This is odd. I was told it was common practice to slow the frequency just a bit during heavy loads, then raise it to compensate during low demand (at night). This was done to keep syncronous motor clocks from gaining or losing time, and that power companies were very careful about maintaining an average frequency of 60 Hz. Maybe it was after this "experiment" that they started doing this?
I'm not sure it was ever common practice, but I don't really know. What was mystifying to me was how they were able to move the whole grid in a controlled fashion.
Someone mentioned the tendency for slower generators to be 'pushed' by faster ones. So perhaps there was a self-synchronizing effect across the grid. The freq. shift must have been quite slow. I've got a contact at Public Service Energy Group who would be in a position to know what happened.
Rotation of Synchronous motor shafts are sync'd to the frequency of the power source. Rob the motor of 2 cycles per second, times 3600 seconds in an hour, and pretty soon you have a fair number of revolutions! Let that go on for 6 or 8 hours, and you've got an error on the order of 7- 10 seconds.
It evidently depends on the number of poles in the armature, and the clock gearing as well, although I never had time to model the whole system.
I think the power company knows very well how we depend upon their 60 HZ being 60 Hz. I would think, however, that lowering the freqency would raise current on brown-out days. It sure decreases the efficiency of the transformers in the circuit.
It is like the politicians who got around the Constitutional requirement to have Congress declare war. Just don't call them wars! Actually, it isn't like that at all. Sorry.
Come on, this is supposed to be an engineering site, think basic electricity. If your generator is loaded down, it slows down. If your grid-tie lags the rest of the grid, you draw current from it. Your electric company buys power just this way. If you have excess power available, speed up the genny and pump the juice to the grid. Your electric company makes money just this way. This buy/sell scheme has almost trumped the goal of keeping a constant number of power cycles (at 3600 a minute) per day. All those millions of clocks? They don't really care if you are on time or not!
They do care a little bit, the IEEE, FERC and NAESB (North American Energy Standards Board) have very tight requirements for the frequency, voltage and other parameters for power quality. Without the standards it would be impossible to have reliable appliances or a distribution grid.
I lived in a neighborhood that was constantly browned out below the industry standard during heat waves. Each time the RMS voltage dipped below the minimum I called Commonwealth Edison to complain because I knew they would need to fill out plenty of documentation with the Feds (for negative reinforcement to be effective it must be immediate and severe). The real issue was load factor correction in a neighborhood that was built long before air conditioning. It only took a little over a week of complaining before they fixed the distribution problem.
First of all, a generator running at 61HZ connected to a generator trying to run at 59HZ will turn the slower generator into a motor. You can't have a national grid with generators running at different speeds (frequencies). I believe I read the expirement in the 60's to vary grid frequency was a typical government attempt at trying to controll the laws of physics. It was ultimately unsuccessful and caused great upset in many industries. Today the national electric grid represents an extraordinally accurate time base. Crystal oscillators before the late 60's or early 70's required constant power to keep the temperature stable and had a significant warm-up time before they could be relied upon to be accurate. When telephone voice and data volume overhwhelmed the available bandwidth of telephone lines, MaBell developed a good-stable non-oven controlled series of crystal oscillators that allowed much better (and cheaper) equipment to permit multiplexing without the need to add infrastructure.
Synchronous motors are very common. You see them on, e.g., bench grinders and drill presses and furnace blowers. The label plate always (here in the US at 60 Hz) shows the RPM as 1725 or 3450 rpm. That's a slight lag from the expected 3600 rpm--that design lag is what causes these motors to rotate; they are trying to "catch up."
Besides the label plate, you can recognize these motors because they are brushless: no sparking commutator on the armature, as you would find in a vacuum cleaner or power drill. Brush-type motors are called universal motors because they can operate on AC or DC. Synchronous motors (sometimes called induction motors because the field current induces a current in the rotor causing rotation) are, of course, AC only.
You may recall an article by Margery Connor a year or so ago, wondering how she could vary the speed of her drill press with a light dimmer. Because it was driven by a synchronous motor, she couldn't. It was 1725 rpm or nothing. Had it been a universal motor, she could have done so, since their speed and torque are related to the rms voltage applied.
those 1725/3450 motors aren't technically "syncronous" thought they follow the line frequency -- the slip varies sloghtly with load and the base slip varies slightly with design, particularly for motors w/ hard to start or impulsive loading.
The true syncronous motors (very low output power devices, as used in older electric clocks and timer gear motors) do sync up exactly to line frequency, and over long terms (days to weeks) the grid frequncy accuracy is VERY GOOD. in the short term it is more variable than say WWV, WWVH, CHU (Canada) or (better still) WWVB time signals, but for most applications this was not an issue.
The case cited here is an exceptional case, though 57-63 Hz shouldn't create issues w/ transfomers and induction motors as say 50 Hz could, but, when seconds are important, can cause the havoc the radio station had. Guess the grid was a little off on their high frequency corrections, but normally they wouldn't be so far off.
Now adays it's easy to come by *good* quartz controlled clock systems w/ WWWVB receivers allowing resync at night when VLF propagation is good across North America from the transmitter in Fort Collins, Colorodo.
Re: the grid business ISO NE has some interesting on line data and N.E. grid information at: http://www.iso-ne.com/sys_ops/index.html
I've designed dozens of clock-based appliances, and they all used line cross for a time base. Most of it is about money; if it plugs into the wall then line cross is free and a crystal costs money, a lot of money. Then there are some practical aspects of using line cross; the software can determine 50Hz or 60Hz using the microcontroller's internal oscillator, multiplexing the display can be synced to the line to prevent flicker with fluorescent lighting, if you know where line cross is you can phase fire triacs, you can "walk" relays by firing them at different parts of the linecycle to minimize relay wear and you know when the power has been interrupted long before VCC fails because line cross has ceased before the power supply bleeds off (handy for saving stuff in EEPROM and shutting the system down nicely).
Perhaps a subtlety that y'all didn't pick up on: THE ENTIRE NATIONAL POWER GRID EAST OF THE ROCKIES WAS INVOLVED IN THIS EXPERIMENT. There being no HV transmission lines over the rockies, the left coast was... left out.
The idea was to see how much fuel could be saved if you ran just a few Hz slower in peak times (rather than having brownouts).... and catching up in null times like 3 am. They may have cared about the clocks, or they wouldn't have tried to catch up. At least, I figued they tried, but missed it by 5-10 seconds.
No big deal for your kitchen clock, unless you wanted the coffeemaker to come on precisely when network news did. :)
The interesting thing to me is that I never saw mention of a final report on the experiment.
The station I worked for had Western Union clocks in the studios. They reset to the correct time once an hour and you could see the second hand either speeding up or holding at the same spot for 2-3 seconds each hour, around 10 minutes before the hour. The sync signal was sent via a dedicated phone line if I understand correctly. We joined the ABC Entertainment Network on the half hour and the clock was always within 1/2 second. (BTW, it was WCWA-AM/FM when I started after high school, the call on FM changed to WIOT and format went to progressive rock from elevator music before I graduated)
Since the network sent a tone 10 seconds prior to the start and went silent for the time before the news started, you could simply open the switch to the mixer board once the tone stopped and the news started (whether you were ready or not). We did also need an ID in that time period as well. I hit the net fairly well while I worked the board but sometimes... the post brings back memories.
The areas of Japan generally southwest of Kyoto are 60Hz, while the rest of the country is 50Hz. http://en.wikipedia.org/wiki/File:Power_Grid_of_Japan.svg It was the result of the US and England helping rebuild the power grid after WW2. Pretty incredible for a modern industrial nation to have 2 line frequencies! Back-to-back AC/DC/AC converters create the intertie between the two (dotted lines on the map in the link).
The US has 3 major, separate grids with a DC intertie in West Texas between them. But the isolation is not due to frequency differences. I believe the grids are kept in pretty close phase sychronization.
Since AC clocks in Japan had to have motors and gearing compatible with the line frequency, they perfected the ubiqitous AA battery-powered wall clock movement.
My dad grew up in the LA area, which was 50Hz until the late 1940's. I remember him talking about the power company taking clocks and converting or trading them for 60Hz operation.
A friend of mine got a Rolex for his birthday from his wife one year. After wearing itfor a few days he noticed it was running a few second slow. He returned it to the store to complain and the gentleman explained that he must have been mistaken and that every Rolex they sold was checked against the time standards at the U.S. Naval Observatory in Washington. My friend then explained to the man behind the counter that he BUILT the machine that the store was using to test the watches.
Hi Rob. Actually the "machine" is a telephone. The Naval Observatory (USNO) in Washington at the time had 27 atomic clocks of various types that were located around in closets, rooms, etc (to account for room temp variations). The time from each clock was averaged and fed to a rack mounted thing called a Nano Clock (I forget who made those) the output of the Nano Clock went to an announcing machine (of which I designed part of) that stored phrases like the words "one", "two", "u. s. naval observatory time is", etc. When you call the USNO telephone number a Weatherchron announcer seizes the line and on the next cycle that starts every five seconds (I think) assembles the words into a complete phrase that sounds resonably natural.
The "voice announcer" number is on the USNO website and anyone can call it (including jewelers). The machine has been in service for several decades pretty much non-stop.
The machine was designed way back when UV erasable EPROMs were common and the voice was stored in a bank of Eproms in 8 bit telephone CODEC format.
The man who owned the company died a couple of years ago and was one of the smartest people I've ever known, and one of the best friends I ever had. He knew (as far as I could tell) just about everything about everything. The DAY AFTER the space shuttle disaster, he told me that the problem was the O-ring seals on the engine. He studied and read just about everything technical he could get his hands.
He would some times look at something I had designed and ask "If this was an airplane would you fly on it?" at which point I would usually make some changes.
One of the early manufacturers of synchronous clocks would send salesmen to the operating engineers of power plants and give them one of their clocks, with the advice "If the clock is slow, just go in and speed up the generator 'til it catches up." Very clever marketing.
The same company went on to become the manufacturer of the first commercially successful electric organ which used a syncronous motor to drive its tone generator. Examples of these nearly 80 years old (I own a 1947 model myself) are still running great -- and in tune!
@Eric Tucker - In addition to salesmen giving clocks to power stations, there was a more formal version: the Warren Telechron Master Clock: A very accurate clock movement driving one hand on a clock dial, while a synchronous motor drove an additional hand, so they just had to adjust the generator frequency to keep the one hand on top of the other.
El-Cheapo clock radios take the 50/6o Hz from the input transformer secondary and use is as a clock frequency instead of a crystal. The Clock/display driver IC usually has a 50/60Hz select pin which is jumpered on he PCB according to the market. These clock radios are notoriously inaccurate, because 50Hz is never quite what it says. It's worse if they go into backup/power out mode using a backup battery because then they rely on an R-C circuit. I understand all this happened 50 years ago, but even then, wouldn't a broadcasting station rely on at least on independent accurate clock source?
What would you have had, 50 years ago, for an alternate clock source? Nothing was available. The pennies-cheap 32,768 kHz crystals and IC counters and LCDs that enable $1.00 watches and giveaway promotional stick-on clocks had not yet been cheaply mass produced. Timepieces were notoriously expensive and inaccurate. Think of the Rolexes and Omegas of the time.
I take issue with your statements regarding the "El-cheapo" clock radios. I have a couple of these (all under $US 10.00), from the usual Asian sources. They rely on the 60 Hx here in North Carolina, which is dead accurate. And if we do have a power outage, the 9-volt battery drives one of those 32,768 kHz crystals, not an RC circuit. I reset these exactly twice a year, at the beginning and end of the Daylight Savings Time period.
I would expect a commercial broadcast station to invest in the required equipment to maintain accurate time. After all, they already have an accurate clock source to drive the transmitter, don't they? This is a commercial enterprise we're talking about, not a HAM operator.
Hard to believe your clock radio normally uses the mains frequency as a time base in preference for its' on-board watch crystal. Which do you think is more accurate?
The mains are more accurate in the US. Those crystals are very accurate when they are on your wrist, held to constant temperature. Less so in an outdoor transmitter hut or a room where air conditioning might not be present or cycles on and off.
Want to test it? Get two of those cheap crystal-controlled watches. Set them to the same time. Leave them in the same environment for a few days to assure they track. Then put one in the refrigerator or freezer and the other in a warm spot, e.g., in proximity to an incandescent light bulb. Check them after two days and notice how far apart they have become.
The mains where I live are not accurate to 20ppm, which is what I have come to expect from a 30$ Casio watch. But I don't use them, I wear 150$ Japanese automatic (mechanical) watches which are accurate to about 10 seonds a day. It's amazing that such intricate engineering can be sold retail at such prices. Anyone who wants a stable frequency uses a crystal oven, which uses a simple feedback controlled heater element in a sealed enclosure to maintain the crystal at a constant frequency - that trick is as old as WW2.
Your proposed "experiment" could be flawed (depending).
Putting two oscillators of the same frequency (mechanical or electronic) near each other will demonstrate a strange affect. They will sync with each other. This gives an additional edge to the two time pieces tracking together in the same environment.
The experiment would be valid if the two time pieces are verified to track closely (but not exactly) with same temperatures - but physically separated enough to eliminate this effect.
The syncing effect is more pronounced with certain types of electronic oscillators. But can be demonstrated at some level in nearly all types.
I suspect using watches with metal cases and a modest distance would provide most of the required isolation to make the experiment valid.
The power systems east and west of the Rocky Mountains have very few ties between them. To avoid problems with massive power flows that would take a tie line out of service, the connections are made through dc connections. At each end of the line is an ac to dc converter station. The operaters agree on how much power is to be transmitted and set the equipment to accordingly. The line can operate as a constant power source or sink depending upon the needs of the utilities.
Synchronous motors and induction motors are not the same. Synchronous motors run at a constant speed that is determined by the power line frequency and the number of poles in the motor. An induction motor's speed does change with loading. Induction motors are less expensive to build than synchronous motors and a quite suitable for numerous tasks like fans, washing machines, and compressors. For 1/4 horsepower to 2 horsepower motors, you may see a No-Load speed around 1795 rpm with the Rated-Load speed of 1750 to 1725 rpm.
The tiny synchronous motors in clocks were used for many, many years. Only when production costs of the electronic clocks fell to near the costs of the motor driven clocks and the public demanded features not readily available in the motor driven clocks did the electronic clocks take over. Even today, some electronic clocks use the power line as the primary time standard with a cheap crystal as a backup for power interruptions.
The power line is poor short term time standard but can be a very good long term standard. Other than the experiment described in the article, the utility maintains a time accuracy of plus or minus 180 cycles (3 seconds). (Utility standards assess time error as a number of cycles.) Over short period, say an hour, a 3 second error is almost 0.1 percent. Over a year, that 3 second error is slightly less than 0.0006 percent.
Synchronus motors and induction motors are not the same. An induction motor runs behind the synchronus speed, a syncrhonus motor is locked to it. An induction motor is the most common motor in the industrial world.
An induction motor (squirrel cage motor) does not have brushes. A synchronus motor does have brushes, and requires a dc voltage on the rotor coil to lock it to the system frequency. Since it does have brushes, and requires a separate power supply for the rotor, they are actually not very common. The generator at the power station is obviously syncrhronus. As noted in this story, some clocks are as well. In 28 years, I think generators and an old motor in the power lab at the University of Wyoming are the only large syncrhonus motors I have ever seen.
The poster is correct that the small throw away motors in vacuum cleaners, mixers, etc are universal motors.
At every power station I have ever been in, there is a meter generally on the back of the board where they track the frequency over a 24 hour period. Generally, the deviation is 10ths or 100ths of a hertz. Since the normal deviation is so small, I think the author showed pretty good trouble shooting skills finding this problem
krruss wrote: "A synchronus motor does have brushes, and requires a dc voltage on the rotor coil to lock it to the system frequency. Since it does have brushes, and requires a separate power supply for the rotor, they are actually not very common."
Wikipedia suggests that there are several varieties of synchronous motor. One type, DC-excited, is produced in sizes greater than 735 W. Three other types of non-excited motors, reluctance motors, hysteresis motors, and permanent magnet motors. do not use DC excitation and do not have brushes. They rely on either induced rotor current or permanent magnets in the rotor. Each has different self-starting characteristics.
My parents had a synchronous clock on the mantel when I was growing up in the 40s & 50's. It kept perfect time, but whenever we had a power outage the clock had to be reset and a small knob on its back had to be rotated to get it going. It was a common prank to hold the knob to stop the clock at noon, and then spin the knob in the opposite direction to make it run in reverse. We could tell what time it was by mentally transposing the numbers on its face, but it used to drive my old man crazy.
Great post Jim. Believe it or not, I had a similar experience with an ERC II (electronic range controller) when I worked for GE. We absolutely could not understand why our on-board clock was losing time. Not much time but it would lose approximately 10 to 15 seconds per day. We contacted Robertshaw and they were baffled also. One of our engineers, a EE type, had come to us from Louisville Power. He made several phone calls and discovered the problem was just as you mentioned an input of 50 to 55 Hz instead of 60 Hz. That solved our test problem. Again, great post
Fifty-six-year-old Pasquale Russo has been doing metalwork for more than 30 years in a tiny southern Italy village. Many craftsmen like him brought with them fabrication skills when they came from the Old World to America.
Focus on Fundamentals consists of 45-minute on-line classes that cover a host of technologies. You learn without leaving the comfort of your desk. All classes are taught by subject-matter experts and all are archived. So if you can't attend live, attend at your convenience.