Everything should have a capacitor battery backup

[Leros]

The power went out today for ten minutes. I spent more than 10 minutes just resetting devices. Microwave, coffee maker, clocks, turning back on devices. It was a pain.

Devices should have a capacitor which keeps the clocks alive when power goes out. It wouldn't need to power the entire device, just the clock portion.

Also, why do car radios lose your favorite stations when you disconnect the battery. That is ridiculous. It would cost them at an extra 25 cents to put the hardware in to remember your favorites.

 

[PottedMeat]

Yeah it would be nice but they'll gladly cut out <$1 components on non-critical functions. Plus those real time clocks are older ( and cheaper ) and consume considerably more idle current than newer RTCs.

 

[DarkWarrior2]

I wholeheartedly agree with you. My power went out a few times in the past year, and I don't even reset the microwave clock anymore - just the answering machine.

Luckily my alarm clock has a 9 volt battery to keep the time when the power fails.

Unluckily, the battery was dead during the most recent power outage, so I had to reset the alarm clock too. :\

Luckily, I should be moving out of this place soon! :awe:

 

[Jeff7]

And they wouldn't need to be powerful capacitors. Most of the outages I experience last less than a second, but that's enough to reset the clocks.


But yeah, it sucks having to take nifty features out of products in the interest of cost cutting, or to stop adding fun new features in the interest of meeting deadlines.
It's the battle between engineering and management.

Engineering: I can make this product awesome.
Management: I want you to make this product marketable.

Sometimes these things don't match up well.



Incidentally, I was working on a product at work that previously lacked backup for its internal clock, but with the addition of an improved RTC, I also tacked in a lithium backup battery.

 

[TechBoyJK]

I'm not an electrical engineer. I don't know much, and I probably still think I know more than I really do.

That said, I think it's a good idea, but I suppose they would have to engineer a bypass to that battery (assuming its inline) so that if that battery ever failed, the device would still at least operate like it would have if it didnt have it in the first place.

 

[Chronoshock]

Wouldn't it be easier to solve it in a centralized way by adding a battery backup to your power lines -> home connection as opposed to adding it to every device?
That way you wouldn't compromise design considerations in smaller devices and presumably get some space and cost benefit from consolidating it into a single component

 

[Number1]

Wouldn't it be easier to solve it in a centralized way by adding a battery backup to your power lines -> home connection as opposed to adding it to every device?
That way you wouldn't compromise design considerations in smaller devices and presumably get some space and cost benefit from consolidating it into a single component

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

 

[0roo0roo]

well no because power lines wouldn't be able to just power the minimum necessary, but the full actual load of everything= giant expensive bank of batteries that require maintenance.

 

[Rubycon]

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

Expensive but essential operations use UBS - Uninterrupted Battery System. When shore mains fail batteries take over. If transfer (bus tie) switchover fails then generator starts. Generator output is DC and charges batteries. Generator sets are powered by No. 2 diesel in large bunkers (1000 tons not uncommon) or may be natural gas powered although in earthquake prone areas natural gas supplies frequently get interrupted.

As long as the juice stays on. (911 call centers, hospitals, casinos, etc.)

Wet caps can be used - several Farads of capacity at 5V are only a few cm² in volume. Configurations will be remembered forever if EEPROM is used. Problem is cost. Cheap things made in CHINA where keeping COST at a MINIMUM is not going to have features of "luxury".

 

[Leros]

Wouldn't it be easier to solve it in a centralized way by adding a battery backup to your power lines -> home connection as opposed to adding it to every device?
That way you wouldn't compromise design considerations in smaller devices and presumably get some space and cost benefit from consolidating it into a single component

No because then you're powering the entire device which draws lots of power.

You only need a few mA maximum to keep the clock alive in a device. You don't need to power the screens, you don't need it to respond to buttons, you just need it to keep tracking time.

 

[Q]

Also, why do car radios lose your favorite stations when you disconnect the battery. That is ridiculous. It would cost them at an extra 25 cents to put the hardware in to remember your favorites.

All that $0.25 adds up though to a lot

 

[Leros]

All that $0.25 adds up though to a lot

They could pass the costs on to the consumer. Heck, recent radios should have some sort of EEPROM already in the device, then its just adding a few more lines of a code to the software.

 

[Red Squirrel]

I've thought of building a UPS for my whole house but it would be insane expensive. Batteries aren't THAT expensive but the rest is. Start off with only a few to last you an hour or so, then keep adding on as you get more funds. The expensive part is the inverter, and the equipment required to make it charge in real time when the power is available. Now what would be really cool is this combined with solar/wind power. So you are constantly charging the batteries while running off them, and the hydro charges whatever the solar/wind can't. I can imagine the price of such system though.

We have a huge UPS at work (hospital) and think it was a couple hundred grand.

I think my microwave has a built in battery or something to keep the time though. The power must of flickered today as I noticed my alarm and my stove was flashing when I got home for lunch but the microwave was not. At first I only noticed the stove so I was scared I might of had an issue with one leg on my wiring or something. I remember when I lived at my parents someone backed into the transformer and half the house had power, the other half did not, was weird. Must of broke one of the hots but not the other. Good thing it did not kill just the neutral. Floating neutral = bad.

 

[Rubycon]

I've thought of building a UPS for my whole house but it would be insane expensive. Batteries aren't THAT expensive but the rest is. Start off with only a few to last you an hour or so, then keep adding on as you get more funds. The expensive part is the inverter, and the equipment required to make it charge in real time when the power is available. Now what would be really cool is this combined with solar/wind power. So you are constantly charging the batteries while running off them, and the hydro charges whatever the solar/wind can't. I can imagine the price of such system though.

We have a huge UPS at work (hospital) and think it was a couple hundred grand.

I think my microwave has a built in battery or something to keep the time though. The power must of flickered today as I noticed my alarm and my stove was flashing when I got home for lunch but the microwave was not. At first I only noticed the stove so I was scared I might of had an issue with one leg on my wiring or something. I remember when I lived at my parents someone backed into the transformer and half the house had power, the other half did not, was weird. Must of broke one of the hots but not the other. Good thing it did not kill just the neutral. Floating neutral = bad.

A 25kVA ferroresonant transformer should handle a residential load. Problem is these are three phase devices and three phase power is not usually served to residents. Just the line conditioning the transformer would provide would be very nice. You would be completely isolated from the utility mains for starters.

 

[Eli]

I've thought of building a UPS for my whole house but it would be insane expensive. Batteries aren't THAT expensive but the rest is. Start off with only a few to last you an hour or so, then keep adding on as you get more funds. The expensive part is the inverter, and the equipment required to make it charge in real time when the power is available. Now what would be really cool is this combined with solar/wind power. So you are constantly charging the batteries while running off them, and the hydro charges whatever the solar/wind can't. I can imagine the price of such system though.

We have a huge UPS at work (hospital) and think it was a couple hundred grand.

I think my microwave has a built in battery or something to keep the time though. The power must of flickered today as I noticed my alarm and my stove was flashing when I got home for lunch but the microwave was not. At first I only noticed the stove so I was scared I might of had an issue with one leg on my wiring or something. I remember when I lived at my parents someone backed into the transformer and half the house had power, the other half did not, was weird. Must of broke one of the hots but not the other. Good thing it did not kill just the neutral. Floating neutral = bad.

Hmm.. I think batteries are by far the biggest expense in a UPS that size. The charge controller might be a pretty penny. The inverter would be a couple grand, but good 200Ah AGM batteries are going to run 300$+ each.

It's also best to buy all your batteries at once. They have a finite life. It sounds good on paper to buy a few a year, but what happens when you're still adding batteries in 5 years? Now you have brand new batteries with 5, 4, 3, 2, and 1 year old ones. Eventually it will become hard to balance your pack, resulting in more maintenance and shorter battery life.

I suppose you could use regular flooded cells and save quite a bit of money, but even more maintenance. Really, it just depends on how big of a system you want/need.

Man... I really need to get my 2000W inverter rockin. Need moar batteries! 210Ah just doesn't cut it.

 

[sjwaste]

A 25kVA ferroresonant transformer should handle a residential load. Problem is these are three phase devices and three phase power is not usually served to residents. Just the line conditioning the transformer would provide would be very nice. You would be completely isolated from the utility mains for starters.

Don't give this guy any ideas. He might get one and mount it in his wooden server rack.

 

[Eli]

Don't give this guy any ideas. He might get one and mount it in his wooden server rack.

I LOL'd.

 

[Rubycon]

Man... I really need to get my 2000W inverter rockin. Need moar batteries! 210Ah just doesn't cut it.

Look into Deka forklift batts.

Don't give this guy any ideas. He might get one and mount it in his wooden server rack.

Wood + fault = farty sparks.

http://www.youtube.com/watch?v=2tCuzys-1P8

 

[sdifox]

except when the cap leaks...

 

[Jeff7]

No because then you're powering the entire device which draws lots of power.

You only need a few mA maximum to keep the clock alive in a device. You don't need to power the screens, you don't need it to respond to buttons, you just need it to keep tracking time.

Or less. That RTC chip I mentioned has a draw of 100nA when on battery power.


I think proper battery backup depends on the RTC having a separate pin dedicated to the battery. That also requires that the RTC is a separate component, which may not be the case - the clock may simply be code or hardware running in the device's main processor.

 

[PottedMeat]

Or less. That RTC chip I mentioned has a draw of 100nA when on battery power.


I think proper battery backup depends on the RTC having a separate pin dedicated to the battery. That also requires that the RTC is a separate component, which may not be the case - the clock may simply be code or hardware running in the device's main processor.

The simplest way to keep the RTC alive is just to have a battery/cap at the RTC power pin, then a series diode to feed power. When Vdd = 0, the cap discharges into the RTC, diode prevents the cap from backfeeding Vdd. Some RTCs have extra charging regulation circuitry for more types of batteries requiring a separate pin.

 

[drinkmorejava]

Just a note. Most cheap clocks have no crystal and track time via the line frequency. Although what you receive may not always be exactly 60 Hz, the power company periodically adjusts this to average out any variance.

A cap could be nice for one second outages, but to maintain the time for more than 2-3 minutes, it's more than just throwing in a $0.04 cap.

 

[Rubycon]

Just a note. Most cheap clocks have no crystal and track time via the line frequency. Although what you receive may not always be exactly 60 Hz, the power company periodically adjusts this to average out any variance.

A cap could be nice for one second outages, but to maintain the time for more than 2-3 minutes, it's more than just throwing in a $0.04 cap.

Yes there is NO cumulative error with synchronus timekeeping on the AC line.

Of course to the unsuspecting guest when there is a problem with line frequency in the middle of the night and their alarm clock is an hour slow they get upset when they miss breakfast, etc. That's what the wakeup call feature on the phone is for! (fortunately this is quite rare - on older ships and freighters it's much more common.)

 

[Red Squirrel]

I've always wanted a clock that plugs into Ethernet and can be configured to get time from a local NTP server. That would rock.

 

[Jeff7]

The simplest way to keep the RTC alive is just to have a battery/cap at the RTC power pin, then a series diode to feed power. When Vdd = 0, the cap discharges into the RTC, diode prevents the cap from backfeeding Vdd. Some RTCs have extra charging regulation circuitry for more types of batteries requiring a separate pin.

Yeah, that'll work too.

Just a note. Most cheap clocks have no crystal and track time via the line frequency. Although what you receive may not always be exactly 60 Hz, the power company periodically adjusts this to average out any variance.

A cap could be nice for one second outages, but to maintain the time for more than 2-3 minutes, it's more than just throwing in a $0.04 cap.

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.

I've always wanted a clock that plugs into Ethernet and can be configured to get time from a local NTP server. That would rock.

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).