[time-nuts] Has anybody checked this? GPSDO in kit

Bob Camp kb8tq at n1k.org
Sun Oct 23 09:08:16 EDT 2016


Hi

You *could* calibrate the nominal frequency a chip very cheaply these days. They may / the 
may not, who knows. It certainly *is* done that way on very low cost wrist watches. 
I’d bet that they do and it leaves the factory set within less than 1 ppm. 

A good generic crystal is still going to be a few ppm sort of proposition once set. Each ppm is 2.6
seconds a month. If you only get bothered by it being off one minute, it’s quite possible
for a simple clock to hang in there for a year. 

The interesting thing is that *most* places set the device to run fast. There are some
great sites on the internet documenting this. You normally are bothered by the 
deliberate offset before you notice the actual accuracy of the clock. 

Bob

> On Oct 22, 2016, at 10:52 PM, Mark Sims <holrum at hotmail.com> wrote:
> 
> Yep,  I recently sorted through a bag of 100 crystals from China ($10, shipped) looking for a "good" one.  They were ALL good...  a complete waste of time.  I was rather amazed at their consistency and performance for a 10 cent part.
> 
> Last year I bought an alarm clock / game from China (looks like 7 sticks of dynamite with an ominous circuit board / LED display strapped to it).  It uses a 40 pin (AVR?) processor driven by a 16 MHz processor crystal.  I have not set it in over a year and it is still within a few seconds.  I suspect they measure the frequency and have a calibration tweak stored in EEPROM... but that seems excessive work for a $20 toy.  I highly doubt they go as far as doing temperature compensation.  Maybe they characterized a bucket of XTALs and use a generic compensation factor?
> 
> ----------------------
> 
>> Yes, you can build gear to do temperature runs on crystals and sort bags full of them. 
> It’s likely that your whole bag of 5,000 came from the same bar and your 
> net result will all look a lot alike…..
> _______________________________________________
> 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