Atomic clocks / General Chatting / Forums

Forums

4hv.org :: Forums :: General Chatting

« Previous topic | Next topic »

Atomic clocks

Move Thread

LAN_403

Conundrum

Wed Jan 20 2016, 11:48PM

Registered Member #96 Joined: Thu Feb 09 2006, 05:37PM
Location: CI, Earth
Posts: 4062

Hi, does anyone here have one of these?

apparently ebay have lots, mostly Rb modules.
May be worth getting one as my frequency counter OCXO is way off.
-A

Kizmo

Thu Jan 21 2016, 12:50AM

Registered Member #599 Joined: Thu Mar 22 2007, 07:40PM
Location: Northern Finland, Rovaniemi
Posts: 624

I dont have one but recently came across this article:

Pretty cool little chip :)

Sulaiman

Thu Jan 21 2016, 07:55AM

Registered Member #162 Joined: Mon Feb 13 2006, 10:25AM
Location: United Kingdom
Posts: 3141

I had a FE5680A Rb oscillator, from eBay when China was upgrading its cellular networks
I checked the frequency against a gps-locked oscillator ... fraction of one Hz difference
(I guestimated <1/16 Hz watching Lissajou on 'scope)
Sold it to a friend at work who had a use for it.

OCXO do drift with time but older units drift more slowly
so it would be worth 'tweaking' your OCXO

Conundrum

Sun Jan 24 2016, 10:18AM

Registered Member #96 Joined: Thu Feb 09 2006, 05:37PM
Location: CI, Earth
Posts: 4062

Cool, thanks.
Do you have any tips on making some sort of calibrator that would give me a rough guide?
Thought maybe lock onto the no-longer-in-Rugby signal and then derive the 1ppm frequency from this using a second oscillator with a PWM to keep it on track.

Sulaiman

Sun Jan 24 2016, 01:08PM

Registered Member #162 Joined: Mon Feb 13 2006, 10:25AM
Location: United Kingdom
Posts: 3141

the 60 kHz MSF and 77 kHz DCF time stations have modulation that I found unsuitable for a direct frequency reference

198 kHz is usable but does have phase modulation as well as the obvious audio amplitude modulation
with a little extra (TA7642 + LM386 + speaker etc.) you can listen to Radio 4 (I don't)
the 500 kW at <30 km gives me a nice signal !

even gps-locked clocks aren't perfect but 1ppm ok

Although Rb clocks are excellent they have an offset from the ideal atomic frequency that drifts a little with age
so not perfect but even an old one should be =/- 1 ppm

more than ever needed

Conundrum

Mon Jan 25 2016, 05:19AM

Registered Member #96 Joined: Thu Feb 09 2006, 05:37PM
Location: CI, Earth
Posts: 4062

Thanks for the tips!
It would be amusing to actually calibrate the thing then put a "calibrated in 2016 by Nygmatec" sticker on it..

EDIT: that chip is a 3 pin replacement for the MK484.
Whats the least complicated method of making a good antenna, thought maybe use ferrite rings in series.
Finally a use for those $^!Â£&!Â£ rings I have about a million of.

-A

woodchuck

Mon Jan 25 2016, 05:03PM

Registered Member #39190 Joined: Sat Oct 26 2013, 09:15AM
Location: Boise National Forest
Posts: 65

I have a number of OCXOs and wanted to get a quick evaluation of their calibration and drift. Most have either 10 MHz or 12.8 MHz output. So these were passed through a flip-flop (functioning as a prescaler) to obtain a clock in the 5.0 to 6.4 MHz range. This was applied to the oscillator input of a Parallax Propeller. The 16X PLL was enabled on the Propeller, which then drives an internal 32-bit counter (as well as all eight processor cores) at 80-102.4 MHz. One of the cores of the processor is then tasked at polling the 1pps output of a GPS receiver with 50ns or better resolution. (Another core can be dedicated to this same task with a second GPS receiver, for comparison.)

It is assumed that short-term deviations will largely be the fault of the GPS, and that long term drift will always be the fault of the OCXO. In this way I've been able to track drift and error with excellent resolution over a wide range of time scales, from seconds to months.

Sulaiman

Mon Jan 25 2016, 07:22PM

Registered Member #162 Joined: Mon Feb 13 2006, 10:25AM
Location: United Kingdom
Posts: 3141

The simplest antenna suitable for 198 kHz is a vertical whip
any length of wire suspended vertically, connections to 'Earth' and the bottom of the whip.
Such an antenna will probably receive a stronger signal from am broadcast transmitters than the 198 kHz
so low pass (or band pass) filtering will be required

Next simplest is a multi-turn loop, 1 ft. to 1 or 2 m diameter, round, square, octagonal etc.
tune the loop to 198 kHz with one or more capacitors
the resonant loop will form a good band pass filter.

Finally a ferrite rod antenna,
no reason why you could not use a 'roll' of stacked ferrite toroids as the core
not as ideal as a solid ferrite rod but good enough.
I prefer NiZn ferrite (non-conductive to a dmm) to MnZn (easily detectable conductivity using any dmm resistance range)
but at 198 kHz I guess even some MnZn ferrites may work ok

I suggest the multi-turn loop,
roughly speaking the important parameters are resonant frequency and Q
more turns and/or larger size increases received voltage
use loop in the vertical plane for LW, max response in the plane of the loop, null on-axis.
Spaced turns (one to six wire diameters air space) give better Q than close turns
Any copper wire will work, 0.4 to 1mm is probably 'optimum', 660/0.04 'Litz' wire would be better, but $$$

Whatever, if you want audio then a receive bandwidth of c9 kHz is required,
so the 'q' required is only 198/9 = 22
whereas for a frequency reference you want the highest Q practical to reduce 'noise/breakthrough/interference'

I tried an op-amp bandpass filter to remove (severely attenuate) phase modulation, but failed.

A bang-bang frequency lock loop should work well, especially if integration times are long (>5s for 1 ppm ?)

P.S. a quote that I vaguely remember from time nuts

A man with a clock knows the time, A man with three clocks is never really sure.

P.P.S. I do not use any of the above antennas, because I receive such a strong signal
I use a 6" pvc pipe with 92 turns of pvc insulated 1 mm2 multi strand tinned copper wire
that I have from some long forgotten experiment, definitely not ideal
but I get 0v5 rms open-circuit when tuned, which is enough for my 'scope etc.

Conundrum

Wed Feb 03 2016, 06:15AM

Registered Member #96 Joined: Thu Feb 09 2006, 05:37PM
Location: CI, Earth
Posts: 4062

Thanks!
I'm actually trying to add a radiation sensor to a clock at the moment, the idea being that it serves as both a battery indicator and shows if something bad (tm) has occurred.
See

also you can make one using pyrolytic graphite and a silicon photodiode but the front contact is in fact the graphite (!)

And

which I can fix trivially with "Andre's patent pending 10M+0.01uF tube series bypass circuit (tm) (c) "

Moderator(s): Chris Russell, Noelle, Alex, Tesladownunder, Dave Marshall, Dave Billington, Bjørn, Steve Conner, Wolfram, Kizmo, Mads Barnkob