1 second UPS

Hi, I?m new to this forum, impressed to see there are still a lot of people out there who love to play with electronics and do not mind sharing some good ideas. So here?s my situation;

Dealing with a cabin with a lot of power outages. I have a trace inverter system providing DC from 12 volts batteries. I also have a desktop computer. When the grid power falls (or comes back) the inverter switches. Most often not fast enough and the computer reboots :( I tried a simple UPS (after the inverter) but it doesn?t recognized the modified sine wave from the Trace as ?clean? power, so it just runs its battery dead after about

20 minutes without consider the Trace is there take over.

My search is to find an ?UPS? that that will bridge the half a second I need for the Trace inverter to kick in. Time being so short, I figure I don?t need clean power, both voltage and frequency don?t need to be stable, just something half a second to fool my computer?s power supply AC is still on.

I was thinking along following lines:

- feed power to computer using a NO contacts on a DPDT relay 110v AC that goes on when AC (Trace) is present.

- charge (before this relay)through a rectifier a 680uf 200v cap up.

- create a 20ma 12v power using a rc bridge and a zenner over the cap to run a

555 at about 60hz

- feed this frequency straight to an n-mosfet 200v 15amps switching the capacitors charge to the NC contacts of the relay .

- put a resistor in series with the relay to run it at about 80V (so it still goes on at regular 110 supply, but falls faster when the power starts going down.)

- so when the power falls, the relay switches the line AC off and feeds a straight square wave at 60 hz for as long as the cap will discharge.

Again this is ugly power, but I only need it for half a second or less. When the Trace inverter kicks in the relay will switch back on and regular power restored.

Figuring charge in a cap is CV2/2 in watts/sec, a 680 uf cap should hold 4 watts/sec. A computer power supply being approx 250 watts/hr, this is 0.07 watts/sec. No I?m not expecting my cap to give me a full minute, that would likely make something explode :)

Comments? Better ideas? I like the challenge of doing this without a transformer!

Sorry for the long post.

StefanV

If you still have the "simple" UPS system that works, why not connect your big battery in place of the battery in the UPS - assuming the voltages are compatible.

--
Dan Hollands
1120 S Creek Dr
Webster NY 14580
585-872-2606
dhollan3@rochester.rr.com
www.QuickScoreRace.com

another problem with relays...

I tried something like this for a sump pump...

the pump connected to the DPDT relay wipers,

the power line to one set of contacts

an inverter to the other set of contacts...

I first tested this with a lamp load and when the relay switched, my inverter blew up...

I suspect the relay contacts.... when opening at the peak of a cycle ....,can arc for a short while and the arc actually bridges the inverter to the power line which promply destroys it...

I decided I would need complex time delays and two relays to absolulty prevent this... In my case I didn't care if the pump was off for a few seconds during the switch over, but I didn't like my inverter blowing up when the realys switched...

So in your case, a realy may not do the trick....

I agree with the other poster, get a big set of batteries for the UPS and just use it directly... it's already desgined to do the switchover... thats what I ended up doing for the sump pump...

Mark

Once upon a time, some power supplies were made with an internal battery (sealed lead acid) so that power supply could continue to run for minutes after a power loss. Did not sell well because consumers need to see the big and separate UPS. Don't know if these power supplies are still available. But it is what you seek.

Computer power supplies will normally maintain power output for a few tens of milliseconds of power loss. You need it to maintain power for hundreds of milliseconds. That typically means some type of battery as part of the computer's power supply. It was not a big battery - maybe about the size of six size D batteries. Once designed a 'PC' wall controller with that battery inside so that power loss left the PC working. They were becoming harder to find even back then.

Otherwise the UPS needs power from some big ass l> Hi, I?m new to this forum, impressed to see there are still a lot

[snip]

A lot of UPS's (APS for example) have an output that signals "Your computer must shut down in x minutes".

Why not use said signal to start an inverter?

Here, in Arizona, I've contemplated such a signal system to start a motor-generator set (I have a very expensive reef tank that wouldn't do well with any sustained power outage :)

...Jim Thompson

--
|  James E.Thompson, P.E.                           |    mens     |
|  Analog Innovations, Inc.                         |     et      |
|  Analog/Mixed-Signal ASIC\'s and Discrete Systems  |    manus    |
|  Phoenix, Arizona            Voice:(480)460-2350  |             |
|  E-mail Address at Website     Fax:(480)460-2142  |  Brass Rat  |
|       http://www.analog-innovations.com           |    1962     |
             
I love to cook with wine.      Sometimes I even put it in the food.

UPS are commonly availble for FREE with bad batteries, which are frequently 12 volts.

So just put your computer on a surplus UPS, go oversize, and use seperate battery/s

I say go oversize because the design run time isnt just battery capacity, its also about component heating.

Oversize is cheap and effective....

make before break will connect the inverter to the power line and blow it up....

even if you use a break before make relay like I did, the arc will jump across sometimes and still blow up the inveter

Mark

The only 100% reliable system I've ever built for this sort of thing used a regulated switching supply to float charge a SLA 12V battery which powered an inverter which ran the computer. Since the first supply is regulated, the batteries are just sitting there being floated until the power fails. When this happens, The secondary inverter simply keeps going off the batteries. No relays, no switches, no glitches. I built that system back in the '70s.

Ted

"stefanv" schreef in bericht news:TZGdnYWjG5qZDQjeRVn snipped-for-privacy@giganews.com...

Stefan,

Guess you have to refresh your basics on electricity. The energy stored in a capacitor is CV/2 in J(oules) or W(att)s(econds). The energy required to power the PC for half a second is 250 * 0.5 = 125Ws, which is much, much more then your capacitor can hold.

petrus bitbyter

You need to make sure you use a make-before-break relay for that application. There are actual mains cut-over contactors which are made for the job.

A satellite station I worked on previously had a big cut-over switch for the mains to generator. In auto-mode it was guaranteed to cut across within a cycle of the mains supply. Apparently the *really* big versions of this (ours was only 20kVA) would break your arm if you got in the way!

Cheers.

Ken

I'm clearly losing it - I made double-sure I wrote the right thing coz I suspected I'd get it arse-about.

Ken

in a

Petrus,

To my knowledge a 250watts power rating are watts per hr. To make sure I checked last months electric bill :) I used less than 400 kw/hr for the month. This is

400,000 Watt/hr. At your claim for a computer supply at 250 watts /sec, 1 computer alone would run that in 1600 seconds or 27 minutes per month. I think you are mixing up W/hr with w/sec. J are w/sec indeed. I'm sure I won't get a long time, but 1/2 second is more than likely if I get the rest of the circuit to run :) A computer uses "only" 0.07w/s

StefanV

So I tried it

I haven't gotten to the relay yet as I have another problem to solve first. I build the circuit and made a schematic at

As expected I get a bad quality square wave that actually runs a transformer really hot in a minute :) but as I need only 1/2 a sec of it, I'm not to worried. But, while I have 135v DC loaded in the capacitor, I only get a 78v "AC" after the mosfet. Any sugestions why and how I can crank this up?

Then I'll try the relay,...

StefanV

I see your point and my mix-up between watts-second and watts per second. I will have to go back to my school books and refresh.

Re my schematic at

could I bring up the voltage using an H-bridge?

Re transfering AC to DC, back to AC and back to DC. Indeed, it sounds crazy, but it's a combination of 2 things. First the chalenge of making an idea work once it's in your head! Next, I know of many people with the same problem who wouldn't mind having one of these (if it ever works) The luxury of just plugging in a device in-line and having the problem fixed, is worth more to them than opening up computers, changing power supplies, opening up UPS, hooking external batteries or buying overtock items and rebuilding them to fit the purpose.

The value for them is in an easy, off-the shelf, plug and play device. Yes, if ever, I would probably build 50 of them and sell them just above cost. I love building things and I like the idea of a compact in-line device vs a bulky one.

It's a chalenge, I appreciate your advice!

StefanV

Well ----- your computer uses ~250 watts every instant it is on. If it is on for one hour, it uses 250 watt_hours. If it is on for 1 second, it uses 250 watt_seconds.

Your posted circuit would give you only about 30 mS (if that), and you need 500 mS. Look at the RC time constant: the PC draws 2.5 amps every instant it is on. At 110V, that's an equivalent resistance of 44 ohms. Your RC is .000680 * 44 or .02992 seconds, at which time the cap will have discharged to about 85 volts and the PC will shut down. (At least, I assume they shut down at around that voltage. Maybe it's higher.)

Ed

Maybe get yourself some nicads and run the LT rails off those during power down. Perhaps monitor the 5&12v rails, and when they get close to min V spec, switch on your nicads, using a solid state switch. This could sit in a spare pci slot, making it nice and neat. It would be a UPS as well as a glitch coverer.

NT

Have you had a look around for a replacement power supply inside the computer. Rip the present one out and replace it with one that will run directly off the 12V batteries.

You are basically converting dc to ac then back to dc again.

sounds crazy to me ......

Your electric bill is in kilowatt-hours, not kilowatts per second.

Your 250 watt computer would run for 1600 hours (not seconds) to equal your 400 kilowatt hour electric bill.

Watts are joules per second. No hours involved.

Joules are watt-sec>To my knowledge a 250watts power rating are watts per hr. To make sure I

What about getting a 12V computer power supply, running that off the battries and charger the batteries with a charger?

Pat