[time-nuts] A Research Proposal

Mon Jul 8 12:19:49 UTC 2019

 > I hope some will accept the technical challenge of gathering data -- 
which is fairly minimal.

Andy,

It may have been before you joined time-nuts, but some years ago we 
compared mains phase/frequency between Albuquerque, NM and Seattle, WA, 
which in spite of being some 1500 miles apart are on the same grid. 
Another test was between two nearby locations, Bellevue, WA and Redmond, 
WA. All data was taken by timing zero-crossings to the microsecond with 
a picPET. You'll find the results fascinating:

http://leapsecond.com/pages/mains-cv/

I'm happy to repeat any of this analysis using recent data from Hal 
(CA), and you (Bellingham, WA), and me (Bellevue, WA).

A third test which is interesting is comparing independent mains 
phase/frequency between different rooms of the same house. In general 
you want to understand local house effects, then neighborhood effects, 
then city / country effects before you jump into interstate grid 
effects. It's similar to establishing baselines and noise floors when 
you make precision clock measurements.

/tvb

On 7/7/2019 11:49 PM, Andy Backus wrote:
> To clarify:
>
> My research proposal is for data to be taken within the same interconnection.
>
> It does not care about frequency.
>
> It does not care about Time Error.
>
> It only seeks to characterize the phase differences between the power line signal presented in regions of the distribution system separated by significant geographical distances.
>
> 180-degree phase reversals and three-phase transformations along the chain of information are easily distinguishable.
>
> Thanks for your interest.  I hope some will accept the technical challenge of gathering data -- which is fairly minimal.
>
> Andy Backus
> Bellingham, WA
> (Western Interconnection)
>
>
> ________________________________
> From: time-nuts<time-nuts-bounces at lists.febo.com>  on behalf of Dave ZL3FJ<2c39a at silverbears.nz>
> Sent: Sunday, July 7, 2019 9:23 PM
> To: 'Discussion of precise time and frequency measurement'
> Subject: Re: [time-nuts] A Research Proposal
>
> I have here a pair of instruments that were part  of a system used  at one time  in a power station here in NZ  to control the time error in one  part of the national grid. It controlled the selected generators and provided a real time display of the time error between a reference standard and the 50 Hz mains frequency.  The system comprised an HP 5280A reversible counter  with two inputs, one from the  mains 50 Hz as generated and one from the reference standard. These two inputs were arranged to add  counts from one input and subtract counts from the other, such that the counter displayed zero while the generated 50 Hz was accurate. Offsets from 50Hz were displayed as positive or negative counts. The reference input was derived from an HP 105A quartz oscillator and the system included provision to manually  synch that to the national standard time standard on an as required basis. The output of the 5280A counter drove an HP 6933B D/A converter, the bi-polar DC output of which was used (both magnitude and sign)  to control the governors on some of the hydro generators. Dual HP 5321B clocks were used to display TOD from both sources.
>   The 6933B is complete but the 5280A counter has been partly disassembled.  The 5321Bs never got this far-neither did the 105A- who knows, it might still be being used  as a reference!
>
> DaveB, NZ
>
> -----Original Message-----
> From: time-nuts [mailto:time-nuts-bounces at lists.febo.com] On Behalf Of Bill Hawkins
> Sent: Monday, July 08, 2019 06:48
> To: Bob via time-nuts
> Subject: Re: [time-nuts] A Research Proposal
>
> Group,
>
> We've discussed this before, but maybe it needs to be said again.
>
> Line frequency is not constant.   There is no master PLL.  Approximate frequency is maintained by a central power dispatching office in each of the four (?) regions tied together by their power distribution grid.  The dispatcher's goal is to create the same number of cycles of AC each day.  IIRC, power is bought and sold by the number of cycles generated.   As the daytime load increases, the generators slow down a bit.  Note that it is not possible for each generating station to control its frequency, as that would not be stable.  Instead, the dispatcher asks various plant operators to generate more or less steam (or water flow) in order to increase the frequency.  When the load drops at night, the generators speed up a bit, and steam has to be reduced. At the end of the day, so to speak, the number of cycles generated is very nearly equal to the number generated if the line frequency had been steady at 60 (or 50) cycles per second.  Synchronous clocks stay accurate although they may be off by a few seconds as dispatchers scramble to get enough steam to keep up.
>
> So yes, you can get phase data within a region but you must compensate timing data as the frequency varies.
>
> The regions are connected to each other for purposes of power sharing with DC transmission lines.  These use inverters to convert between AC and DC. The AC frequency is controlled by the grid that it is tied to.  Phase angle can be changed to change the amount and direction of the power transferred.
>
> So no, you can't compare data from different regions, unless you want to know which way DC power is flowing.
>
> I hope this was informative.
>
> Bill Hawkins
>
>
> On Thu, Jul 4, 2019, at 2:00 PM, Andy Backus wrote:
>> Historically, and even today, the steady frequency of AC power has
>> been used for timekeeping.  So there may be interest here in the
>> following research proposal:
>>
>> Within a given power distribution grid, several observers as widely
>> separated geographically as possible, time stamp the first two zero
>> crossings of the power line after each UTC second – over the course of
>> 24 hours (86,400 pairs of data).
>>
>> Popularly conceived, all the components of a power distribution grid
>> are phase locked – though, of course, power is taken in and out by
>> varying degrees of lead or lag.  Frequency is maintained by a constant
>> balancing act between load and generation.
>>
>> Typical power distribution grids, however, are sized on a scale of
>> thousands of miles.  “Locking phase,” then, is problematic simply on
>> the basis of the limits of information transmission rate.  Even at c,
>> every 1000 miles takes 5 ms, which represents a third to a quarter of
>> the period of the AC power waveform.
>>
>> Many interesting phenomena might result from that reality, which
>> suggests a certain constrained flexibility over large distances –
>> almost as if the system is like a large lake of viscous liquid.  When
>> there are local disturbances such as rapid load changes or sudden
>> generation adjustments, for example, it is quite possible harmonic
>> ripples could be propagated through the system.
>>
>> Such effects could be observed by comparing phase data across
>> significant distances within a distribution grid.
>>
>> Andy Backus
>> Bellingham, WA
>> USA
>>
>> ________________________________
>> From: time-nuts<time-nuts-bounces at lists.febo.com>  on behalf of Thomas
>> D. Erb<tde at electrictime.com>
>> Sent: Thursday, July 4, 2019 5:23 AM
>> To:time-nuts at lists.febo.com
>> Subject: Re: [time-nuts] 60 Hz frequency and phase measurement
>>
> snip
>> I had a recent tour of a power station - the operators had no idea the
>> output was synchronized to a time standard - they just synchronize
>> with the local grid.
>> Thomas D. Erb
>> p:        508-359-4396
>> f:        508-359-4482
>> a:        97 West Street, Medfield, MA 02052 USA
>> e:tde at electrictime.com
>> w:www.electrictime.com<http://www.electrictime.com<http://www.electrictime.com<http://www.electrictime.com>>
>> Tower & Street Clocks Since 1928
>>
> _______________________________________________
> time-nuts mailing list --time-nuts at lists.febo.com  To unsubscribe, go tohttp://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
> and follow the instructions there.
>
>
>
> _______________________________________________
> time-nuts mailing list --time-nuts at lists.febo.com
> To unsubscribe, go tohttp://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
> and follow the instructions there.
> _______________________________________________
> time-nuts mailing list --time-nuts at lists.febo.com
> To unsubscribe, go tohttp://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
> and follow the instructions there.
>