Grid Frequency Experiment

Status
Not open for further replies.
If one google's "Grid Frequency Experiment", all results originate from newspapers or TV outlets. So far, there's little info besides, 'your clock may not be correct anymore' or more in that vein.

Is this a joke or is there any concrete info available that details the limits of variation in frequency? (that's besides the notion that if two generators are at two slightly different frequencies, and you 'parallel' them, the 'slower' one will shift it's load to the 'faster' one until frequencies are the same.)

For example, if Grid Frequency was allowed to drop from 60Hz to 50Hz, my Fire Pump would not turn at the 'design' rpm. Air Handlers that are older squirrel cage rotor motors would turn slower, pushing air a little less. Not sure what VFD's would do, but think they would turn their motor at the dialed in Hz/rpm.

Thanks for any info.
 
generators and UPS's

generators and UPS's

There are 8 generators on campus, all Emergency or Standby - so I'm not worried about them. However, the IT dept has a 'large' UPS and i'm not sure of the 'trip' settings that would disconnect it from the campus grid - phase loss, low and high voltage and frequency are setable. If we were 'Co-Gen' with the utility, I'd start to worry.
Hopefully, TVA frequency will stay within 59Hz to 61Hz.
 

gar

Senior Member
Location
Ann Arbor, Michigan
Occupation
EE
110625-2225 EDT

See my post #12 from 2008 ided by 080820-0757 EST at
http://forums.mikeholt.com/archive/index.php/t-102830.html
for frequency measurements I made from the local power line.

The whole eastern part of the United States will be close to that. Different parts of the system can not deviate very much.

The long term average over a year is held very accurate by reference to the NBS (NIST) time standard.

Under normal conditions, unless there is a system failure, you won't see much frequency variation. But you must use a correct measuring technique to get accurate results.

.
 

gar

Senior Member
Location
Ann Arbor, Michigan
Occupation
EE
110625-2236 EDT

Now I have read one of your references. Probably mostly scare.

There is no detailed information and therefore no way to know the effect. Many systems today do not depend upon power line frequency for master clock information. So called "Atomic Clocks", time reference from the Internet, and your cellphone come from NIST one way or another. For a long time my old computers have had a real time clock with its own crystal oscillator that provides the real time information. This does not drift too much so only rarely do I update the base.

A time logging device that I make has its own internal crystal based clock and only needs updating once in a while. This clock has an internal battery that continues its operation when no AC power is applied. Dallas Semiconductor the manufacturer of the IC claims that the internal battery can maintain clock operation for 10 years without external power.

It is very likely that power frequency would not become too loose.

Where is there some direct information on this experiment?

.
 

don_resqcapt19

Moderator
Staff member
Location
Illinois
Occupation
retired electrician
It sounds like they are going to eliminate the now required time correction. With the required correction, a clock that uses the grid frequency will be very accurate over long periods of time.
There is information here.
 

broadgage

Senior Member
Location
London, England
I would not expect variations in grid frequency to be of any concern in a developed country as they would be very small indeed.

Here in the UK where we use 50 cycles, the legal limits are from 49.5 cycles to 50.5 cycles, and the normal operational range is from 49.8 to 50.2.

I would expect that USA tolerances are broadly similar.

The frequency of small portable generators varies much more, but is still fine in practice.
Large modern generators are nearly as good as the grid, with some actually more stable.
Some large mobile generators have a choice of 48 cycle or 50 cycle output, both being accuratly maintained.
 
NERC info

NERC info

Thank you Don, NERC link is what I was looking for!

It appears that devices/operations depending on TEC for long durations (weeks to months) will feel the effects. Otherwise, not noticeable. I'll be following this as it proceeds, as well as the NERC System Protection Initiative ( http://www.nerc.com/page.php?cid=5|308 )

Again, thanks to Gar and Broadgage as well.

Another distribution connected story, relevant to the future of 'blackout' prevention: http://www.publicpower.org/Media/magazine/ArticleDetail.cfm?ItemNumber=31745
 

gar

Senior Member
Location
Ann Arbor, Michigan
Occupation
EE
110626-1058 EDT

I have glanced at the NERC notes. and I do not know what they are saying.

Other than the heading "TEC Elimination" I did not see any statement that the intention is to eliminate long term cumulative time error correction. What was provided was some information on time periods of frequency deviation beyond limits.

Why can you not maintain long term moderate time accuracy, for example +/-10 second maximum deviation over periods longer than a month. It is a matter of power input. It seems they want to loosen the limits on short time frequency variation. But does that also mean long term can not be reasonably good?

If you consider the 59.95 Hz lower limit that is a deviation of 0.08333 %. If I had a 1 hour timer, had this as the input, and I designed for 60.0000 Hz, then I have an error of 3 seconds in one hour.

Their data does not really tell me what kind of short time timing errors I would encounter today under present rules, or in the future under relaxed rules. My data that I referenced above is quite good.

Even today or yesterday if I wanted high timing accuracy I would not depend upon line frequency.

For clocks around the house I have several "Atomic" based and several line frequency based. The "Atomic" are quantized to 1 second and synced at least once per day. The line based have a display quantized to 1 minute, but internally probably are quantized to 1/60 second, and manually set and therefore not more accurate than 1 minute or so.

My cellphone is quantized to the minute on the display, and quantized better than this internally. At this moment the minute rollover between the cellphone and the "Atomic" was within 1 second.

For an industrial type timer that is line frequency based I would think that an important criteria would be cumulative error over some time period.

I am using the word timer to mean something different than a clock. If a device is controlling something thru a daily cycle, then I want to call this a clock. This clock device (possibly called a timer in a catalog) should not have its time of day shift from 8 AM on Dec 5, 2010 to 7:45 AM on Dec 5, 2011 because of cumulative frequency errors.

Another area of consideration is a disk record player, or cassette player. These are line driven devices and a change of frequency will produce a pitch change. Here moderately short time variations can be a problem as well as long time variations.

.
 

chris kennedy

Senior Member
Location
Miami Fla.
Occupation
60 yr old tool twisting electrician
If a device is controlling something thru a daily cycle, then I want to call this a clock. This clock device (possibly called a timer in a catalog) should not have its time of day shift from 8 AM on Dec 5, 2010 to 7:45 AM on Dec 5, 2011 because of cumulative frequency errors.

I have ATS's in govt buildings set to exercise the generators under load when support staff arrives in the morning but before all the computers and such get fired up. I can see the scenario described above as being an issue.
 

hurk27

Senior Member
I have ATS's in govt buildings set to exercise the generators under load when support staff arrives in the morning but before all the computers and such get fired up. I can see the scenario described above as being an issue.

I would believe the controls (including the timer) are powered through the DC buss on the generator, I know Kohler and Generac both are, since the timer has to keep ticking away even if the generator is stand alone it must have its own reference oscillator as does almost all timers fed from a DC source, from digital clock radios to TV's micro-waves even the clock on a my stove has a small DC power supply.

If it has a DC supply it has its own time reference.

I don't see this change any more of a problem then power and frequency fluctuations one might experience from running on a generator for a long period, which in a few cases I have.

But I am seeing from a few news sites the response of people thinking the sky is falling, and I think it will get blown out of portion before its over.
 

gar

Senior Member
Location
Ann Arbor, Michigan
Occupation
EE
110626-1317 EDT

Don:

Thanks for the second reference. Very good historical discussion.

Prior to government legal requirements it appears that Time Error Correction was not a system reliability problem. One sentence in the whole report seemed to imply this. But now with government rules it is apparently a problem.

There are a lot of qualitative statements made in the document, but no data to prove those statements.

There may be good reason to start to phase out equipment that requires a stable frequency and small cumulative time error.

.
 

gar

Senior Member
Location
Ann Arbor, Michigan
Occupation
EE
110626-1329 EDT

hurk27:

A clock run from a DC source does not mean that that DC source remains when AC power is lost.

The original IBM PC had a pseudo real time clock so long as power was on and the processor was running. Only later, possibly with the first XT, did a battery backed up clock come into existence from IBM. It was a very bad design with short battery life. IBM a power hungry Motorola chip. Had IBM used an existing Dallas Semiconductor product the results would have been much superior.

.
 

hurk27

Senior Member
110626-1329 EDT

hurk27:

A clock run from a DC source does not mean that that DC source remains when AC power is lost.

The original IBM PC had a pseudo real time clock so long as power was on and the processor was running. Only later, possibly with the first XT, did a battery backed up clock come into existence from IBM. It was a very bad design with short battery life. IBM a power hungry Motorola chip. Had IBM used an existing Dallas Semiconductor product the results would have been much superior.

.

And where did you get that from my post?

A clock on a DC supply will will have to have its own reference oscillator or will have to source it out such as those in atomic clocks, I have yet to find a electronic clock design the uses the AC 60hz power line as a reference when supplied via DC, even my simple alarm clock I have in my bedroom has a oscillator in it, and many of these are battery backed up so this is a must, this has nothing to do with whether or not a clock will keep working if you remove the power to it:?
 

gar

Senior Member
Location
Ann Arbor, Michigan
Occupation
EE
110626-1606 EDT

hurk27:

I got it from here
If it has a DC supply it has its own time reference.
I have two digital clocks that obviously operate from a DC source, but get their timing information from the AC line. These do not have battery backup, and must be manually reset on power loss. These clocks are probably 40 years old.

Both my microwave oven and regular oven, 10 and 20 years old, require manual reset on power loss.

.
 

Besoeker

Senior Member
Location
UK
FWIW, the clock in my car is accurate to the second. It is precicely in time with the "pips".
http://en.wikipedia.org/wiki/Greenwich_Time_Signal

It updates for the fall back/spring forward. And no grid connection.

Much equipment these days is designed for the international market thus clock frequency cannot be tied to grid frequency.

That said, tight control of grid frequency has to be maintained.
 
Status
Not open for further replies.
Top