333000 Hz frequency generator to 1 Hz accuracy - need assistance

[David Lowrance]

I am trying to either build a frequency marker unit, or obtain proper test gear to produce a 333000 Hz injector wave that is accurate to 1 Hz.
I have several pieces of older test gear, F gens, frequency counter, and none of my equipment is stable enough to generate a 333000 Hz signal to 1 Hz accuracy. The digital synthesizers miss the frequency by from 35 to over 100 Hz but are very stable. The analog ones drift around over time by 100 Hz or more. Here is the rundown on my current test gear.

<**broken link removed**

My latest thought was to use a 33.3000 MHz crystal oscillator and then add some divider chips to drop by 100x in steps of 5, 5 and 4, and the output of the 333000 Hz should then become more accurate and stable. The snag is locating the crystal oscillator at 33.30000 MHz.

Anyone have test gear [F gen] that can produce a 333000 Hz wave form to 1 Hz accuracy. That is 6 digit accuracy or about 30 ppm as I count it?
I will buy a new F gen if I have a correct model number that can do this for me.
Or does anyone have a source for a crystal oscillator at 33.3000 MHz, or any other kind of oscillator that is that stable and accurate?
Thanks,
Dave L

 

[Nigel Goodwin]

You don't get local sources that accurate - the technique would be to phase lock it to a highly accurate radio transmission, such as Rugby in the UK

 

[JimB]

David Lowrance said:

Here is the rundown on my current test gear.
< https://www.resonantfractals.org/PCC/AOM.html>

That link does not work for me.

The digital synthesizers miss the frequency by from 35 to over 100 Hz but are very stable.

What digital synthesiser do you have?
Maybe it just needs a better reference oscillator.
An OCXO (Oven Controlled Crystal Oscillator) should give the required stability, but would need to be adjusted to the correct frequency by reference to either a calibrated frequency counter or an off-air frequency standard.

Nigel Goodwin said:

You don't get local sources that accurate - the technique would be to phase lock it to a highly accurate radio transmission, such as Rugby in the UK

You are a bit out of date Nigel, the standard frequency transmitter is not at Rugby anymore, it moved some years ago.
It is now located at Anthorne in Cumbria.

JimB

 

[kubeek]

My latest thought was to use a 33.3000 MHz crystal oscillator and then add some divider chips to drop by 100x in steps of 5, 5 and 4, and the output of the 333000 Hz should then become more accurate and stable. The snag is locating the crystal oscillator at 33.30000 MHz.

Easiest way I se would be some form of digital sinthesis: the crystal is stable enough, but the base accuracy is off, then you can use a counter or DDS to get you to the required frequency quite accurately. Some small and cheap FPGA would probably be the best way to do this, unless you can find a dedicated chip - there are some programable dividers etc available.

 

[Musicmanager]

Hi

The link to resonantfractals will work if you remove the > after the .html

S

 

[kubeek]

Hi, The link to resonantfractals will work if you remove the > after the .html
S

Oh, now I regeret spending my time to write the previous reply. What a load of BS that site.

 

[jpanhalt]

Here's the main site: https://www.resonantfractals.org
One to do with the powerful effects of copper: **broken link removed**
The original link: **broken link removed**

You may have to go to the home site first, then dig through it.

+/- 1 Hz seems awfully imprecise for those uses. I wonder whether better effects could be obtained with higher precision, such as +/- 0.001 Hz? Or, if +/- 1 Hz is OK, maybe the best frequency is slightly off that center frequency, but needs better precision?

John

 

[JimB]

Oh, now I regeret spending my time to write the previous reply. What a load of BS that site.

I completely agree!

JimB

 

[Pommie]

Anyone have test gear [F gen] that can produce a 333000 Hz wave form to 1 Hz accuracy. That is 6 digit accuracy or about 30 ppm as I count it?
I will buy a new F gen if I have a correct model number that can do this for me.
Or does anyone have a source for a crystal oscillator at 33.3000 MHz, or any other kind of oscillator that is that stable and accurate?
Thanks,
Dave L

1 in 333,000 is 3 in 1,000,000. So isn't that 3 ppm.

Why do you need such accuracy?

Mike.

 

[David Lowrance]

Thanks all. I have it figured out. Thanks for your expert assistance. Alternate energy research is not for everyone, and granted building something on the word of a channeler may be iffy at best.. at least I will know if the dual cone system can work as stated. Walter Russel came up with a similar system which supposedly worked. Since the dual cones have a sharp resonance dip, now I can prune them to perfection and see what they will do.
**broken link removed**

 

[ronsimpson]

Here is 33.3mhz at +/-50ppm which is not good enough. But it is off the shelf.
**broken link removed**
Here is 33.3mhz at 2.5ppm.
https://www.digikey.com/product-detail/en/FXTC-HE73PR-33.3 MHZ/FXTC-HE73PR-33.3 MHZ-ND/3725526
Here is 33.3mhz at 1ppm
https://www.digikey.com/product-detail/en/FXTC-HE73TC-33.3 MHZ/FXTC-HE73TC-33.3 MHZ-ND/3725594

I have a 10mhz oscillator with a 10 turn adjustment. I use WWV and a shortwave radio to calibrate it. I can get the 10mhz to +/- 0.1hz.
I looked and can not find new oscillators like that.

 

[AnalogKid]

David - the problem with many such sites is that the writers do not understand the most basic principles of physics or electronics, have a particular blind spot regarding how test equipment actually works, and thus have no idea when they stray from the possible to the comical. For example, on one of the pages the writer connects an HP frequency counter's reference output to its input to measure the reference, and is amazed that it measures perfectly. It's the reference. It always will measure perfectly, even if it is broken.

At it's core, all test equipment performs some form of a comparison between an internal reference and an external signal. Voltmeter, frequency counter, bathroom scale, whatever, it does not matter. ALL measurement depends on a reference value or quantity of some manner, and the quality of the reference is the single most important factor in the quality of the instrument. Therefore, when the instrument measures its own reference, the result always is perfect unless the instrument is seriously busted.

For example, if the reference oscillator actually were 9 MHz instead of 10 MHz, it still would measure itself as 10 because it thinks it is a 10, and it would measure an external, true 10 MHz source as 11.111111 MHz. This is 5th grade arithmetic; if you have not mastered this, your results will vary.

If you really want to know the accuracy of your frequency measurement system, purchase a rubidium frequency standard (a low-cost atomic clock), turn it on and let it settle for a few weeks, then measure its output. SRS has a nice one for $1500 plus the cost of a decent power supply.

ak

 

[David Lowrance]

"1 in 333,000 is 3 in 1,000,000. So isn't that 3 ppm."

Hz are per second, ppm is pulses per minute. Are you a factor of 60 off or am I?

So if I start with a programmable Oscillator like Digikey CPP series, have them program it for me, set at 33.3000 MHz with +- 25 ppm accuracy [The lowest you can get]
https://www.digikey.com/product-detail/en/0/CPPT1-HT5RT-ND

I then run through a couple divide by 10 chips 74HC4017E good to 60 Mhz, that will drop both the frequency and the ppm down by a factor of 100 to 333000 Hz with a variance of +- .25 ppm.

.25 ppm is +- 0.0041667 Hz
Should then be able to get 8 digit accuracy if the primary crystal oscillator is programmed accurately.

Did I calculate it correctly?


Dave L

 

[kubeek]

ppm is usually parts per million, so a 30Mhz crystal with 25 ppm accuracy will be off by +/- 750 Hz.

 

[David Lowrance]

Thanks, my calculations need to be corrected then.

33.3000 Mhz +- 25 parts per million means the range is what then? Per million what? Per million of maximum f0.
.0000333 Mhz = one millionth
33.3 Hz x 25 = 832.5 Hz
832.5 Hz / 100 = 8.325 Hz off at 333000 Hz

Range will be inside 333008.325 Hz to 332991.675 Hz.

"Here is 33.3mhz at 1ppm"
https://www.digikey.com/product-detail/en/FXTC-HE73TC-33.3 MHZ/FXTC-HE73TC-33.3 MHZ-ND/3725594

Then at 333000 Hz it comes out +- 0.333 Hz x 1
Seems like that one should do it to 1 Hz accuracy.

Expected Range 333000.333 Hz to 332999.667 Hz

Are we there? LOL!

 

[kubeek]

per million of the original value. The same way as one per cent is a hudredth of the original value, and one promile is one thousandth.

Beware of the difference between stability and accuracy. With those specs "Frequency Stability: ±1ppm" the base frequency of that crystal could be say 5% off, but it will stay at that frequency ±1ppm for the next few months or years. The OP was after both accuracy and stability, the accuracy can be cured with initial calibration, but the stability can not.

 

[dr pepper]

MSF at anthorn is not a good signal to phase lock too, noise is a real problem as the signal is carrier switched and dissapears every second, I know I tried to build a freq standard off it, I used radio 4 instead on 198kc which is accurate to 0.0001 ppm.

Subject in hand though, if you want an accurate reg then a ublox neo6 or 7 satellite rx is all you need, these come with a programmable freq ref, if you download the free software from it from ublox you can have any freq you want up to 10mc, there is a built in fll, its fll not pll so at very low freq's theres somphase noise, however the freq is super accurate.

 

[AnalogKid]

I then run through a couple divide by 10 chips 74HC4017E good to 60 Mhz, that will drop both the frequency and the ppm down by a factor of 100 to 333000 Hz with a variance of +- .25 ppm.

Did I calculate it correctly?

No. Because ppm is a ratio, it does not change when the base frequency is merely divided. The 330 kHz will have a tolerance of +/- 25 ppm.

ak

 

[Tony Stewart]

I have designed many frequency synthesizers to 1e-11 short term and 1e-10 per year and 1e-6 or 1ppm over a wide temp. range.

GPS 1Hz , 10MHz better than 1e-10 accuracy , low cost
OCXO 10MHz better than 1e-10 accuracy. high cost ($300 new)
TCXO 1e-6 or 1ppm tuneable to <0.1 ppm at a fixed temperature and voltage. cheap (<$3)
XO 30~50 ppm , tuneable to 0.1ppm if insulated at a fixed temp ($0.30 for Xtal)

However I don't know anything about your project and why you need 0.03 ppm or 1Hz/33MHz.

I would generate a reference signal like 10MHz from GPS Rx and calibrate the counter to 1e-10, since I believe it has an OCXO.
Then Generate the reference signal for the Sig Gen to calibrate it to 1ppm or better if possible. (not sure if possible with autocal. and ref f)

Crystals based on AT cut and MEMS technology are typically 50 ppm and with Tempcomp can be tuned bought for 1ppm
Crystals based on the SC cut crystal are exclusively used in OCXO's and cost 100x more and also 1000x more stable.

The cheapest solution is a 1ppm TCXO for $3. A tuneable one is better. VT-TCXO $13

The PLL synthesizer is the easy part.

 

[Tony Stewart]

IF you are trying to transfer wireless energy, this is not the best way to do it.
My client has been doing transferring over 10 Watts with 1% ferrite wound coils and 1% PU Cap resonant to distances equal to the diameter of the ferrite ( 10cm) for wireless illumination.

High Q torroidal air coils have been around for a long time and mechanical rigidity and interwinding symmetry are key to high Q, but this is only reactive power, for real power, Q of 100 is plenty overkill as the coil gets loaded. Meaning that frequency tolerance of 6 digits is meaningless.