Author Topic: Cheap low jitter(<10ps) low speed clock reference recommendation?  (Read 1335 times)

0 Members and 1 Guest are viewing this topic.

Offline andylithiaTopic starter

  • Newbie
  • Posts: 1
  • Country: cn
Greetings all,
Recently, I've got a new task of designing an extremely low jitter pulse generator running at <1MHz speed. A sub-10-picosecond-jitter clock is required. Now the only device I have that can work in such a low jitter is a Tektronix 11801 with 4GHz sampling heads, which is a fragile piece of equipment and quite difficult to use for clock generation (by which I mean using the internal clock outputs as a clock reference).
The other clock sources I have are HP8656B and HP3314A. HP8656B has a jitter of over 100ps at a 100MHz output, which is unacceptable. I haven't tested the HP3314A. Since HP didn't even bother listing it in the specifications, I'd assume it's bad.
Any recommendation on buying or DIY-ing a low jitter clock?
 

Offline Marco

  • Super Contributor
  • ***
  • Posts: 6714
  • Country: nl
Re: Cheap low jitter(<10ps) low speed clock reference recommendation?
« Reply #1 on: November 20, 2018, 01:07:03 pm »
Random LVDS oscillator? That requirement isn't exactly stringent.
 

Offline LapTop006

  • Supporter
  • ****
  • Posts: 467
  • Country: au
Re: Cheap low jitter(<10ps) low speed clock reference recommendation?
« Reply #2 on: November 21, 2018, 11:43:21 am »
Random LVDS oscillator? That requirement isn't exactly stringent.

Which reminds me of one of my current favourite pieces of kit, the Stanford Research CG635.

Not a pulse generator though.
 

Offline David Hess

  • Super Contributor
  • ***
  • Posts: 16600
  • Country: us
  • DavidH
Re: Cheap low jitter(<10ps) low speed clock reference recommendation?
« Reply #3 on: November 21, 2018, 02:27:02 pm »
Low jitter goes along with low phase noise so that is where I would start.  Take the output from a low phase noise quartz crystal oscillator and *carefully* condition it into a logic level signal.

Single ended logic like TTL and CMOS has poor (meaning no) power supply rejection so differential input logic like comparators and some ECL are especially useful in this application.  If you must use single ended logic, then the supply voltage and ground must be noise free.
 

Offline Leo Bodnar

  • Frequent Contributor
  • **
  • Posts: 803
  • Country: gb
Re: Cheap low jitter(<10ps) low speed clock reference recommendation?
« Reply #4 on: November 29, 2018, 10:47:40 pm »
Are you sure you are talking about hundreds of picoseconds jitter and not hundreds of femtoseconds, andylithia?

Second important question is how are you measuring it? RMS or peak-to-peak?  And over what frequency interval?  Typical measurement is RMS over 12kHz to 20MHz span.

I would say that almost any modern TCXO or OCXO has sub-ps jitter in the region of 100fs - 1ps.  100ps jitter sounds awful.

If your TCXO/OCXO has only phase noise data, use jitter calculator to get your jitter estimation.

Few examples of current state of the art:
run of the mill TCXO - https://www.digikey.co.uk/product-detail/en/ecs-inc/ETXO-H33CL-100.000/XC2134-ND/5875502  1ps jitter

[reasonably expensive] modern 100MHz OCXO - https://abracon.com/Precisiontiming/AOCJY7TQ.pdf  20fs  jitter

Leo
« Last Edit: November 29, 2018, 10:52:58 pm by Leo Bodnar »
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf