Author Topic: 500khz 50% square wave!  (Read 9288 times)

0 Members and 1 Guest are viewing this topic.

Offline GregoryTopic starter

  • Regular Contributor
  • *
  • Posts: 59
  • Country: br
    • All Electronics Channel
500khz 50% square wave!
« on: March 11, 2013, 06:32:38 am »
Hello guys! I need to generate a perfect 50% square wave with short term stability (i will use only 8 cycles of it). I tried with a NE555, ir works great for lower frequencys but near 500khz it starts to get crap with not a 50% duty cycle. Maybe I need to use a CMOS 555 or maybe some other chip? thanks!

Offline midasgossye

  • Contributor
  • Posts: 32
  • Country: be
    • Bike2Charge
Re: 500khz 50% square wave!
« Reply #1 on: March 11, 2013, 06:40:59 am »
Wich values of caps and resistors did you use?
 

Offline amspire

  • Super Contributor
  • ***
  • Posts: 3802
  • Country: au
Re: 500khz 50% square wave!
« Reply #2 on: March 11, 2013, 06:51:32 am »
Hello guys! I need to generate a perfect 50% square wave with short term stability (i will use only 8 cycles of it). I tried with a NE555, ir works great for lower frequencys but near 500khz it starts to get crap with not a 50% duty cycle. Maybe I need to use a CMOS 555 or maybe some other chip? thanks!
The way you get a perfect 50% duty cycle is to generate 1MHz and add a divide-by-2 flip flop IC. It thn does not matter what the oscillator duty cycle is, the output of the flip-flop will be perfect.
 

Offline grumpydoc

  • Super Contributor
  • ***
  • Posts: 2905
  • Country: gb
Re: 500khz 50% square wave!
« Reply #3 on: March 11, 2013, 08:04:58 am »
Why does it have to be "perfect"?
 

Offline sub

  • Regular Contributor
  • *
  • Posts: 107
  • Country: au
Re: 500khz 50% square wave!
« Reply #4 on: March 11, 2013, 08:41:24 am »
My suggestion would be a Pierce oscillator using a 1MHz crystal followed by a T (or JK) flip-flop, assuming you need not a significant range of tuning.
 

Offline Eonir

  • Contributor
  • Posts: 15
  • Country: se
Re: 500khz 50% square wave!
« Reply #5 on: March 11, 2013, 09:16:44 am »
The 555 is completely out of question. It's an outdated design. The maximum possible frequency might be something like 300kHz, but the duty cycle will be far from 50%.

One of the easiest ways of generating a very reliable square wave is getting a crystal and dividing it down to your desired frequency using something like a 4060 chip. It's basically a chain of cascaded D flip flops configured as frequency dividers. Its usual applications are counters, but I found them very useful for generating various frequencies for e.g. ultrasonic and optical sensors modulated with 38kHz or what not.

Get a 16MHz crystal and divide it down 32 times, and you'll get a nice clean 500kHz square wave. Moderns 4000 series ICs work fine up to 20MHz without a hitch.

You could of course use a 1MHz crystal and a single flip flop - but these come in quad sets in 16DIP packages anyway, so you're not going to save any space. The 4060 chip is small and extremely cheap, so using separate flip flop is not going to save you much money.
« Last Edit: March 11, 2013, 09:18:48 am by Eonir »
 

Offline BravoV

  • Super Contributor
  • ***
  • Posts: 7547
  • Country: 00
  • +++ ATH1
Re: 500khz 50% square wave!
« Reply #6 on: March 11, 2013, 09:25:58 am »
The 555 is completely out of question. It's an outdated design. The maximum possible frequency might be something like 300kHz, but the duty cycle will be far from 50%.

Ever heard of 555 "cmos" version ?  ;)


One of the easiest ways of generating a very reliable square wave is getting a crystal and dividing it down to your desired frequency using something like a 4060 chip. It's basically a chain of cascaded D flip flops configured as frequency dividers. Its usual applications are counters, but I found them very useful for generating various frequencies for e.g. ultrasonic and optical sensors modulated with 38kHz or what not.

Get a 16MHz crystal and divide it down 32 times, and you'll get a nice clean 500kHz square wave. Moderns 4000 series ICs work fine up to 20MHz without a hitch.

You could of course use a 1MHz crystal and a single flip flop - but these come in quad sets in 16DIP packages anyway, so you're not going to save any space. The 4060 chip is small and extremely cheap, so using separate flip flop is not going to save you much money.

Naah... read the 1st post  again, the op is requesting "specifically" an exact 50% duty cycle while only for "near" 500 Khz, definitely xtal accuracy is not needed.

A 2 ICs solution like a simple "cmos" 555 (again, not your grand dad bipolar 555) + 4027 and few cheapo discretes like amspire's suggestion above should more than enough. Ok, maybe using mica cap for better frequency stability.

Thats all, a dirt cheap solution while fits nicely with requirement.
« Last Edit: March 11, 2013, 09:27:43 am by BravoV »
 

Online fcb

  • Super Contributor
  • ***
  • Posts: 2117
  • Country: gb
  • Test instrument designer/G1YWC
    • Electron Plus
Re: 500khz 50% square wave!
« Reply #7 on: March 11, 2013, 09:36:37 am »
4047B, one resistor and one capacitor.

The 4047 is brill. It's got the oscillator, divide-by-2, reset logic all done.  All you have to do is count your 8 pulses out (which could be done with another single logic IC).



https://electron.plus Power Analysers, VI Signature Testers, Voltage References, Picoammeters, Curve Tracers.
 

Offline BravoV

  • Super Contributor
  • ***
  • Posts: 7547
  • Country: 00
  • +++ ATH1
Re: 500khz 50% square wave!
« Reply #8 on: March 11, 2013, 10:17:12 am »
4047B, one resistor and one capacitor.

The 4047 is brill. It's got the oscillator, divide-by-2, reset logic all done.  All you have to do is count your 8 pulses out (which could be done with another single logic IC).

If its workable, this is even better.  :-+

Btw does 4047 capable of running at 1 Mhz reliably ?

Online fcb

  • Super Contributor
  • ***
  • Posts: 2117
  • Country: gb
  • Test instrument designer/G1YWC
    • Electron Plus
Re: 500khz 50% square wave!
« Reply #9 on: March 11, 2013, 10:35:14 am »
Yeah, it should be fine. The Philips HEF process parts might have lower consumption at 1MHz than the cooking parts.

That's the 'D' bit of R&D though :)
https://electron.plus Power Analysers, VI Signature Testers, Voltage References, Picoammeters, Curve Tracers.
 

Offline GregoryTopic starter

  • Regular Contributor
  • *
  • Posts: 59
  • Country: br
    • All Electronics Channel
Re: 500khz 50% square wave!
« Reply #10 on: March 11, 2013, 05:07:12 pm »
ALD555 is one CMOS 555 that can run higher frequency. Maybe I don't need a really PERFECT 50%, but it will be a starting generator to a voltage-fed phase-shifted inverter that I'm desing for a new type os tesla coil I'm creating. I really need 50% to not inject DC components on my output LC tank. A counter will run 4 or 8 cycles of it and will change to lock on regenerative current feedback from the output LC. thanks!

Offline Jay_Diddy_B

  • Super Contributor
  • ***
  • Posts: 2731
  • Country: ca
Re: 500khz 50% square wave!
« Reply #11 on: March 11, 2013, 05:29:52 pm »
Hi,

You can use the LTC6990. This chip has a silicon oscillator and a divider in a SOT23 package.

Link: http://www.linear.com/product/LTC6990

Schematic:


I have attached a zip file with the LTspice model.

Jay_Diddy_B

 

Offline GregoryTopic starter

  • Regular Contributor
  • *
  • Posts: 59
  • Country: br
    • All Electronics Channel
Re: 500khz 50% square wave!
« Reply #12 on: March 11, 2013, 06:52:51 pm »
ohh, very thanks. I was thinking to use a simple 74hc14 oscilator and a flip-flop to generate the 50% because I have this two components left in the circuit.


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf