measuring frequency

Well, God bless.

Cheers

Phil Hobbs

Depends for what. SDRs have really sucky phase noise, which I often care about. And they only cost 1000 times as much as a superhet!

Cheers

Phil Hobbs

"How do I do this thing where you..."

"Well I've done that a few times, you have to..."

"DONT YOU tell me what to do you're not the boss of me"

Frequency is free! It's voltage that takes energy.

Since you have coal, you could use it, or frack for your own NG. Or be green and cold in the dark.

There was a great bumper sticker in Louisiana:

LET THE BASTARDS FREEZE IN THE DARK

Am 07.03.2018 um 23:01 schrieb John Larkin:

Funny. I wonder where you get your disinformation.

I can't remember having seen a blackout ever.

In fact, the first half of 2017 was the first time that Germany imported NO electrical energy, not on a single day.

Germany exports electricity in the volume of 3/4 the capacity of all its nuclear power plants, of which there still are plenty.

Last year, 33% of all electrical energy was green. No need for fracking, and if the trend continues, Putin can shove his new gas pipelines into, yes. Just when they will be ready, finally.

Gerhard.

That's the (arguably) best approach. The 'isolation' is easy (no need for a transformer, just capacitive put-wires-nearby is good for CMOS inputs). The 'zero-crossing' can be quite noisy (remember X-10 modules, which inject carriers at the zero-crossing times?).

There's undoubtedly lots of Labjack-style modules with interrogatable counter sections. At 50 Hz, a day's worth of cycles is about 22 bits of count, plenty of margin to get a part-per-million resolution even with slow USB.

Better would be side-by-side counters for the AC pickoff and for a reference clock, with appropriate latch/display logic; you could just push a button and see the numbers.

Two of these counters and a bit of logic would bypass the computer and soundcard entirely

Am 08.03.2018 um 00:06 schrieb Gerhard Hoffmann:

Ich mache mal die Ingrid!

Just for the fun of it:

It didn't take more than an old 9V AC wall wart and a voltage divider to 4 Vpp. The counter can do 12 digits in a second.

Now, later in the night, the frequency has risen to > 50.02, they have to catch up.

regards,

Gerhard

We were discussing the UK, not Germany.

How much does a KWH cost in Germany?

-----------

There are several sites like this one for grid frequency.

.... Phil

Y'er welcome and sorry for the delay. Paying work comes first.

I downloaded the power line frequency data from the above URL and dumped it on my web pile: It's 13.2MB and consists of everything on the page in 15 minute intervals for 4 years. I then fed the CSV file into LibraOffice 5.1 and found that there were 115,000 rows of data. I removed some columns leaving date, time, grid demand, and grid frequency and generated a graph. However, working with large XML data files is slow, and I didn't see anything unusual in the 4 year graph when it finally appeared. So, I started shrinking the number of rows, until my computer became more responsive. Then, I discovered some problems with the Chart program, so I switched back to MS Excel 2003.

The spreadsheet, which covers Jan 1, 2018 to Mar 7, 2018 is at: A JPG of the graph is at: I'll throw in some more stuff as time permits.

The graph shows daily excursions from 49.900 to 50.100 Hz. Previous graphs showed a slightly larger range, from 49.850 to 50.150 Hz.

Frequency and other data is available from the grid operators. I don't know much about European power but here are some of the US ISO (independent service operator) sites:

that will only tell you how well your soundcard clock aligns with the mains.

youll be able to compare the skew between samples that way, also you could feed the PPS from the GPS into the other sound-card channel...

there's a web site...

They drop both. The difficulty in turning the generators round increases with the load so if there isn't enough torque available they slow down and the voltage decreases proportionately. The rate that the wires in the generators are cutting magnetic lines of flux decreases.

It is quite a good demo to take a stepper motor and a large pulley driving a small pulley to make electricity and have people try to light either a high efficiency 1W LED or a 5W low voltage filament bulb.

It is more impressive with an alternator and a stationary bike with larger scale mains lighting. Teenagers rapidly get tired of peddling like crazy just to keep a 60W bulb glowing dimly and remember to switch house lights off for a while after trying. LED is easy by comparison.

Changing frequency to match the total load is the natural way that the generators stay in sync once they have joined the grid. Power delivered to the external load going as the square of the voltage out.

No. You missed the "including timestamps". The timestamps time (some of) the samples to the real-time clock. The calculation of the zero crossings would be relative to those timestamps. The sample clock is irrelevant.

It isn't required. I have done microsecond-accurate soundcard output using ALSA and it works just fine without sync channel and without locking the sample clock to a reference (as I originally intended to do).

I have now used the data from the website instead of measuring it myself.

As I mentioned, I first averaged the frequency for all the measurements on a single day (after correcting some errors in the file at 2015-03-20) and then worked from those daily averages. This because while there are fluctiations during the day, the long-term value should be more precise. But currently it isn't.

The cost always depends a lot on where you define your system boundaries. The superhet as a PCB is different from a complete radio with power supply and case, and similarly for an SDR you could consider only the cost of the front-end and leave out the cost of the processor and software, because the processor is already required for other things and the software is free or has fixed development cost.

That way, you can now make a receiver for the 25-1600 MHz range using only a $10 RTL2838 DVB-T USB stick plugged into a PC or Raspberry Pi that you already have. With free software it can receive "anything" in that range within 2MHz bandwidth. That is hard to beat with a superhet.

Sure, but my understanding is now that it does not work that way. That is, there is no "lock" between the line frequency and the generator RPM for those generators, like with a stepper motor. The output frequency (or better: the phase) "slips" proportionally to the load.

The frequency stays in lock automatically because with increased load the slip increases. When the phase of the generator would get ahead of the grid it would dump more power into the grid, the slip would increase and the phase difference decreases, even without an RPM change.

No, the large coal and nuclear generators with steam turbines are synchronous, i.e. a big magnet (electromagnet) rotating inside some coils that produce the output. The shaft angle is always the same when the voltage of a given phase is positive. The reason why the frequency is not exactly what is desired might be for example that arbitrarily large amounts of torque are not always available in order to turn the generator at exactly 3000 RPM (due to that being pretty much the definition of something that is hard work to do), or because doing so would exceed the current rating of the windings or cables and damage them, etc.

If you slightly drop the voltage of the generators, there are automatic tap changing transformers between the generators and the consumers so that the voltage at the consumers will stay the same, but the current at the generators will then increase, so dropping the voltage will not make the generators easier to turn. Dropping the voltage supplied to consumers generally wouldn't help for a similar reason - thermostats of ovens would stay on more of the time, switched mode power supplies would draw more current, and so the power supplied to consumers would not drop as much as you might expect. Old filament lamps would use less power, and there might be somewhat less iron loss in some badly designed transformers and motors, and un-controlled heaters would use less power, but wiring losses would increase. There are also legal minimum voltages that are supposed to be supplied, and probably that would constrain any power savings that could be achieved by this method. Also, the people who could do this probably get paid less if consumers use less power, so they might find it fairly uninteresting as a topic of investigation.

Some wind turbines are asynchronous, but then the entire power output has to be put through a frequency changer built with semiconductors.

You can make an asynchronous generator using an induction motor, and some people do that for micro-hydroelectric systems. The main motivation is that small scrap induction motors are more commonly found than scrap synchronous motors or generators.

Big mains generators are all synchronous machines. There is no slip. Allowing a synchronous machine to change phase w.r.t the network would change the power factor, make it reactive, which is wasteful, so that is avoided.

Power injected into the network is adjusted by changing the field excitation. The setpoint is derived from the frequency error. There are lots of limits and safeguards built into the process.

Jeroen Belleman

An SDR with a sample clock with a lot of phase noise will have all kinds of reciprocal mixing issues. The same applies to a superhet with a dirty local oscillator.

Some chips use stupid designs such as using noisy PLLs to move the clock frequency higher. A cleaner approach would be using an overtone crystal to generate the required frequency directly. The drawback is that you would need an LC circuit to force the correct overtone.