[time-nuts] Rubidium oscillator controlled clock

ulmann at vaxman.de ulmann at vaxman.de
Mon Jul 12 17:57:14 UTC 2010


 Hello Magnus, hello Ulrich - 
thank you very much for your replies and your suggestions! I will dig
deeper into that issue (I begin to get fascinated about precise timing
- what a change for someone who once built amplifiers for measuring
signals with periods down to 1/100 Hz :-) ). It will take some time
(some weeks, since I am preparing for a conference etc.) until I will
find time again to improve my oscillator, but I will let you know when
I have decided on a circuit design and ask you for advice before
actually building it.
 All the best - have a great day - Bernd. :-)

>Message: 3
>Date: Mon, 12 Jul 2010 14:18:35 +0200
>From: Magnus Danielson <magnus at rubidium.dyndns.org>
>Subject: Re: [time-nuts] Rubidium oscillator controlled clock
>To: time-nuts at febo.com
>Message-ID: <4C3B081B.3090101 at rubidium.dyndns.org>
>Content-Type: text/plain; charset=ISO-8859-1; format=flowed
>
>Bernd,
>
>On 07/12/2010 02:05 PM, ulmann at vaxman.de wrote:
>>   Hello Ulrich -
>> thank you very much for kind reply - this is a great suggestion. Would you
>> recommend simple 74xx74 flipflops driven by the original 10 MHz signal
>> delivered by the rubidium oscillator for resynching?
>
>That should do the trick. Even if I doubt the gain would be significant, 
>you can use both the DFFs in there in series. It is the common way to 
>reduce the effect of unsynchronized signal into DFFs as there may be 
>meta-stability, so using two DFFs in series helps reducing the added 
>noise if the first DFF goes meta-stable. The '74 should be protected to 
>some degree if I recall things correctly, but it's there and free... so 
>why not?
>
>Cheers,
>Magnus
>
>------------------------------
>
>Message: 4
>Date: Mon, 12 Jul 2010 14:47:29 +0200
>From: "Ulrich Bangert" <df6jb at ulrich-bangert.de>
>Subject: Re: [time-nuts] Rubidium oscillator controlled clock
>To: "'Discussion of precise time and frequency measurement'"
>	<time-nuts at febo.com>
>Message-ID: <87AA1E4D813C493A9F325307AC987648 at athlon>
>Content-Type: text/plain;	charset="us-ascii"
>
>Bernd,
>
>> Would you recommend simple 74xx74 flipflops 
>> driven by the original 10 MHz signal delivered by the 
>> rubidium oscillator for resynching? 
>
>In principle, yes. However not directly clocked by the rubidium itself but
>from a clean ttl signal that has been made out of the LPRO's sine. There has
>been a long discussion here about the "howto" of this. As a good starting
>point I suggest this:
>
>http://www.ko4bb.com/~bruce/CLKSHPR.html
>
>While the ADCMP600 are available well in Germany (for example from FARNELL)
>the MSOP enclosure is very tricky to solder and for a bit less in frequency
>performance a LT1016 will do it for you too.
>
>Best regards
>Ulrich Bangert
>



More information about the time-nuts mailing list