[time-nuts] ks... answers The emails are getting long

Bob Camp kb8tq at n1k.org
Tue Nov 18 20:47:48 EST 2014


Hi

Yes Bob is back on his silly “don’t go overboard” rant ….

We often get excited about things that matter at the 1x10^-15 level when looking at what is a 1x10^-12 level problem. One good example is distribution amplifiers. Another are things like multipliers. What ever horrid thing they are doing in the REF-0 box to get 10 MHz, it works reasonably well from an ADEV standpoint past 10 seconds. They made a decision that what ever it is was good enough for their needs. Past 10 seconds it’s as good as anything you could do compared to the 5 MHz source it’s self. At 1 second, it degrades the 5 MHz slightly. Phase noise wise, it’s utter garbage anyplace you look. It’s not as bad as a telecom Rb, but it’s certainly in that league. 

Keeping a goal in mind is a really good idea on any project. Coming up with a “everything for everybody” solution is wonderful. HP could not afford to do that at a > $50K sell price on a 5071A. Tradeoff against a specific target is the realistic way to do even a TimeNut effort. ADEV past 1 second does not require -170 dbc / Hz phase noise at a 100 Hz offset from carrier. 

High Q is indeed an issue, so are bandpass as compared to lowpass filters. Bandwidth does matter. Stable parts (no weird ferrite) are part of the proper approach. What ever you do, it must be done with thought and with care. You also need to test the result. The effort for any of this needs to go into the little things (like testing) rather than into this or that magic circuit. Magic circuit or not, you still need to verify the result ….

Bob 

> On Nov 18, 2014, at 8:26 PM, Magnus Danielson <magnus at rubidium.dyndns.org> wrote:
> 
> Indeed. I was just pointing out that high-Q approaches might not always works as you would think, but there is a way around it if you need.
> 
> Cheers,
> Magnus
> 
> On 11/19/2014 02:17 AM, Bob Camp wrote:
>> Hi
>> 
>> For the limited temperature swings that most labs have, and the sort of ADEV that the OCXO has, a Q up to 10 can be tolerated pretty well. Keeping the parts count down to two to four coils  and a few capacitors lets you do it without a lot of complex design tools. They did it with charts and slide rules back in the 1940’s. You can do it with a few clicks on a web based calculator app these days.
>> 
>> Bob
>> 
>>> On Nov 18, 2014, at 8:10 PM, Magnus Danielson <magnus at rubidium.dyndns.org> wrote:
>>> 
>>> Hi,
>>> 
>>> The low Q values has the benefit of being very phase stable.
>>> 
>>> Low Q dips can be put in to null out particular overtones, LCR-series loading the signal to ground at suitable place. Also very traditional 1930-1950 era.
>>> 
>>> Cheers,
>>> Magnus
>>> 
>>> On 11/19/2014 01:30 AM, Bob Camp wrote:
>>>> Hi
>>>> 
>>>> Single stage NPN 2N5179 or 2N918 or 2N2222  or … biased to generate a collector current that looks like a “raised cosine” (see your 1952 RF transmitter textbook). Tune the collector to the desired output. Set the Q to something like 10 or so. Linear “class A” stage after that. Tune the collector with another Q 3 to 10 tank. Match it to the load with a PI or a Tee.
>>>> 
>>>> Bob
>>>> 
>>>>> On Nov 18, 2014, at 6:21 PM, Doug Ronald <doug at dougronald.com> wrote:
>>>>> 
>>>>> I doubled my Austron 1250A from 5 MHz to 10 MHz with 2 NPNs fed in push-pull and output in parallel with a tuned circuit. Before buffering it with an LH0063, I fed the signal through a 10 MHz crystal, purchased at a local surplus electronics store for $0.95. The 2nd harmonic is about 70 dB down.
>>>>> 
>>>>> -Doug, AE6SY
>>>>> 
>>>>> -----Original Message-----
>>>>> From: time-nuts [mailto:time-nuts-bounces at febo.com] On Behalf Of Bob Camp
>>>>> Sent: Tuesday, November 18, 2014 3:11 PM
>>>>> To: Discussion of precise time and frequency measurement
>>>>> Subject: Re: [time-nuts] ks... answers The emails are getting long
>>>>> 
>>>>> HI
>>>>> 
>>>>> Actually there are three different approaches to the “double 5 MHz” question.
>>>>> 
>>>>> Roughly 99.99% of all doubler / tripler OCXO’s out there do the trick with a simple single transistor stage and a  tuned tank. Follow it up with a tuned single transistor output amp. Cheap, easy, not very fancy, gets people mad when mentioned. It works plenty good enough. The basic design approach dates back to tube based multipliers done in the 1920’s. Just about any transmitter design textbook from 1930 through 1960 has design charts and tube based examples.
>>>>> 
>>>>> Bob
>>>>> 
>>>>>> On Nov 18, 2014, at 4:26 PM, paul swed <paulswedb at gmail.com> wrote:
>>>>>> 
>>>>>> Don
>>>>>> The two units work together by a buggy cable from what I have heard so far.
>>>>>> The Ref0 is the master that drives stuff it connects to REF1 that
>>>>>> disciplines REF0.
>>>>>> If GPS goes away or the ref1 it all keeps ticking. This is been teh
>>>>>> system design for about 20 years.
>>>>>> 
>>>>>> Many of us have the ref1 only and the internal osc is 5 Mhz so thats
>>>>>> why you have seen discussions on multipliers here. Tap the 5 Mhz
>>>>>> double it filter it and buffer it. there seem to be 2 schools of
>>>>>> thoughts on the process. Balanced mixer or Wenzel.
>>>>>> I would have hacked the answer already. But I need to keep my ref1
>>>>>> running so that it ages in. ASI will refund your money if bad. But only in 30 days.
>>>>>> I do like what I see from the KS-... Can never remember the number.
>>>>>> Regards
>>>>>> Paul
>>>>>> WB8TSL
>>>>>> _______________________________________________
>>>>>> time-nuts mailing list -- time-nuts at febo.com To unsubscribe, go to
>>>>>> https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
>>>>>> and follow the instructions there.
>>>>> 
>>>>> _______________________________________________
>>>>> time-nuts mailing list -- time-nuts at febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
>>>>> and follow the instructions there.
>>>>> 
>>>>> _______________________________________________
>>>>> time-nuts mailing list -- time-nuts at febo.com
>>>>> To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
>>>>> and follow the instructions there.
>>>> 
>>>> _______________________________________________
>>>> time-nuts mailing list -- time-nuts at febo.com
>>>> To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
>>>> and follow the instructions there.
>>>> 
>>> _______________________________________________
>>> time-nuts mailing list -- time-nuts at febo.com
>>> To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
>>> and follow the instructions there.
>> 
>> _______________________________________________
>> time-nuts mailing list -- time-nuts at febo.com
>> To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
>> and follow the instructions there.
>> 
> _______________________________________________
> time-nuts mailing list -- time-nuts at febo.com
> To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
> and follow the instructions there.



More information about the time-nuts mailing list