frequency standard ideas ?

[Flump]

Thanks again everyone for their input on this subject

babysitter has been trimming a TCXO for me and it is currently on its way to me  O0
it needed trimming because I wanted a slightly under 10mhz reference,
I will leave it to him to give any detail on it because he did all the testing.

If a few cheapo 74HC390 dividers were fitted then it would provide tap off points for 10MHz, 5MHz, 2MHz, 1MHz etc etc down to below 1kHz. So there would be outputs that would fall within the frequency range of the DMM and also the timebase ranges of the scope.

I would love this to be the next stage for my frequency standard is it easy to implement ?
I have been googling frequency dividers but not found anything yet basic enough for me
to try and build.

[linux-works]

http://www.leapsecond.com/pic/picdiv.htm

I have some spare chips, so if you need one burned, I could probably help you out.  or, just buy from the guy who wrote the code (probably better that way, since he deserves credit for his hard work).

[babysitter]

You should go the PICDIV Route definitely if you are interested in using it for gate times and other slow events.
(Remember the Consider the 10 Hz error then!

Another thing you might look into is buying some DDS Module from ebay and feed it from the TCXO; these would allow you to set a ratio and get a signal at your TCXOs Frequency times DDS ratio. Good to provide a test signal e.g. in the other ham bands.

If you need a real 10 MHz Reference for this, feel free to trim a Oscillator exactly 10 Hz above the one you got from me. Easy with a ham radio sw rig and using soundcard + analyzer software or your counter.

BR

Babysitter

[Flump]

Just wanted to update this incase someone else reads this hoping to findout how to do a 10:1 divider

the 74LS90N which is easy to get hold of can do a 10-1 division and it tells you how to in the data sheet, at the top.

and I wanted to show you what babysitter made for me
its a tcxo which he adjusted to be 10hz under 10Mhz
which I used to calibrate my frequency counter to

Thanks Hendrik 

[KF5OBS]

This is what I use: https://baltic-lab.com/2012/10/review-jackson-labs-gpstcxo-eval-board/

Pretty affordable even for a home-lab!

EDIT:

[...] and I wanted to show you what babysitter made for me

For a second I hoped you had bad grammar there and I was about to ask where I can hire that babysitter...

[KF5OBS]

Sorry for double post but I just remembered that I have something laying around that fits very well in the topic: A brand new GPS disciplined Atomic Clock Standard. Atomic Clock for precision, GPS for accuracy (clock edge on UTC second).

[grumpydoc]

Very nice - from whence and how much?

[David Hess]

They are expensive and apparently they have had problem with the chip scale atomic clock modules:

Affected Products: Jackson Labs or Symmetricom time/frequency reference boards based on Symmetricom chip scale atomic clock, most common unit is the CSAC GPSDO sold by both symmetricom vendors and direct from Jackson labs

Type of issue: Short and Long-term reliability, mfr couldn’t assure that if it works for 1week, 1month or 1year it is not an affected unit

Issue: CSAC Loss of vacuum, sudden onset failure mode

Result: CSAC unable to control the temperature, Indefinite warm-up state, ultimately the unit fails to reach a lock state

Failure Rate: Vendor noted ~5% of all units, internally it has been 1 of maybe a dozen or less units to date

Resolution Status: Issue acknowledged by manufacturer but not resolved, still affecting all current stock and current production units, Jackson Labs will replace the CSAC part on the board and return without a fee though admittedly the new part may also fail
 
Details:
While resolving an issue with a Jackson Labs CSAC GPS Disciplined Oscillator board, we were made aware of a reliability issue affecting all timing boards built around the new Symmetricom Chip Scale Atomic Clocks (a.k.a. CSAC).  There is a fundamental design flaw that can result in the loss of vacuum in the CSAC, which prevents the unit from being able to effectively control the
part’s operating temperature. The end result is a unit that is in an indefinite warm-up state. We were informed that all current and even currently produced units can fail in this way. The number we were given was ~5% of all units will fail, so we happened to get one of those 5%. The failure mode is a very sudden onset failure, for our case one day it worked and the next day (during a test of course) it did not. Also this is not a case of ‘if it works the first week then it will be fine’, it can fail anytime (in my case I had used it on a regular basis for several months prior to the failure). Our unit also spent 99% of its life in an air conditioned lab so certainly not subjected to even a remotely harsh environment.

[G0HZU]

The original requirement was to calibrate a 100ppm DMM or a scope timebase!

A cheap OCXO is all that is needed.

However, if anyone in the UK wants a novel quick and dirty 'standard' that is (kind of) traceable in the UK you could make a little 198kHz trf receiver from junk parts and run it into a (very) hard limiter and measure the frequency of Radio 4 when averaged over a few seconds with a reciprocal counter. I used an active trf stage ahead of the limiter section of a salvaged Philips NE604/SA604 FM IF/DET chip.

I just tried this and could reliably get 198.000001kHz on an old Philips reciprocal counter (run from a decent external reference) with a 10s measurement time. I've shown a 1 at the end of the 198.000001kHz but sometimes it shows a zero here or flicks down 1 LSD. If I watch it over several minutes there is the odd rogue count that jumps up about 5-10 LSDs but most counts are as good as the digits I posted above. It will probably work more reliably during daylight hours.

It 'works' within a few seconds (as opposed to 30-60minutes for a big OCXO or GPS) and runs on flea power.

I bet there's hardly anybody on this forum that 'needs' anything better than this for home use.

About 20yrs ago I designed a cheap and cheerful offair standard that locked a 10MHz reference to the 198kHz signal from R4. It's built in dead bug style and is (now) fitted inside an old plastic ice cream tub. It works a bit better than the crude trf receiver but I'm not sure I really need the extra performance. I never ever  bothered making a neat version beyond the ugly prototype because it only got used a couple of times a year at most.

[G0HZU]

I tried it with my other reciprocal counter and set it to a 1 second gate time. This counter didn't cope as well with the jittery output from the limiter so I tried averaging the readings over a few minutes.

Obviously, it jumps around like crazy like this for each 1 second reading because the limiter can't remove all the modulation but I logged the counter and averaged the readings with a PC and was able to get a 198.000000kHz average within 200 counts.
Obviously, not everybody has access to a reciprocal counter and some counters might not cope as well as mine have. But it is certainly a cheap and cheerful option for people in the UK.

For anyone who doesn't know about BBC Radio 4 in the UK the 198kHz transmission frequency from Droitwich is locked to a very accurate time standard. So this adds some traceability to the cheapo trf receiver + limiter approach.

[Andy Watson]

However, if anyone in the UK wants a novel quick and dirty 'standard' that is (kind of) traceable in the UK you could make a little 198kHz trf receiver from junk parts and run it into a (very) hard limiter and measure the frequency of Radio 4 when averaged over a few seconds with a reciprocal counter. I used an active trf stage ahead of the limiter section of a salvaged Philips NE604/SA604 FM IF/DET chip.

What a coincidence! I was looking at Radio 4's time keeping abilities earlier today. It's not so dirty if you can lock to it, and once-upon-a-time it used to be traceable to the NPL standards.
http://tinyurl.com/lzurwkm
Looks like NPL stop monitoring frequency and time in 2012. The BBC now say that the Radio 4 frequency is locked to a rubidium frequency standard.

What I was actually trying to find out was whether Radio 4 longwave is still phase modulated with time signals.

[KF5OBS]

They are expensive and apparently they have had problem with the chip scale atomic clock modules:

Expensive is relative to what you are doing. Just because you can get cheap burnt out rubidium standard from eBay for $ 100, you shouldn't distort the fact that they once cost just as much as the CSACs. If I remember correctly, Jackson labs responded and said they will replace Units affected by the issue free of charge even if they're out of their normal warranty time already.

[David Hess]

They are expensive and apparently they have had problem with the chip scale atomic clock modules:

Expensive is relative to what you are doing. Just because you can get cheap burnt out rubidium standard from eBay for $ 100, you shouldn't distort the fact that they once cost just as much as the CSACs.

Well, I meant expensive in the sense of "if you have to ask, then you cannot afford it".   I made some effort to find prices but stopped when I found that their comparable GPSDOs are in the $1000+ range.  One of their low end units can be had for like $400 or so.

A GPS disciplined crystal oscillator will be a lot more economical if you can live with its limitations.

If I remember correctly, Jackson labs responded and said they will replace Units affected by the issue free of charge even if they're out of their normal warranty time already.

They did which I have to give them credit for and they were not the only ones affected by the Symmetricom CSAC issue.

[bingo600]

They are expensive and apparently they have had problem with the chip scale atomic clock modules:

Expensive is relative to what you are doing. Just because you can get cheap burnt out rubidium standard from eBay for $ 100, you shouldn't distort the fact that they once cost just as much as the CSACs. If I remember correctly, Jackson labs responded and said they will replace Units affected by the issue free of charge even if they're out of their normal warranty time already.

It does say eval sample on the CSAC , does that mean you got one from Said for free ?

Then expensive is really relative 

/Bingo

[KF5OBS]

It does say eval sample on the CSAC , does that mean you got one from Said for free ?

I haven't paid for 99 % of the things on my blog. That does include the CSAC. The CSAC was a real challenge for Said because the CSAC itself is not Said's product. It's the GPSDO part. The CSAC is the biggest cost factor though. So he approached Symmetricom, got a free CSAC and then modified one of their PCBs and the SW to make it work with the "off standard" 10.24 MHz frequency. Thought that was pretty nice and impressive to do that just because.

However, I think they sell for around $ 1500 now. And like I said earlier, a good nuclear standard is definitely worth that.  I'm sure the rubidium standards all over the place from eBay used to cost a similar chunk of money back when they were new and had acceptable phase noise. Usually, if you need that amount of precision, odds are that you have the cash.