[time-nuts] time-nuts equipment verification from scratch (was: WTB: GPSDO)

Bob Camp kb8tq at n1k.org
Tue Mar 21 07:36:23 EDT 2017


Hi


> On Mar 21, 2017, at 1:07 AM, Chris Albertson <albertson.chris at gmail.com> wrote:
> 
> I built one of these using a PWM DAC also.   The design was posted to this
> list so I can't take credit for the idea.   But we used two PWM output
> pins.   The PWM provides more voltage range than is needed by the OXO's
> EFC.  To the output was scaled by a voltage divider.  This also scaled down
> thew step size.     The second PWN output was scale down even more, like
> maybe 100X more.  The two PWM outputs were added.  One does course
> adjustment the other fine.   The software first sets the course PWM and
> then the fine one takes over.
> 
> But the PWM output was just run through an RC filter with a very long time
> constants low pass filter with corner freq. < 1 Hz.   The goal was to build
> a VERY low cost GPSDO and adding a good external DAC would add to the cost.
> 
> Someone here recently suggested that one could do as well by simply
> adjusting a good oversized crystal with a screw driver as they could using
> a simple GPSDO.   Well, before building the GPSDO I tried keeping my OXO in
> sync with my Tunderbolt using just a screw driver and a dual trace analog
> tektronix scope.  It is REALLY hard to do with a screw driver.

You need a finer adjust pot on the EFC. It’s no different than the process you 
describe above with the PWM’s. With a PPS that is good to 10 ns, you can get
to 0.1 ppb in 1000 seconds with margin. You will have a pretty good idea of 
what is going on in 100 seconds. 

If you are going to rig it up, a 20 turn wire wound pot with a dial on it is the high
end approach. Set up the pot with a ~1x10^-8 full scale range. You then have 
roughly 5x10^-10 per turn. The cool part is that you can log the readings and
work out what’s going on with the OCXO.

Bob

>   Some tines
> I'd think I had it right then I'd look at the scope after 30 minutes and
> fine one sine wave had gained 1/4 cycle on the other.  Lots of reason for
> this, perhaps one of my voltage regulators are temperature sensitive,
> "stiction" in the screw I was turning.  Who knows.     But my simple GPSDO
> would notice the 1/4 cycle error and fix it automatically
> 
> It real life for practical purposes I use the Rb clock, was lucky to get
> one at the old $35 price.
> 
> On Mon, Mar 20, 2017 at 4:43 PM, Mark Sims <holrum at hotmail.com> wrote:
> 
>> The "DAC" was PWM based, but used a separate voltage regulator for the
>> "reference".  I never tried it using the USB power as the reference.
>> 
>> The OCXO (+board) uses less than 500 mA warming up (which it does rather
>> quickly).  It's in a small hermetic package about twice the size of a
>> standard DIP-14 oscillator package.  There was a Ebay seller several years
>> ago offering them at $15 each or 10 for $100.
>> 
>> The Chinese "Arduino" board (it's not really and Arduino,  just a MEGA 328
>> and a proto area)  has a micro-USB connector for power input but does not
>> implement USB data.   I used the processor serial port with a level shifter
>> dongle.   The firmware was a cheap and dirty hack and I didn't implement
>> much in the way of control or monitoring... never got around to improving
>> it.  The project was basically "Hey, I forgot I had those parts...  Hmmm,
>> one could build a simple GPSDO... why not?
>> 
>> -----------------
>> 
>>> Did you use the Arduino's PWM output plus a LPF for the DAC, or a
>> separate
>> DAC? If PWM, did you have problems with noise or sensitivity to the
>> USB-provided supply voltage?
>> _______________________________________________
>> 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.
>> 
> 
> 
> 
> -- 
> 
> Chris Albertson
> Redondo Beach, California
> _______________________________________________
> 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