presumable because of the capacitance of the
diodes.
What is a good method of doing this? Any ideas are appreciated.
Click to expand...
Thank you for all your responses. It seems that the diode/resistor thingy is
the most used solution, and I guess all those designeds know a lot more than
I do. I don't need a perfect sine, though. I'm currently using the TTL
blockwave out of the 629, and driving a filter or other tunable element via a
small C or not so small R, using a log detector you get quite usable results.
So basically anything that looks more like a sine than a block or triangle
would be fine. I get the feeling I'd be spending too much effort for a small
result, I might go with the mixer method instead. The disadvantage of that is
that it is difficult to get a fairly linear sweep.

I wonder what happens if I put the triangle into a log amp (I guess I'd need
to handle positive and negative side seperately and add them again. Surely
this would result in a more rounded-off triangle? Has anyone tried this
technique?

Thanks, Mat Nieuwenhoven
 
You must log in or register to reply here.

Welcome to EDABoard.com

Sponsor

Back
Top