I'm a bit non-plussed Jim. How can a power outage smoke a UPS? My APC Smart ones (3 x 1500, 1 x 1000), are set to auto close down attached PC's upon 10 minutes of power failure, and close themselves down when less than 5 minutes of battery power is left. How loaded are your UPS's? Sometimes AC drops below a point where the UPS kicks in, power loss, brownouts etc, and when it transfers to battery it can't handle the load. And the batteries only last 3-5 years anyway.

We get very few power cuts here, but an amount of "dirty" mains, and circuit switching surges. That's why I mainly have mine.

Yeah, I'm puzzled myself. After power outage, UPS (APC 2000XL) refused to power up. A quick test revealed that the 100a fuse on the batteries is blown. Replaced with a spare, but as soon as batteries were reconnected, the replacement 100a fuse blew as well, on an unloaded power-up.
UPS loading was definitely borderline, in that I recently added a new cruncher (HP Z600) which pushed the UPS load over 80% as indicated on the control panel. This was enough to keep the internal fan on the UPS running, which normally would not happen.
FWIW, I didn't have PowerChute software running, or any other mechanism of shutting PCs off when UPS exhausts. When I did the remodel here that added the "man-cave", the new room I added got a second set of outlets connected to that UPS, for PCs, monitors, TV, Cable modem, router, etc.
Batteries are pretty new, perhaps 2-3 years old, and definitely oversized at 33ah each x 4 for a 48v string externally mounted.
My suspicion is that the load was enough, or the UPS was old enough, that when we went on battery due to the 25.5" of snow here taking out the power lines, it caused the inverter to fail in such a manner that it's showing low resistance or a short to the batteries. Haven't pulled it apart for a look-see yet, but I suspect just age.
Found a APC 2200va on eBay for a decent price, should see that in a week or so, though the shipping was a killer. In the meantime, took the 4 batteries (tested good)and paralleled them 2 per to a 400va and 600va UPS I had laying around, giving me 66ah for each UPS. That gets my main workstation and the TV/Cable covered. The HP work stations (2xZ400, 1xZ600) will have to just deal without UPS for now ...
Should get me through the next iteration of this insane Colorado weather, where 70 degree days and feet of snow seem to alternate at this time of year...
Later, Jim ...

---

Yeah, I'm puzzled myself. After power outage, UPS (APC 2000XL) refused to power up. A quick test revealed that the 100a fuse on the batteries is blown. Replaced with a spare, but as soon as batteries were reconnected, the replacement 100a fuse blew as well.

100amp fuses blowing indicates a dead short somewhere, or a major fryup in the innards!

UPS loading was definitely excessive, in that I recently added a new cruncher (HP Z600) which pushed the UPS load over 80% as indicated on the control panel. This was enough to keep the internal fan on the UPS running, which normally would not happen.

Maybe that seems like your answer. All APC UPS's have 5 green load lights on the front panel. I never run any of mine above three of them. The point for people to remember is that the model number does not equate with power load capability in Watts. It is stated in VA

2000XL = 1400W
1500 = 980W
1000 = 700W

FWIW, I didn't have PowerChute software running, or any other mechanism of shutting PCs off when UPS exhausts.

It is useful to have that, and there is a special workround for Win 7 PC's to get ot working.

Batteries are pretty new, perhaps 2-3 years old, and definitely oversized at 33ah each x 4 for a 48v string externally mounted.

The APC units have an auto self test and a battery meter on the front panel. You can see when they are getting end of life, and Powerchute also gives you a red alert on the taskbar.

No expert, just a savvy end user :-))

With these older UPSs, the connection is serial for power chute, which doesn't work too well for me given the physical layout. Not new enough hardware to have the ethernet connectivity. I did have a workaround there, though the old BBS system, but I lost that when I lost my domain name and pulled the plug.
Loadwise, I was 4/5, and agree that 3/5 is safest.
But I agree, most likely a major internal fry-up, and probably not one I could have avoided. The 2000XL is old hardware. Don't see a date code, but based on what I know of its history, it's probably at least a dozen years old ...
Too much fun . . .

---

The best way to buy a UPS is to go on Ebay and get a working one WITHOUT batteries.

Indeed. Just what I did. $180 usd for a APC2200XL, with no batteries. Shipping still probably $80, just on weight.

The price of replacement batteries is an industry rip-off in my opinion. 50% dearer than a heavy duty car battery. There are people that sell "kits" of assemble it yourself batteries, but usually they are of an inferior make and quality.

We have a chain here called Batteries Plus, and their prices for UPS style batteries isn't put of sight. Given the avoidance of (direct) shipping costs, usually they are cheapest. At one point I had gotten marine slow discharge (deep cycle) batteries from Walmart. Not a good plan; didn't take the constant float environment of a UPS well; only lasted a couple years before the UPS declared them bad.

Bit of a minefield all round really but I estimate that probably 80% of all UPS sales are to companies and corporate enterprises that write off the price against allowable operating costs in their tax returns. i.e. they don't care what they pay, it's all essentially free anyway. It's us home consumers that get stung for trying to be sensible.

Yep. I just look at $/ah, and go from there. Since I'm hanging the batteries externally, physical size is no big deal. I do miss working in the phone biz, as from time to time I could come across old free battery systems from PBXs. Those GNB Absolyte II batteries were amazing ... 10 year lifespan on average ..

---

Meanwhile London is getting battering with more to come tomorrow, it is hissing down :-(

Chris its Easter bank holiday monday tomorrow what else would you expect ;-)

Well, at least the guys on board the Ross Revenge have fun, doing their Easter Radio Caroline broadcast in the middle of this bad weather. You can listen to them via their website (click the Easter icon). :-)

---

Pretty typical Bank Holiday weather! Barometer down to 958 Mb and falling ....

Talking about falling and falling over, Seti is in trouble, most of the servers are off air for maintenance or something, but the website and boards are up. Obviously unplanned, but at 10.45pm on a Sunday night they might not try anything like remote reboots until the morning their time.

Always pays to have 3-5 days cache of work :-)

Guess I better get over my aggravation with Raistmer's attitude and go grab some Beta work ...
[edit] bummer, no work there either ...

---

Too much fun! Batten down the hatches!!

---

And we are down.

I hope staff has fix for AP database server...

At 2 PM PDT we are still down.

After a long long load time, I see the forums are back.
Edit: but the status page disagrees.
Edit two: ah, now it agrees.

---

Seems SETI is down

Just checked with "isitdown" and it agrees.

Yup I get the same.

I wonder how long? Is anyone starting a sweepstake on when?

About 10 minutes after the lab opens. Everything's unreachable - web server, scheduler, upload server. Those are usually the easiest ones to fix - maybe not literally plugging one wire back in, but to a tecchie something equivalent. And with SETI, we usually see that everything has been running just fine behind the broken wire, as soon as we can see it.

Unless anyone's heard anything about the San Andreas fault?

That must be a campus or Colo fault not Seti?

Is this server we are conversing on not also in the same Colo?

That must be a campus or Colo fault not Seti?

Is this server we are conversing on not also in the same Colo?

I would think so. The IP address of boinc.berkeley.edu is 208.68.240.115 and the IP address of setiathome.berkeley.edu is 208.68.240.110 so they are on the same network.

The funny thing is that both IP addresses respond to a ping so it appears it is not a network problem.

Edit: Interesting that

setiathome.berkeley.edu = 208.68.240.110
208.68.240.110 = muarae1.ssl.berkeley.edu

boinc.berkeley.edu = 208.68.240.115
208.68.240.115 = isaac.ssl.berkeley.edu

I am not familiar with either of those host names.

Uploads are stuck!!! :-(

---

Have TARDIS, will travel...
Come along K-9!
Join Calm Chaos
Pluto is still a planet

Isaac is the BOINC server, the one we're using for this message board.

Or, I suspect Jord will pop in to say, Isaac is the name that has been given to a succession of BOINC servers down the years.

Ah, how tragic in that I was starting to snag a few of those tasty APs to crunch.

Isaac is a new one on me.

You're posting on it: it's the BOINC domain server, and the second one already after the first one mysteriously died from one moment to the next.

Edit: the "new" isaac server is what was formerly muarae3; muarae2, 3, and 4 are/were general compute servers.

---

Obviously fixing this unscheduled outrage is requiring more than a quick flip of a switch.