[time-nuts] FE-5680A Results so far

ws at Yahoo warrensjmail-one at yahoo.com
Wed Dec 28 14:05:10 EST 2011

Chris posted:
> So how to test it?  anyone have any creative ideas?

That's easy, get a couple more of them.
The using a scope, trigger on one unit and watch the phase drift of the 
With most basic scopes you can detect sub ns phase changes of 10MHz signals 
this way, which allows you to see a 1e-11 frequency difference in a couple 
minutes. (1ns/sec change = 1e-9 freq offset)

Or you could use the basic x y display and count how many minutes it takes 
to complete one cycle.
That will allow you to see the same resolution after a few hours. (a 100ns 
cycle per sec = 1e-7 freq offset)



[time-nuts] FE-5680A Results so far
Chris Albertson albertson.chris at gmail.com
Wed Dec 28 18:21:54 UTC 2011

Just got my fe5680 yesterday.   Seems to work but the fe5680 is the most
accurate reference I have.  So how to test it?   I plan to buy a
Thunderbolt but until then, anyone have any creative ideas?

I have an HP5328A with unknown calibration.  It reads 10,000,026 Hz
I have an FCC 1 kit I assembled a few years ago and it says 9,999,993 Hz.
(I can make the FCC1 read exactly 10MHz if I place it on top of the hot
FE5680, the FCC1 is very temperature sensitive.
My old Tek scope says the period is "about" 0.1 uS and the sine wave looks
pretty good.

BTW, 5 VDC is 76.5 mA with RS232 not connected.    15 VDC is 1.5 ~ 0.8 A.
 The unit does get warm.  A 3/16 x 5 x 16 inch aluminum plate worked is a
"good enough" heat sink.

Going to have to get a Thunderbolt.  I have an Oncore UT+ the provides a
good pulse per second but that will not help

Chris Albertson
Redondo Beach, California 

More information about the time-nuts mailing list