Everything should have a capacitor battery backup

[Red Squirrel]

I remember hearing in college that the 60Hz frequency has to be maintained pretty accurately to ensure that there's no problems with multiphase lines, or sync issues caused by linking together multiple power plants via the grid, or some such thing.

That would make sense. In fact I've always wondered how they do it with people who sell energy back to the grid. You can't just plug in your inverter right in and assume it will be in phase. There must be some special syncing equipment involved.

Someone was telling me he once was told from a friend who worked at a power plant that he was about to bring a turbine online, and he had to flip the switch at the right time so that it's in sync, and he flicked it too fast and it was 180 degrees out of sync. Now imagine this massive turbine, the shaft being a couple feet diameter, suddently turning into a motor turning the opposite way the water is forcing it at, at lightning speed. This shaft broke like a straw. I would guess that today such processes are fully computer controlled so there's less chance of human error.


And no, I would not put a giant transformer in a server rack. Something about magnetism and servers don't play very well.

 

[Jeff7]

That would make sense. In fact I've always wondered how they do it with people who sell energy back to the grid. You can't just plug in your inverter right in and assume it will be in phase. There must be some special syncing equipment involved.

Yup, you need a "grid-tie inverter" for that.
1) It can ensure smooth switchover from line voltage to battery power.
2) It ensures that power going back into the grid is clean and synchronized.
3) It ensures that, if external power is down, it does NOT feed power into the grid. One of the last things a service tech wants to find is that a supposedly dead line is quite thoroughly electrified, courtesy of your battery bank.

Someone was telling me he once was told from a friend who worked at a power plant that he was about to bring a turbine online, and he had to flip the switch at the right time so that it's in sync, and he flicked it too fast and it was 180 degrees out of sync. Now imagine this massive turbine, the shaft being a couple feet diameter, suddently turning into a motor turning the opposite way the water is forcing it at, at lightning speed. This shaft broke like a straw. I would guess that today such processes are fully computer controlled so there's less chance of human error.

I...I'd think that the timing has to be more accurate than "flick a switch at the right time." The line is at 60Hz, and to get it in sync, you'd need to have even faster reaction time than that.
I could be wrong, but that seems like something's missing from that story - or else I'm missing something.

And no, I would not put a giant transformer in a server rack. Something about magnetism and servers don't play very well.

What do you think is stepping down the power for those servers?
Keep everything properly shielded, and it should work.*



* Not responsible for data loss or infertility.

 

[Leros]

Apps like that exist for Windows, which allow it to sync to a timeserver of your choice, and I'd expect that apps would exist which allow any PC to act as a timeserver for a LAN.

As for a standalone clock......I don't know, I think you can use TCP/IP communications with some PIC chips or other microcontrollers. I assume you could use that to do basic NTP protocol...stuff.

Another option would be to use a GPS module. GPS satellites use atomic clocks for timekeeping, which is critical to ensure accuracy of their positioning calculations. At the simplest, a GPS module can output one pulse per second for an external RTC to use; alternately, they put out a sentence, GPRMC, which includes time and date information (GMT).

Or you could just use the NIST radio station like lots of clocks do.

http://tf.nist.gov/stations/wwvb.htm

 

[bobdole369]

Nope - would need a significant amount - somewhere around a farad to keep them up for 10 minutes. But a simple 4700 ufd cap would keep most things alive for about 15 seconds - and would handle most everything besides teh perhaps thrice yearly power outages that last longer than that.

 

[bobdole369]

Someone was telling me he once was told from a friend who worked at a power plant that he was about to bring a turbine online, and he had to flip the switch at the right time so that it's in sync, and he flicked it too fast and it was 180 degrees out of sync. Now imagine this massive turbine, the shaft being a couple feet diameter, suddently turning into a motor turning the opposite way the water is forcing it at, at lightning speed. This shaft broke like a straw. I would guess that today such processes are fully computer controlled so there's less chance of human error.

We are doing something REALLY close to that. synchronizing generator phase with shorepower phase, and also both generators against each other.

Think of it this way -

Your waveform from your generator doesn't happen exactly the same as the one coming from the pole does it. So this is why most people cannot sell power back to the grid, and standby generators cannot be connected to the breakers.

Imagine 180V peak happening on your gen, the instant it's 0 volts on the mains.

Instant fuse blown at the least, possible blown windings or fire at the worst.

They compensate by ever so slightly slowing down or speeding up the RPM of the generator.

60hz 3phase generator = 1800RPM. 50hz = 1500rpm.

 

[Jeff7]

Or you could just use the NIST radio station like lots of clocks do.

http://tf.nist.gov/stations/wwvb.htm

Yeah, though it's not available 24/7 in a lot of areas. I only get updates for my clock around 2-3am. I think it has to do with the ionosphere being less energetic at night, and thus it is more conducive to reflecting radio signals from CO back down to the ground.

 

[bobdole369]

Yeah, though it's not available 24/7 in a lot of areas. I only get updates for my clock around 2-3am. I think it has to do with the ionosphere being less energetic at night, and thus it is more conducive to reflecting radio signals from CO back down to the ground.

It works like this.

The ionosphere has 4 layers during the day. D, E, F1, F2.

The D layer absorbs most everything below 10 mhz or so. THus during the day, AM broadcast, and the lower freq HF stuff just stops skipping.

At night F1 and F2 merge to make F, and D quickly dissipates, allowing for <10mhz skip, and E typically goes pretty far away - depends on the solar output - the more its gone the higher the MUF.

 

[LTC8K6]

My Casio watch has little trouble setting itself at night from Ft. Collins, which is about 2,500 miles away.

Also, the last few cars I have owned have had radios with non volatile memory.

 

[Jeff7]

It works like this.

The ionosphere has 4 layers during the day. D, E, F1, F2.

The D layer absorbs most everything below 10 mhz or so. THus during the day, AM broadcast, and the lower freq HF stuff just stops skipping.

At night F1 and F2 merge to make F, and D quickly dissipates, allowing for <10mhz skip, and E typically goes pretty far away - depends on the solar output - the more its gone the higher the MUF.

Damn puny human lifespans...too much interesting stuff to learn, way way WAY too little time.

 

[Colt45]

a supercap and a RTC is going to be at least a buck, even in mass quantity, i'd think.

add that someone has to add the code to the thing, etc, etc.

 

[boomerang]

Somewhat related since the thread seems to be about clocks now. I just picked up an atomic bedside clock. My previous one automatically compensated for DST changes but was hard coded for the previously used dates. Twice a year it created a bit of havoc.

In looking for one I liked, I noticed that they all had weather functions. Didn't think much of it. So the one I decided on arrived and while unpacking and reading the instructions I see that there is a cigarette pack sized remote sensor which I'm pretty certain is for the outside temperature display. It is, but ....... it also functions to transmit the time to the main unit.

The remote sensor is powered by a battery. They want you to place it in a location that is out of direct sunlight and moisture. It also needs to face the main unit minimizing obstructions such as doors, walls and furniture. It should be in a location with a clear view of the sky, away from metallic objects and should be close to the main unit as during winter, below freezing temps may affect battery performance and signal transmission. (The preceding is a paraphrase of the instruction manual)

While I'm trying to figure out where this ideal location might possibly exist, it's sitting next to the main unit.

I'm trying to determine where outside I can put it to meet the criteria and have it accessible in the winter for battery changes if needed.

I may need to hire an engineering firm.

 

[Jeff7]

...
The remote sensor is powered by a battery. They want you to place it in a location that is out of direct sunlight and moisture. It also needs to face the main unit minimizing obstructions such as doors, walls and furniture. It should be in a location with a clear view of the sky, away from metallic objects and should be close to the main unit as during winter, below freezing temps may affect battery performance and signal transmission. (The preceding is a paraphrase of the instruction manual)

While I'm trying to figure out where this ideal location might possibly exist, it's sitting next to the main unit.

I'm trying to determine where outside I can put it to meet the criteria and have it accessible in the winter for battery changes if needed.

I may need to hire an engineering firm.

Wait, are you trying to design this sensor module, or simply place it somewhere?

Build it something with an aluminum shield over it so it's out of direct sunlight, then put a solar panel + high-quality charging circuit on top, and hook that to the (NiMH or Li-ion) batteries. No direct sunlight, and it shouldn't need many battery changes.

 

[darkxshade]

I think a UPS for an entire house would be a rather expensive proposition.

In the eyes of the corporate world, it's an expense they would rather have you incur over them.

 

[Red Squirrel]

I...I'd think that the timing has to be more accurate than "flick a switch at the right time." The line is at 60Hz, and to get it in sync, you'd need to have even faster reaction time than that.
I could be wrong, but that seems like something's missing from that story - or else I'm missing something.

Yeah I'm not sure exactly how it worked, but he had about a second to work with, not 1/60th. Think it still had it's own syncing to some extent but it still had to be done at the right time. The way he described it is there was two light bulbs that flashed about every second and he had to flip the switch when they were in sync. Not sure how far this dates though, obviously I would think it's more sophisticated now. Could also be that the turbines were actually running at a very low frequency and it was just changed later on in the plant.

 

[Jeff7]

Yeah I'm not sure exactly how it worked, but he had about a second to work with, not 1/60th. Think it still had it's own syncing to some extent but it still had to be done at the right time. The way he described it is there was two light bulbs that flashed about every second and he had to flip the switch when they were in sync. Not sure how far this dates though, obviously I would think it's more sophisticated now. Could also be that the turbines were actually running at a very low frequency and it was just changed later on in the plant.

Well heck, with that, I think you could rig up some relays or something, wired into the light sockets. When both lights would receive power, something would trip to close the other circuit.

Now I'm sure most of it would be done with electronic controls.

 

[bobdole369]

two light bulbs that flashed about every second and he had to flip the switch when they were in sync

This is exactly how its still done today (at least on boats)

There is a bank of LED's labeled "Synch" - and when the LED is lit where indicated, you throw the switch. While "synching" the LED moves back and forth (or around in a circle) - like KiTT's LED on Knight Rider.

The LED indicates the phase difference, and when activate the generator varies its RPM +- a few percent. This changes the freq of the generator power, effectively changing the phase back and forth a bunch.

When the switch is thrown the RPM is locked in, the bus are tied together and very little power is lost/equipment failed.

It is very very expensive to do this automatically. (at least on boats, most stuff seems stuck in the 70's)

 

[DaWhim]

I have two UPS in the house. I bought the first one to keep the cable modem and router alive in the event of power outrage.

Since the power in my house is little unstable sometimes like the power would be interrupted for a second or so. I had to reset the alarm clock next to my bed. I end up buying a 2nd UPS just for that. well, it was cheap like 20 bucks.

 

[Red Squirrel]

Actually I wonder if this would work for a small scale solution:

Get a bank of car batteries, hook them up in parallel to a smart charger that can dedect when the batteries are at full charge (to avoid overcharging) AND an inverter (you can get 1kw ones for an affordable price - about 500 bucks) then plug the load on the inverter and the charger into the wall. Basically it would be a very long lasting line interactive UPS. You could power a couple servers off of it easily and when power goes out or blips it could last a few hours I'm sure.

An 1800w inverter would be even better as it could be protected from overload by a standard 15 amp breaker.

 

[Colt45]

that's pretty much what a UPS is.

 

[Modelworks]

Except it wouldn't only be 25 cents to add. For something like a clock radio you could probably get by with just a supercap. Microwaves, set top boxes, car radios all have custom processors that usually are not designed to use the least power possible so it would take a lot of supercaps , supercaps cost $1 - $2 each for low capacity to several thousand $for higher capacity.

They could add a philips RTC chip that cost about $1 each but then you need the layout and parts for that so back to $2-4 . Add the normal 275&#37; markup for large product runs and by the time it gets to consumers you pay $6 -$12 more for the feature.

 

[Paperdoc]

OP asked for a capacitor that would keep the clock circuit in a device going for 10 minutes. Pretty quickly the discussion became batteries, inverters, etc, a lot of it centered on larger loads. But the key answer to OP is that a capacitor cannot store enough charge to provide a discharge current of the magnitude needed for anything like 10 minutes! It really does require a battery in which the energy stored is in a chemical form, not electrostatic charge. That requires more expensive components, and any battery actually has a finite lifetime and then must be replaced. Some clocks, for example, plan for this by having you insert your own 9v battery used to maintain the clock accuracy (not necessarily even its display) through a power outage of seconds to hours. But you are responsible for periodic replacement to prevent total failure in the power outage. The fact that these designs use a standard commercial battery design easy to buy and replace is important.

 

[Modelworks]

But the key answer to OP is that a capacitor cannot store enough charge to provide a discharge current of the magnitude needed for anything like 10 minutes!

A supercap can keep the clock going on a processor for days in something like a clock. The display would be off but the cap will allow the processor to keep time till power is restored.

 

[Leros]

A supercap can keep the clock going on a processor for days in something like a clock. The display would be off but the cap will allow the processor to keep time till power is restored.

Exactly You can run a microprocessor on much less than 1mA. I've seen microprocessors which run on ten of nano Amps. All you need to do is provide up to 1mA to keep the clock ticking.