[time-nuts] WWVB Chronverter update progress

Bob kb8tq kb8tq at n1k.org
Mon Aug 27 11:57:48 EDT 2018


Hi

The temperature spec on a typical 60 KHz crystal is in the 100’s of ppm
range. The temperature coefficient can hit multiple ppm / C at fairly rational
temperatures.  That all adds on top of the set tolerance of the crystal. Simply 
to keep it passing signal while the room changes temperature would require a 
bandwidth into the 10’s of Hz. 

For reasonable loss in the filter you would want a Ql/Qu ratio of at least ten … 
again back into the 10’s of Hz range. 

The gotcha of course is that some of the tolerance stacks up all in one direction.
You might pass 20 Hz high just fine, but nuke 2 Hz low on this or that device. 

Bob

> On Aug 27, 2018, at 11:12 AM, Tim Shoppa <tshoppa at gmail.com> wrote:
> 
> The consumer WWVB wall clocks use a single 60kHz crystal as a front end
> filter (not as an oscillator).
> 
> Unloaded Q of a small tuning fork crystal is often 30,000 or so. (You can
> actually observe this order of magnitude when a 32kHz crystal used in an
> oscillator - remove power and the crystal continues vibrating for a couple
> seconds before dying off.).
> 
> This gives expected bandwidth of WWVB front end crystal filter as a more
> than a couple Hz. Well, that should be expected too, because they are
> demodulating a time code with a 1 bit per second data rate using amplitude
> modulated pulses requiring differentiating 0.2s pulses from 0.8s pulses.
> And anyway these tuning fork crystals are only specified to 20 or 30 ppm.
> 
> It's only a single pole filter so if a WWVB emulator were 10Hz off and
> "loud", I'm sure the clock will respond just fine.
> 
> Now, if this is a fancy-pants consumer WWVB wall clock that also expects
> the phase shift keying (not just amplitude keying) then I have no idea how
> it'll respond to just plain amplitude keying for the time code. You might
> think the WWVB BPSK chip will respond to loud and clear amplitude
> modulation but I have no experience with this. Nor do I even know if a
> single consumer clock was ever sold with BPSK ability.
> 
> Tim N3QE
> 
> On Mon, Aug 27, 2018 at 10:40 AM Bob kb8tq <kb8tq at n1k.org> wrote:
> 
>> Hi
>> 
>> One Hertz out of a million Hertz is one part out of a million parts.
>> 
>> One Hertz out of 100,000 Hertz is ten parts per million.
>> 
>> One Hz out of 10KHz is 100 PPM.
>> 
>> One Hertz out of 60,000 Hertz is 16.667 parts per million.
>> 
>> If you are running into a device that swings it’s OCXO to lock to the
>> incoming signal,
>> it’s unlikely that the OCXO has a very wide tuning range. Numbers in the
>> sub half ppm
>> range are not uncommon.
>> 
>> So - what are we trying to do? If it’s just feeding the cheap wall clock,
>> they aren’t phase locking
>> an OCXO. The vast majority of them just want to do AM reception. If this
>> is going to feed
>> TimeNuts gear for various purposes then indeed locking up an OCXO may be
>> part of
>> the deal.
>> 
>> Bob
>> 
>>> On Aug 27, 2018, at 10:19 AM, paul swed <paulswedb at gmail.com> wrote:
>>> 
>>> OK you sort of lost me here. 16 PPM would be great I might find something
>>> that works without a power sucking oven.
>>> But Bob mentioned online calculators and there are PPM calculators. Mr
>>> Google seems to know all.
>>> That said the lock range of any of the professional receivers is tight.
>> +/-
>>> .6 Hz. Used a HP 3336 gen and moved up and down by .1 Hz to figure out
>> the
>>> lock range. 3336 locked to GPSDO.
>>> I may have used the calculator wrong but at 60 KHz it seems like its .05
>>> PPM or was it .5ppm. Anyhow don't like the answer.
>>> I have numbers of 6 MHz xtals and also clock gens. The clock gens were
>> off
>>> by 5Hz when divided down and out of the lock range. It is interesting
>> that
>>> all 3 were of by different amounts but all negative. These are not TCXOs.
>>> A home brew xtal oscillator will not have temperature compensation.
>>> So this has presented a bit more of a challenge then hoped for.
>>> Regards
>>> Paul
>>> WB8TSL
>>> 
>>> 
>>> On Mon, Aug 27, 2018 at 7:29 AM, Mike Feher <mfeher at eozinc.com> wrote:
>>> 
>>>> Simple error. It is 1.667. 73 - Mike
>>>> 
>>>> 
>>>> 
>>>> Mike B. Feher, N4FS
>>>> 
>>>> 89 Arnold Blvd.
>>>> 
>>>> Howell NJ 07731
>>>> 
>>>> 848-245-9115
>>>> 
>>>> 
>>>> 
>>>> -----Original Message-----
>>>> From: time-nuts <time-nuts-bounces at lists.febo.com> On Behalf Of Bob
>> kb8tq
>>>> Sent: Sunday, August 26, 2018 6:34 PM
>>>> To: Discussion of precise time and frequency measurement <
>>>> time-nuts at lists.febo.com>
>>>> Subject: Re: [time-nuts] WWVB Chronverter update progress
>>>> 
>>>> 
>>>> 
>>>> Hi
>>>> 
>>>> 
>>>> 
>>>> Well …. 1Hz at 60KHz is …. errr … (off to Google Calc …) …. 16.67 ppm.
>>>> That’s a pretty major lock range for a TCXO. It’s likely > 10X what an
>> OCXO
>>>> will do. What the bare crystal in a watch will do … who knows.
>>>> 
>>>> 
>>>> 
>>>> Bob
>>>> 
>>>> 
>>>> 
>>>>> On Aug 26, 2018, at 5:52 PM, paul swed < <mailto:paulswedb at gmail.com>
>>>> paulswedb at gmail.com> wrote:
>>>> 
>>>>> 
>>>> 
>>>>> Looks like the signal has to be with in 1 hz. I new it was tight.
>>>> 
>>>>> Need a bit more experimentation across the several receivers though
>>>> 
>>>>> they all use the same approach with a sharp Xtal filter.
>>>> 
>>>>> 
>>>> 
>>>>> 
>>>> 
>>>>> On Sat, Aug 25, 2018 at 8:24 PM, paul swed < <mailto:
>> paulswedb at gmail.com>
>>>> paulswedb at gmail.com> wrote:
>>>> 
>>>>> 
>>>> 
>>>>>> My mistake the chronverter uses a pic 12f1840.
>>>> 
>>>>>> 
>>>> 
>>>>>> On Sat, Aug 25, 2018 at 8:23 PM, paul swed < <mailto:
>>>> paulswedb at gmail.com> paulswedb at gmail.com> wrote:
>>>> 
>>>>>> 
>>>> 
>>>>>>> Have to look at what oscillators and crystals I have then a divide.
>>>> 
>>>>>>> Thats old style and most likely the junk box has everything.
>>>> 
>>>>>>> Have to look in my box of tricks as I built several chains for the
>>>> 
>>>>>>> de-psk-er experiments.
>>>> 
>>>>>>> Have to agree on the micro the chronverter uses a pIC 12F64 as I
>>>> 
>>>>>>> recall and it has quite a nice set of dividers in it. Just not an
>>>> 
>>>>>>> external crystal input so a better oscillator can be used. Getting
>>>> lazy these days.
>>>> 
>>>>>>> I have some number of tcxos from a company called vectron. Several
>>>> 
>>>>>>> cards of each. Can't imagine how I came by those.
>>>> 
>>>>>>> :-) Believe the lowest freq was 12 MHz.
>>>> 
>>>>>>> Regards
>>>> 
>>>>>>> Paul.
>>>> 
>>>>>>> 
>>>> 
>>>>>>> 
>>>> 
>>>>>>> On Sat, Aug 25, 2018 at 8:07 PM, Bob kb8tq < <mailto:kb8tq at n1k.org>
>>>> kb8tq at n1k.org> wrote:
>>>> 
>>>>>>> 
>>>> 
>>>>>>>> Hi
>>>> 
>>>>>>>> 
>>>> 
>>>>>>>> A pretty simple modest Q tuned circuit on the output of a gate will
>>>> 
>>>>>>>> turn the square wave into a sine. Not much to it since there are a
>>>> 
>>>>>>>> lot of power supply components out there that work fine at 60 KHz.
>>>> 
>>>>>>>> In order to get a stable 60KHz, a divider is the way to go. I’d
>>>> 
>>>>>>>> probably use one that’s built into an MCU as part of a timer or
>>>> 
>>>>>>>> something similar.
>>>> 
>>>>>>>> 
>>>> 
>>>>>>>> Bob
>>>> 
>>>>>>>> 
>>>> 
>>>>>>>>> On Aug 25, 2018, at 7:00 PM, paul swed < <mailto:
>> paulswedb at gmail.com>
>>>> paulswedb at gmail.com> wrote:
>>>> 
>>>>>>>>> 
>>>> 
>>>>>>>>> Bob The generator sig is 60 Khz sinewave.but as a I build up a
>>>> 
>>>>>>>>> divider chain and such your idea makes great sense.
>>>> 
>>>>>>>>> Then filter to 60 KHz.
>>>> 
>>>>>>>>> But I see the issue with the Chronverter carrier its 120-150 Hz
>> high.
>>>> 
>>>>>>>>> Bounces around a fair amount.
>>>> 
>>>>>>>>> There could be a secondary issue in that the carrier may not
>>>> 
>>>>>>>>> actually
>>>> 
>>>>>>>> be
>>>> 
>>>>>>>>> phase stable.
>>>> 
>>>>>>>>> It really doesn't matter as for the Truetimes and Spectracoms its
>>>> 
>>>>>>>> simply
>>>> 
>>>>>>>>> not good enough.
>>>> 
>>>>>>>>> I still need to try the cheapy clocks as that is what it was
>>>> 
>>>>>>>>> intended
>>>> 
>>>>>>>> for.
>>>> 
>>>>>>>>> Not throwing stones as so much of what it does saves me a bunch of
>>>> 
>>>>>>>> coding
>>>> 
>>>>>>>>> and doing exactly what Dave did.
>>>> 
>>>>>>>>> Regards
>>>> 
>>>>>>>>> Paul
>>>> 
>>>>>>>>> WB8TSL
>>>> 
>>>>>>>>> 
>>>> 
>>>>>>>>> On Sat, Aug 25, 2018 at 5:31 PM, Bob kb8tq < <mailto:kb8tq at n1k.org
>>> 
>>>> kb8tq at n1k.org> wrote:
>>>> 
>>>>>>>>> 
>>>> 
>>>>>>>>>> Hi
>>>> 
>>>>>>>>>> 
>>>> 
>>>>>>>>>> You can do the 14 db deep modulation with a tristate gate and a
>>>> 
>>>>>>>>>> pair
>>>> 
>>>>>>>> of
>>>> 
>>>>>>>>>> resistors.
>>>> 
>>>>>>>>>> Ground the input and feed the “modulation” signal to the tristate
>>>> 
>>>>>>>> control.
>>>> 
>>>>>>>>>> 
>>>> 
>>>>>>>>>> Bob
>>>> 
>>>>>>>>>> 
>>>> 
>>>>>>>>>>> On Aug 25, 2018, at 4:35 PM, paul swed < <mailto:
>>>> paulswedb at gmail.com> paulswedb at gmail.com> wrote:
>>>> 
>>>>>>>>>>> 
>>>> 
>>>>>>>>>>> From the earlier threads OOK modulation does not work for high
>>>> 
>>>>>>>>>>> end
>>>> 
>>>>>>>> clocks
>>>> 
>>>>>>>>>>> like spectracoms and truetimes. Have not tried it on the cheapy
>>>> 
>>>>>>>> clocks
>>>> 
>>>>>>>>>> yet.
>>>> 
>>>>>>>>>>> 
>>>> 
>>>>>>>>>>> I added an external modulator. A dg419 analog switch and then
>>>> 
>>>>>>>>>>> with a
>>>> 
>>>>>>>> few
>>>> 
>>>>>>>>>>> resistors added DC offset and anttenuation so that the carrier
>>>> 
>>>>>>>>>>> drops
>>>> 
>>>>>>>> by
>>>> 
>>>>>>>>>> -14
>>>> 
>>>>>>>>>>> db per the wwvb spec. The logic control is driven from the
>>>> 
>>>>>>>> chronverter
>>>> 
>>>>>>>>>> and
>>>> 
>>>>>>>>>>> it totally works. Spectracom Netclock 2 locked up in the normal
>>>> time.
>>>> 
>>>>>>>>>>> There are far better modern single supply analog gates that will
>>>> 
>>>>>>>> work.
>>>> 
>>>>>>>>>> Just
>>>> 
>>>>>>>>>>> did not have any.
>>>> 
>>>>>>>>>>> 
>>>> 
>>>>>>>>>>> The 60 KHz carrier is supplied by a fluke 6060 generator. This
>>>> 
>>>>>>>> eliminates
>>>> 
>>>>>>>>>>> any chance of being off frequency and is not the long term
>> answer.
>>>> 
>>>>>>>>>>> 
>>>> 
>>>>>>>>>>> The chronverter actually uses a dallas semiconductor clock chip
>>>> 
>>>>>>>>>>> thats reasonably accurate.(dangerous to say on time-nuts) So
>>>> 
>>>>>>>>>>> that even if
>>>> 
>>>>>>>> you
>>>> 
>>>>>>>>>> do
>>>> 
>>>>>>>>>>> not add the GPS receiver it will run quite some time correctly
>>>> 
>>>>>>>>>>> after setting the time.
>>>> 
>>>>>>>>>>> 
>>>> 
>>>>>>>>>>> Chronverter draws 6 ma from a 5 V supply. I always like low
>> power.
>>>> 
>>>>>>>>>>> 
>>>> 
>>>>>>>>>>> All in all a very nice answer to what do we do if WWVB goes away.
>>>> 
>>>>>>>>>>> As mentioned it handles timezones and DST and you can change DST
>>>> 
>>>>>>>>>>> if
>>>> 
>>>>>>>> those
>>>> 
>>>>>>>>>>> silly politicians screw with it again.
>>>> 
>>>>>>>>>>> 
>>>> 
>>>>>>>>>>> Speculation
>>>> 
>>>>>>>>>>> The chronverter actually puts out carriers for the popular LF
>>>> 
>>>>>>>> stations
>>>> 
>>>>>>>>>>> along with the time codes for them.
>>>> 
>>>>>>>>>>> 
>>>> 
>>>>>>>>>>> Though the modulator allows quality clocks to lock I speculate
>>>> 
>>>>>>>>>>> the frequency from the chronverter is not tight enough. All of
>>>> 
>>>>>>>>>>> these
>>>> 
>>>>>>>> clocks
>>>> 
>>>>>>>>>>> have very sharp crystal filters in them. 10 Hz. Plan to measure
>>>> 
>>>>>>>>>>> the
>>>> 
>>>>>>>>>> output
>>>> 
>>>>>>>>>>> of the chronverter see what the 60 KHz is actually at. Though
>>>> 
>>>>>>>> listening
>>>> 
>>>>>>>>>> on
>>>> 
>>>>>>>>>>> a HP3586 receiver it appears pretty close to frequency maybe its
>>>> 
>>>>>>>> just not
>>>> 
>>>>>>>>>>> good enough. Or the fact is OOK modulation does not work because
>>>> 
>>>>>>>>>>> it
>>>> 
>>>>>>>> gaps
>>>> 
>>>>>>>>>>> the PLL in these clocks.
>>>> 
>>>>>>>>>>> 
>>>> 
>>>>>>>>>>> Next steps
>>>> 
>>>>>>>>>>> 
>>>> 
>>>>>>>>>>> - Determine and lay to rest the OOK/need for external modulator.
>>>> 
>>>>>>>>>>> - Power amplifier and loopstick transmitting antenna staying at
>>>> 
>>>>>>>>>>> or
>>>> 
>>>>>>>>>> below
>>>> 
>>>>>>>>>>> 30uv at 30 meters.
>>>> 
>>>>>>>>>>> - Integrate a neo6 GPS receiver in. (Seriously easy)
>>>> 
>>>>>>>>>>> -
>>>> 
>>>>>>>>>>> - Box it up.
>>>> 
>>>>>>>>>>> 
>>>> 
>>>>>>>>>>> Have my eye on one of the Truetime DC468 units and the classy
>>>> 
>>>>>>>> panelplex
>>>> 
>>>>>>>>>>> display.
>>>> 
>>>>>>>>>>> 
>>>> 
>>>>>>>>>>> Hope this helps those of you concerned with what to do should
>>>> 
>>>>>>>>>>> WWVB go
>>>> 
>>>>>>>>>> away.
>>>> 
>>>>>>>>>>> Regards
>>>> 
>>>>>>>>>>> Paul
>>>> 
>>>>>>>>>>> WB8TSL
>>>> 
>>>>>>>>>>> _______________________________________________
>>>> 
>>>>>>>>>>> time-nuts mailing list --  <mailto:time-nuts at lists.febo.com>
>>>> time-nuts at lists.febo.com To
>>>> 
>>>>>>>>>>> unsubscribe, go to  <http://lists.febo.com/mailman/>
>>>> http://lists.febo.com/mailman/
>>>> 
>>>>>>>>>> listinfo/time-nuts_lists.febo.com
>>>> 
>>>>>>>>>>> and follow the instructions there.
>>>> 
>>>>>>>>>> 
>>>> 
>>>>>>>>>> 
>>>> 
>>>>>>>>>> _______________________________________________
>>>> 
>>>>>>>>>> time-nuts mailing list --  <mailto:time-nuts at lists.febo.com>
>>>> time-nuts at lists.febo.com To
>>>> 
>>>>>>>>>> unsubscribe, go to  <http://lists.febo.com/mailman/>
>>>> http://lists.febo.com/mailman/
>>>> 
>>>>>>>>>> listinfo/time-nuts_lists.febo.com and follow the instructions
>>>> 
>>>>>>>>>> there.
>>>> 
>>>>>>>>>> 
>>>> 
>>>>>>>>> _______________________________________________
>>>> 
>>>>>>>>> time-nuts mailing list --  <mailto:time-nuts at lists.febo.com>
>>>> time-nuts at lists.febo.com To unsubscribe,
>>>> 
>>>>>>>>> go to  <http://lists.febo.com/mailman/>
>>>> http://lists.febo.com/mailman/
>>>> 
>>>>>>>> listinfo/time-nuts_lists.febo.com
>>>> 
>>>>>>>>> and follow the instructions there.
>>>> 
>>>>>>>> 
>>>> 
>>>>>>>> 
>>>> 
>>>>>>>> _______________________________________________
>>>> 
>>>>>>>> time-nuts mailing list --  <mailto:time-nuts at lists.febo.com>
>>>> time-nuts at lists.febo.com To unsubscribe,
>>>> 
>>>>>>>> go to  <http://lists.febo.com/mailman/>
>>>> http://lists.febo.com/mailman/
>>>> 
>>>>>>>> listinfo/time-nuts_lists.febo.com and follow the instructions
>>>> 
>>>>>>>> there.
>>>> 
>>>>>>>> 
>>>> 
>>>>>>> 
>>>> 
>>>>>>> 
>>>> 
>>>>>> 
>>>> 
>>>>> _______________________________________________
>>>> 
>>>>> time-nuts mailing list --  <mailto:time-nuts at lists.febo.com>
>>>> time-nuts at lists.febo.com To unsubscribe, go
>>>> 
>>>>> to  <http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com>
>>>> http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
>>>> 
>>>>> and follow the instructions there.
>>>> 
>>>> 
>>>> 
>>>> 
>>>> 
>>>> _______________________________________________
>>>> 
>>>> time-nuts mailing list --  <mailto:time-nuts at lists.febo.com>
>>>> time-nuts at lists.febo.com To unsubscribe, go to  <http://lists.febo.com/
>>>> mailman/listinfo/time-nuts_lists.febo.com>
>> http://lists.febo.com/mailman/
>>>> listinfo/time-nuts_lists.febo.com
>>>> 
>>>> and follow the instructions there.
>>>> 
>>>> _______________________________________________
>>>> time-nuts mailing list -- time-nuts at lists.febo.com
>>>> To unsubscribe, go to http://lists.febo.com/mailman/
>>>> listinfo/time-nuts_lists.febo.com
>>>> and follow the instructions there.
>>>> 
>>> _______________________________________________
>>> time-nuts mailing list -- time-nuts at lists.febo.com
>>> To unsubscribe, go to
>> http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
>>> and follow the instructions there.
>> 
>> 
>> _______________________________________________
>> time-nuts mailing list -- time-nuts at lists.febo.com
>> To unsubscribe, go to
>> http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
>> and follow the instructions there.
>> 
> _______________________________________________
> time-nuts mailing list -- time-nuts at lists.febo.com
> To unsubscribe, go to http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
> and follow the instructions there.




More information about the time-nuts mailing list