DIY square wave oscillator

[zigzagflux]

Anyone have a decent quality design for a square wave generator, running from 1 kHz to 2.5 MHz ? Of all the designs I've looked at, the VCO's out there are way up in the Megahertz range, no where close to the range I am looking for. 555 timer based circuits do not work well at 2.5MHz, so that's out.

Ideally, I would like to build my own circuit board containing a simple square wave generator. There will be an additional circuit following this square wave, of course, but that portion I have figured out. Entire assembly will mount inside an enclosure, producing 1.21 gigawatts. :-O Voltage rails are flexible, but 5V would be ideal. I would buy a dedicated IC if I could find one appropriate for the range of operation I am looking for.

Other requirements I can think of is 50% duty cycle. I can use a divide by 2 circuit to guarantee this, but then I will need a 2k to 5M range of operation. Stable frequency. Simple control, pot or encoder. Range switches are not too desirable, but I can deal with it if necessary.

Thanks for the ideas.

 

[Steerpike]

How about a Wien bridge oscillator based on any general purpose wideband op amp (you probably will require several ranges, achieved by switching the timing capacitors, otherwise the frequency control pot will behave in a very coarse manner);
followd by a CMOS or TTL Schmitt buffer to turn it into a square wave.

 

[wakibaki]

There are a few threads out there about square wave generators, it depends what frequency and stability you want.

I'd be tempted to build a DDS, convert the sine to square (you can probably feed it direct into a bit of ECL) and then divide that down to get in the range you want. With a synchronous counter you get octaves at every tap, so you only need an octave of tuning range. No analog design involved, you can get a few DDS boards from ham radio sources.

w

 

[thespeakerguy]

There's many ways

One difficulty is resolution over such a wide range. I would make a basic ring oscillator with 74HC14, and use a potentiometer to adjust frequency across a 16x or so range, perhaps 150kHz to 2.5MHz.

Use an MC14040 12 bit divider to provide 16x ranges or so by tapping Q4 (/16) Q8 (/256) and Q12 (/4096). THis puts the minimum frequency around 36Hz.

A SP4T switch selects the range you want

 

[bwaslo]

The VCO section of the cheap 74HC4046 PLL chips makes a nice square wave generator (and it can be voltage controlled, you can use a pot to adjust frequency). Can cover a very wide range, but for very low frequencies here's a second vote for the suggestion made above-- to use a bank of frequency dividers. That gives a guaranteed 50% duty cycle, and you need to only calibrate frequency once, since the divide by 2^N is precise.

 

[wrenchone]

Ordinary bipolar 555 timers don't work to well at high frequency, but you may want to try the Texas Instruments TLC555 - it's much, much faster.

 

[Conrad Hoffman]

Look at the Linear Technology app notes. There are some very wide range VCO circuits there. I think you can also still get Exar chips- they also have some nice oscillators that will probably do what you want.

 

[Pjotr]

It is out of production but maybe you can get a hand on a MAX038

http://www.maxim-ic.com/quick_view2.cfm/qv_pk/1257

 

[zigzagflux]

Thanks guys. I agree the divide by N counter is probably the best solution. I'll pursue a dedicated IC capable of a few MHz, and divide down from there.

 

[quadtech]

Why not do it using one of those small microcontrollers?
Many of them go to 20MHz clock speeds or more. It
seems your desired freq range could be possible. You
could hook up a rotary encoder and LCD display too. Or you can use
a analog input to read a pot to control the freq.

 

[astouffer]

How about 1Khz to 33Mhz?

http://www.linear.com/pc/productDetail.jsp?navId=H0,C1,C1010,C1096,P2069

Get one of those little SOT-23 adapter boards and you are all set.