[time-nuts] HP5061B Versus HP5071 Cesium Line Frequencies
Richard (Rick) Karlquist
richard at karlquist.com
Fri Jun 2 15:10:30 EDT 2017
I said no *manufacturer* does it this way. NBS is not
a manufacturer. In a one-off money-is-no-object non-portable
standard, you can make direct multiplication work. It
will not work well in a 5061, because of RF leakage
issues specific to the 5061 that are well documented.
Bolting on a different synthesizer does nothing to change that.
The decision not to use direct multiplication has nothing to
do with not being able to figure out how to synthesize the
correct frequency. Certainly by the time we did the 5071A,
we were already using DDS, and it wouldn't have been any
problem to synthesize for direct multiplication if we had
wanted to do that. You seem to be doing it the hard way
(pre DDS) involving Diophantine equations. So it's easier
to do direct multiply than it used to be, but that doesn't
necessarily mean you should do it that way.
On 6/2/2017 10:59 AM, Donald E. Pauly wrote:
> A guy by the name of David W. Allan used direct multiplication to
> build NBS-4 and NBS-5, see http://tf.nist.gov/general/pdf/65.pdf . He
> didn't see anything wrong with it. He used a commercial frequency
> standard modified from 5 mc to 5.006880 mc. That in turn was
> multiplied by 1836. This was a multiplier chain of 2·2·3·3·3·17.
> When multiplied to 9192 mc, this is 90 cycles low so the standard
> would be forced high by 0.05 cps.. They measured the locked frequency
> standard to determine the actual frequency of the cesium line. I
> propose NO multiplier chain.
> What are the supposed problems in using a direct submultiple of the
> cesium resonance? It seems to me that all other techniques result in
> more phase noise there. I found the relationship 91.92631770
> mc·(137,075/126,008)=99,999,999.98992 cps=100,000.000--0.01008 cps.
> It is low by 0.1 ppb and therefore cannot be adjusted by C field
> current. The C field can only lower the frequency. There is another
> relationship that gives a higher frequency of a fraction of a part per
> billion which is easily adjustable. Perhaps HP was unaware that such
> a frequency exists.
> ---------- Forwarded message ----------
> From: Richard (Rick) Karlquist <richard at karlquist.com>
> Date: Thu, Jun 1, 2017 at 10:01 PM
> Subject: Re: [time-nuts] HP5061B Versus HP5071 Cesium Line Frequencies
> To: Discussion of precise time and frequency measurement
> <time-nuts at febo.com>, "Donald E. Pauly" <trojancowboy at gmail.com>,
> "rward0 at aol.com" <rward0 at aol.com>
> Direct multiplication to 9192 MHz isn't used
> by any manufacturer of any atomic clock that I
> know of, due to its well known disadvantages.
> I can state for a fact that it was summarily
> rejected by the designers of the 5060/5061
> (Cutler, et al). In the 5071, I (being the
> RF designer) also summarily rejected it.
> The architecture that is instead used is indeed
> complex and expensive as you say. It is
> also ACCURATE.
> On 6/1/2017 7:04 PM, Donald E. Pauly wrote:
>> The lock system on the HP5071 is complex and expensive. My plan to
>> improve the HP5061B is to to use a pair of third overtone crystals
>> running at 91.9 mc and 100 mc. I have come up with the magic numbers
>> to lock them together. This eliminates all multipliers with the
>> exception of the A4 board. The 12.61 mc synthesizer input presently
>> wastes half the microwave power produced by the 90 mc input in the
>> unused lower sideband. Therefore only half the 91.9 mc drive is
>> Eight bit ECL dividers in one package are available to perform the
>> necessary lock. When multiplied by 100 to the cesium resonance line,
>> the 91.9 mc frequency is a few cycles high so that C field currents
>> are reasonable. With crystal cuts for zero temperature coefficient at
>> 25°C, it is possible to get along without an oven. Room temperature
>> performance at 25°C±5°C is ±15·10^-9. Oscillator warm up time would
>> be measured in seconds.
>> Square wave modulation of variable frequency and amplitude shows
>> promise for reducing the noise effects of the beam tube. You can
>> smoothly change the lock time constant, deviation and frequency. This
>> would avoid the big disturbance of the HP5061B when you switch from
>> OPR to LTC. (OPR=operate with 1 second time constant, LTC=operate with
>> 100 second time constant)
>> time-nuts mailing list -- time-nuts at febo.com
>> To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
>> and follow the instructions there.
> time-nuts mailing list -- time-nuts at febo.com
> To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
> and follow the instructions there.
More information about the time-nuts