Frequency standard, radio 4 or msf?

[dr pepper]

I want to build a frequency standard, not too bothered about the actual frequency, maybe 1 mhz, for callibrating my test gear and maybe connecting up to my frequency counter (would need surgery), also I have a pendulum master clock from a bt telephone exchange that I'd like to sync too.
I've seen these rubidium oscillators on ebay, however I suspect that these are probably toward the end of their lamp life.
I have a schematic for syncing a 10mhz xtzl osc to radio 4, here in the uk is broadcast on 198kc and the carrier is locked to an atomic standard.
I've also considered building a receiver for the msf time code signal on 60kc, but I dont know if I'd be able to get that to work well as the transmission uses amplitude shift keying, I thought of just receiving the signal and then squaring it, but I suspect theres going to be issues with noise with the signal strength all over the place.
I have built several unusual clocks using a symtrick msf receiver, I get a good signal as I'm about 40 miles from the tx.
Any suggestions?

 

[JimB]

If MSF at Anthorn is in your back yard, go for it.

Simply amplifying and squaring the carrier will not work well.
I have an MSF frequency standard which I built a couple of years ago, it is based on a design in the RSGB magazine RADCOM from April/May 1994.
The design is a bit dated now, but does work.
I had to make a few changes to the RF section to improve its selectivity because of a very strong signal on 51.95kHz from Crimond, about 10 miles away.

If you are interested, send me a e-mail address in a PM and will send a scan of the article to you.

JimB

 

[dr pepper]

Yes Jim if you dont mind please scan and send me the article, that would be very interesting, I've spent a couple of hours trawling through the net and couldnt find anything at all on using msf as a frequency standard.
My comms receiver doesnt pick anything up in the lf band, in fact not a lot below 100kc so I dont think I'll have a problem.
I like the radio 4 pll standard, what concerns me about that however is the fact that once the transmitter goes down the BBC are no longer going to broadcast on LW.
I dont know if theres anything I can do in return, I have a stock of odd ball ic's from tv's and industrial equipment, if you ever need an ic no longer in prod mail me.

Doh I'm forum illeterate couldnt work out PM, my email is blah blah blah.

 

[Diver300]

The 1 Hz output from a GPS receiver is at atomic clock accuracy. Why don't you just calibrate to that?

 

[JimB]

Doh I'm forum illeterate couldnt work out PM,

The PM did work, and sent me an e-mail notification.

I have sent the info and some photos of my unit by return e-mail.

JimB

 

[mvs sarma]

i suppose that the horizontal frequency (15.625KHz in case of UK) recovered from a TV is a good enough as it is locked to atomic level be the transmitting authorities. Perhaps a small coil hooked to side of the boy can take the clock by induction.

 

[dr pepper]

MVS - good idea, did run through my mind, however where I live I cannot easily receive an analogue tv signal, everything has gone dvb.

I have an article on gps sync and I have a old pocket gps that has a multipin connector at the base I'd considered that, short term stability is as far as I can tell not as good locking to 1 hz.

Thanks Jum thats great printed out.

Like you of course I couldnt build the eact same circuit, I thought to do away with the counter chips and part of the vco, use a pic 12f675 with a 10mc xtal, and use a varactor on the xtal to make the pic's clock voltage controlled, then within software divide 10mc's down to 60kc's to be used by the pll, instead or ordering 'hc66 xors I'll use that section of a 4046.
The secondary function of the pic could also be to control the oven for its own 10mc xtal.
I like the welding rod case, I have a couple of those about.

Thanks very much for taking the time to dig that out and scan it for me Jim, favour owed.

 

[mvs sarma]

GPS modules are there that output 1Hz and 10MHz in some. we need to search

please search for siliconchip magazine article of GPS conditioned clock frequency. perhaps the same was also published by EPE magazine of UK

 

[mvs sarma]

**broken link removed**
please check whether the board under the ad can help

 

[dr pepper]

Jim, did you know theres a mk2 of this, the author doesnt seem right confident about either of them for some reason:

http://www.g4jnt.com/msfmk2.pdf

 

[JimB]

Yes, I am aware that the author is not completely satisfied with its performance, but he is using it to generate signals at 10GHz (10,000 MHz), so there is a fair bit of multiplication going on with his application.

The practicalities of my application are as follows:
I use the standard for checking my Racal 1999 frequency counter, which I have fitted with a TCXO (Temperature Compensated Crystal Oscillator).
Measuring the 1 MHz output of my "standard" the jitter on the display of the counter is roughly +/- 30mHz (milli Hertz).
The counter and the standard agree with each other (+/- 30mHz) over extended periods.
That is good enough for my purposes.

Translating the 30mHz jitter to the authors case, 30 milli-Hertz multiplied by 10,000 is 300Hz.
If you are working with SSB bandwidths, that is awfull!

Am I inclined to try something better? Maybe, one day.
Until then I consider what I have to be very much better than what the average amateur (and many professionals) have available.

JimB

 

[dr pepper]

I'm inclined to agree.
The 100meg resistor charging a bank of caps doesnt appeal to me, you'd need to have some caps with zero leakage for that.

 

[dr pepper]

I coulndt see all of your post, this pc doesnt work with the new site, I had to interpolate, if I missed something I'm not being ignorant.

 

[BeerBelly]

I remember when PLL circuits first became popular. They were said to be able to demodulate a frequency which was below the noise level. With the MSF signal in your back yard, I think I would try that as a fun project

 

[dr pepper]

That mustve been a while back, I've a 4046 thats dated '79.
Cant you get wwvb, thats also on 60kc and broadcasts all over the US.
There are 2 signals on air here in the uk that are precise, both of them are 'messed up', msf on 60kc has amplitude shift keying and radio 4 has phase modulation.
A second hand rubidium oscillator is probably not such a bad idea.
For now though I am working on the msf standard, I might put together a simple radio4 standard too, but not 10mhz just the same as the carrier 198kc, I need a standard to check the standard.
The issue with the msf circuit linked to above is that it can lock to itself, once setup its fine getting to that point I might need verification.

 

[WTP Pepper]

The problem with using a time standard such as MSF is the free wheeling your PLL circuit will have to do in the absence of the carrier (as it's ASK) making uneccessary phase noise.
Have a scan using a HF receiver in the 1-30MHz band. There used to be many highly accurate carriers at around 10MHz that would be ideal.
50Hz is highly accurate but over a 24 hour period which isn't any use.
Many GPS receivers do have a 10MHz output which is probably as good as it gets.
Rubidium & Caesium standards will be expensive.

 

[dr pepper]

The only frequencies I know about I can get easily are 60kc msf and 198kc radio 4.
There are some russian freq's at 4.999 and 9.999 mc's, I cant get them and I dont really want to have an inverted v aerial over my house to do so, and even they are modulated.
BBC TV used to maintain highly accurate sync pulses, but analogue reception here has now ended, and so are most of the others in this country.
I can get a rubidium source for around 50 quid, however how long is left for the lamp life is unknown.

 

[dr pepper]

OK, so I prototyped a circuit on breadboard to sync an xtal osc to 198 kc radio 4.
Performance is well rubbish, the o/p wanders by 10 to 15 hz at the loop filter rate.

The carrier is received by a simple trf receiver mounted remotely, this goes to a cmos gate where its squared then divided down to 2 kc, a 10mc xtal is also divided down to 2kc, the pair of 2kc signals go into a 4046 pll, the type 2 phase comparator goes into a 0.01 hz lpf which then biases a 500pf varicap diode in one leg of the xtal, there is a trim cap to get the o/p smack on 10mc with the control volts in the centre of the 5v rail.
But the o/p wanders around by 10hz or more, I think the mod on the carrier is making the pll go up and down.

In an attempt to remove the modulation I replaced the gate that squares the 200mv from the receiver with a ne567 pll tone decoder, a tranny on pin 5 takes off the synthsised signal which is a exact copy of the incoming signal, only a 0.1 hz loop filter is placed in the vco loop which I thought would remove pretty much all the mod, but it doesnt the system still wanders like its been on beer.

Is there such thing as a 198kc crystal filter?, I think it will take something like this to remove the am sidebands and phase mod from the carrier, maybe I could use a pll to multiply the carrier by 18 which gets close to 3.579 mcs uk colourburst frequency and use a ladder of these.

 

[JimB]

Is there such thing as a 198kc crystal filter?

Only if you build it!

I think it will take something like this to remove the am sidebands and phase mod from the carrier, maybe I could use a pll to multiply the carrier by 18 which gets close to 3.579 mcs uk colourburst frequency and use a ladder of these.

Have you ever heard the saying "When you find yourself at the bottom of a deep hole, stop digging" ?
I think that it may apply in this case.

Have a look here for an interesting technique for using the 198kHz transmission:
https://www.qsl.net/pa2ohh/07freqstd.htm

JimB

 

[dr pepper]

Yes maybe your right, I didnt know how deep the hole was.

I had allready looked at the droitwich standard project, and this one:

https://www.epemag3.com/lib/free_projects/lab_equipment/0602- Frequency Standard Generator.pdf

the same droitwich project has been converted to work off the french standard frequency station alluis, and also dcf77, the latter is interesting as it has ask mod.

https://www.qsl.net/pa2ohh/12freqstd.htm

I had tried the circuit on breadboard and wasnt really convinced with the results, however maybe its worth a revisit, being audio the system will not be affected by r4's phase modulation, a lm567 tone decoder could also be used here to indicate a signal and to generate a stable frequency back from the receiver.
Maybe I should revisit that attempt, this ones wasting time.