>>> They didn't have gotcha-problems with irrelevant info to teach you
>>> to "read the whole problem before trying to solve it?" :)
>>>
>>Probably.
>
>In our defense, once you get past the first post and don't have a full
>quote, you can't read the OP unless you have two windows. :)
>
>Mark
>
Yes this is an old thread but, the original posters question is very
similar to a GPS disciplined oscillator problem. Where the input is 1 Hz
aka 1PPS from a GPS and the output is 10Mhz from a disciplined oscillator.
For an example look at
http://ve2zaz.net/GPS_Std/Downloads/VE2ZAZ_GPS_Derived_Std_QEX_09_10_2006.pdf
(figure 1).
Depending on the frequency you would adjust both the pre scale of the
input from divide by 16 to perhaps divide by 32 and also the 16 bit value
of Latch value from 6800 hex to another value calculated to get 10Mhz. If
you arrive at a less than prefect multiplication you could vary the
calculated value between the two fixed values using either more one than
the other, any jiter would be lost by the low pass filter.
Stanley
Reply by markt●June 4, 20082008-06-04
>> They didn't have gotcha-problems with irrelevant info to teach you
>> to "read the whole problem before trying to solve it?" :)
>>
>Probably.
In our defense, once you get past the first post and don't have a full
quote, you can't read the OP unless you have two windows. :)
Mark
Reply by Tim Wescott●June 4, 20082008-06-04
Ben Bradley wrote:
> On Tue, 03 Jun 2008 15:18:35 -0700, Tim Wescott <tim@seemywebsite.com>
> wrote:
>
>> Jerry Avins wrote:
>>> Ben Bradley wrote:
>
>>>> The original poster lists freqnencies from 10Hz to 16Hz each with
>>>> 15 decimal places, yet in the original post he says "I would need
>>>> roughly a 0.01Hz frequency resolution" which would indicate only two
>>>> decimal places would be needed for each frequency.
>>>>
>>>> Looks like a typical Usenet question, not even the OP knows what
>>>> he's asking.
>>> He knows what he's asking. He just doesn't seem to know what he's
>>> talking about.
>>>
>>> Jerry
>> Which is why he's asking.
>>
>> I missed the precision specifier, and got caught up in all the digits --
>> comes from a too-strict education in freshman physics, I suppose.
>
> They didn't have gotcha-problems with irrelevant info to teach you
> to "read the whole problem before trying to solve it?" :)
>
On Tue, 03 Jun 2008 15:18:35 -0700, Tim Wescott <tim@seemywebsite.com>
wrote:
>Jerry Avins wrote:
>> Ben Bradley wrote:
>>> The original poster lists freqnencies from 10Hz to 16Hz each with
>>> 15 decimal places, yet in the original post he says "I would need
>>> roughly a 0.01Hz frequency resolution" which would indicate only two
>>> decimal places would be needed for each frequency.
>>>
>>> Looks like a typical Usenet question, not even the OP knows what
>>> he's asking.
>>
>> He knows what he's asking. He just doesn't seem to know what he's
>> talking about.
>>
>> Jerry
>
>Which is why he's asking.
>
>I missed the precision specifier, and got caught up in all the digits --
>comes from a too-strict education in freshman physics, I suppose.
They didn't have gotcha-problems with irrelevant info to teach you
to "read the whole problem before trying to solve it?" :)
Reply by Tim Wescott●June 3, 20082008-06-03
Jerry Avins wrote:
> Ben Bradley wrote:
>> On Tue, 03 Jun 2008 13:07:00 -0500, "markt" <takatz@pericle.com>
>> wrote:
>>
>>>> Just how much precision do you need? You're showing 16 significant
>>>> digits -- do you really need to be locked to 0.0001 parts per
>>>> trillion? And how long do you have to wait to achieve lock on a
>>>> 10-20Hz signal?
>>>> And how stable is it?
>>>>
>>>> I'm not even sure that world-class labs get timing down that
>>>> precise, and if they do it's _only_ world-class labs.
>>> Cesium frequency standards are only accurate to 10e-12. I'm thinking
>>> there's a red herring being chased here.
>>
>> The original poster lists freqnencies from 10Hz to 16Hz each with
>> 15 decimal places, yet in the original post he says "I would need
>> roughly a 0.01Hz frequency resolution" which would indicate only two
>> decimal places would be needed for each frequency.
>>
>> Looks like a typical Usenet question, not even the OP knows what
>> he's asking.
>
> He knows what he's asking. He just doesn't seem to know what he's
> talking about.
>
> Jerry
Which is why he's asking.
I missed the precision specifier, and got caught up in all the digits --
comes from a too-strict education in freshman physics, I suppose.
--
Tim Wescott
Wescott Design Services
http://www.wescottdesign.com
Do you need to implement control loops in software?
"Applied Control Theory for Embedded Systems" gives you just what it says.
See details at http://www.wescottdesign.com/actfes/actfes.html
Reply by Jerry Avins●June 3, 20082008-06-03
Ben Bradley wrote:
> On Tue, 03 Jun 2008 13:07:00 -0500, "markt" <takatz@pericle.com>
> wrote:
>
>>> Just how much precision do you need? You're showing 16 significant
>>> digits -- do you really need to be locked to 0.0001 parts per trillion?
>>> And how long do you have to wait to achieve lock on a 10-20Hz signal?
>>> And how stable is it?
>>>
>>> I'm not even sure that world-class labs get timing down that precise,
>>> and if they do it's _only_ world-class labs.
>> Cesium frequency standards are only accurate to 10e-12. I'm thinking
>> there's a red herring being chased here.
>
> The original poster lists freqnencies from 10Hz to 16Hz each with
> 15 decimal places, yet in the original post he says "I would need
> roughly a 0.01Hz frequency resolution" which would indicate only two
> decimal places would be needed for each frequency.
>
> Looks like a typical Usenet question, not even the OP knows what
> he's asking.
He knows what he's asking. He just doesn't seem to know what he's
talking about.
Jerry
--
Engineering is the art of making what you want from things you can get.
�����������������������������������������������������������������������
Reply by Ben Bradley●June 3, 20082008-06-03
On Tue, 03 Jun 2008 13:07:00 -0500, "markt" <takatz@pericle.com>
wrote:
>>Just how much precision do you need? You're showing 16 significant
>>digits -- do you really need to be locked to 0.0001 parts per trillion?
>> And how long do you have to wait to achieve lock on a 10-20Hz signal?
>
>> And how stable is it?
>>
>>I'm not even sure that world-class labs get timing down that precise,
>>and if they do it's _only_ world-class labs.
>
>Cesium frequency standards are only accurate to 10e-12. I'm thinking
>there's a red herring being chased here.
The original poster lists freqnencies from 10Hz to 16Hz each with
15 decimal places, yet in the original post he says "I would need
roughly a 0.01Hz frequency resolution" which would indicate only two
decimal places would be needed for each frequency.
Looks like a typical Usenet question, not even the OP knows what
he's asking.
>
>Mark
Reply by markt●June 3, 20082008-06-03
>Just how much precision do you need? You're showing 16 significant
>digits -- do you really need to be locked to 0.0001 parts per trillion?
> And how long do you have to wait to achieve lock on a 10-20Hz signal?
> And how stable is it?
>
>I'm not even sure that world-class labs get timing down that precise,
>and if they do it's _only_ world-class labs.
Cesium frequency standards are only accurate to 10e-12. I'm thinking
there's a red herring being chased here.
Mark
Reply by Tim Wescott●June 3, 20082008-06-03
paryanz wrote:
> John,
>
> Thanks for your previous reply it was very useful. I would change my
> question, after talking to my supervisors I have been told to generate a
> constant frequency of 10Mhz as an output based on the following low
> frequencies as an input.
>
> 16.863406408094434 Hz
> 16.722408026755854
> 16.694490818030051
> 13.755158184319120
> 12.562814070351759
> 12.531328320802006
> 12.062726176115802
> 11.148272017837234
> 10.405827263267431
> 10.060362173038229
> 10.040160642570282
> 10.010010010010012
>
>
> So my frequencies are very low and I have to generate a constant output of
> 10 Mhz by using the above input frequency as an input.
>
> Would a fractional PLL be a solution to this problem.
>
> Thanks for your help in advance
> Aryan
Do you mean you have to use one of the above frequencies for a reference
and generate 10MHz from it, in a way that is somehow locked?
In that case, one method you could use would be to use a DDS as the
divider in a PLL, to generate, for example, a nominal 12.562...Hz
feedback signal that you then lock using the 12.562...Hz reference.
One good way to do this would be to sample the reference signal in a
DSP, using an ADC and DDS that is clocked by the 10MHz oscillator, and
servo the 10MHz oscillator from the DSP. All in all it would be a fun
project to do.
Just how much precision do you need? You're showing 16 significant
digits -- do you really need to be locked to 0.0001 parts per trillion?
And how long do you have to wait to achieve lock on a 10-20Hz signal?
And how stable is it?
I'm not even sure that world-class labs get timing down that precise,
and if they do it's _only_ world-class labs.
--
Tim Wescott
Wescott Design Services
http://www.wescottdesign.com
Do you need to implement control loops in software?
"Applied Control Theory for Embedded Systems" gives you just what it says.
See details at http://www.wescottdesign.com/actfes/actfes.html
Reply by Robert Lacoste●June 3, 20082008-06-03
"paryanz" <dr.asthana@gmail.com> a �crit dans le message de news:
puGdnXtMwJncg9jVnZ2dnUVZ_hOdnZ2d@giganews.com...
> Hi,
>
> This is my first post to this forum. I have a reference oscillator of 12.7
> MhZ, how do I generate the following frequencies using this oscillator
>
> 16.863406408094434 Hz
> 16.722408026755854
> 16.694490818030051
> 13.755158184319120
> 12.562814070351759
> 12.531328320802006
> 12.062726176115802
> 11.148272017837234
> 10.405827263267431
> 10.060362173038229
> 10.040160642570282
> 10.010010010010012
>
> I would need roughly a 0.01Hz frequency resolution. Would a fractional PLL
> be a solution to this problem.
>
> Most of the papers I have read deal with increasing the output freq to
> several Mhz or Ghz for implementing a PLL. But this implementation of mine
> requires to reduce the generated freqs, I would appreciate any help advice
> or some kind of code which would help me solve this problem.
Dear Aryan,
A direct digital synthetizer (DDS) would be far more suited to your problem
than a PLL. For example with a chip like the AD9832 (5$/1k budgetary price
on Analog's web site) and clocked with your 12.7MHz source you will get a 32
bit tuning word meaning a resolution better than 0.05Hz from DC to around
5MHz. And will require only 10mA. See www.analog.com/dds.
Friendly yours,
Robert Lacoste
www.alciom.com