|
|
| Author |
Message |
| 25 new of 83 responses total. |
mdw
|
|
response 50 of 83:
| 
May 19 06:40 UTC 1999 |
1.5KVA is a common power rating, because it's the most power you can get
through an ordinary 110 V 15 A circuit without doing any rewiring.
There's a pretty good chance this will be a drop-in replacement for the
current power conditioner.
There is certainly one test we ought to do when we install it though;
and that's to pull the power plug and make sure nothing crashes when
this is done. At one point, some of the cheaper low-end UPSs had the
nasty habit of glitching the output power when switching, a bad thing.
The popularity of computers has probably forced these off the market,
but it's still a good sanity check to try this & see what happens.
|
mary
|
|
response 51 of 83:
|
May 19 10:27 UTC 1999 |
That's what I was hoping to hear, that this unit wouldn't
require an upgrade to our electrical service. At some point
we may need to do that but it would be a good idea to find out
what all would be involved before getting the hardware.
|
steve
|
|
response 52 of 83:
|
May 19 13:34 UTC 1999 |
Right, I don't want to see us change things either, unless we really
have to.
We'll do the test Marcus, and see just how long the batteries last.
This isn't an APC UPS however, and isn't likely to do strange things.
|
rtg
|
|
response 53 of 83:
|
May 19 13:55 UTC 1999 |
I would recommend an initial test with a smaller, non-critical load, such as
all the consoles, terminals, and lights in the pumpkin. Let the batteries
charge at least overnight, then pull the plug and see how long they continue to
run. Do we have the spare/test sun in bootable condition? If so, I would add
it to the test load, just in case Suns are more sensitive to transfer glitches
than the PC consoles.
I'm assuming that the unit has a status port which will give us at least
contact closures for'
inverter running' and 'low battery' states, and possibly more information.
Next project would be to connect that port to a spare serial port, and
implement powerd to give us an orderly shutdown in the event of an extended
power outage.
|
steve
|
|
response 54 of 83:
|
May 19 18:41 UTC 1999 |
Yup. I'm trying to borrow a line monitor that I know of, so we can watch
the wave form it creates. Failing getting one of those we can get a scope.
The unit has an ethernet address, so I'm assuming that we;ll have some already
written software for it, at least snmp and maybe more.
|
djf
|
|
response 55 of 83:
|
May 20 03:53 UTC 1999 |
Lead acid batteries will degrade if they are left uncharged for a long
period of time. I know that some large UPSes which I've dealt with
had big warning stickers on them that the warranty will be void if the
unit is not connected to line current and recharged within 120 days
(if I remember correctly) of the ship date.
Lead acid batteries really like being held properly at or near full
charge. They are damaged by extreme discharge, which in the case of
lead acid is something like <1.1V/cell if I remember correctly. I
don't know what the internal discharge rate of these batteries is or
how long they sat around, but they may well may be shot.
I hope very much that this isn't the case, but if you experience low
runtimes even after an appropriate period of charging for this unit,
that could be the problem. Lieberts are generally considered good
units. As such it may sound an alarm if it determines, by charge or
discharge characteristics, that the batteries aren't up to snuff.
However given the awesome price, this wouldn't be the end of the
world. I assume you could get some salvage fee for the lead in the
current batteries and put that towards new ones...
BTW, as a corollary to the above, in your testing I'd recommend not
letting the UPS discharge until it shuts off. Even though it won't
let you discharge the batteries into the truly damaging range, it is
my understanding that unnecessary deep cycle discharges can shorten
useful battery life.
It's also a good idea to have the software shut everything down when
there are still several minutes left on the UPS. This helps prevent
fully cycling batteries during every >5-7 minute power outage.
|
steve
|
|
response 56 of 83:
|
May 20 17:12 UTC 1999 |
I am expecting that these batteries are in decent shape but not as
good as brand new ones. We got this unit from a group of people who
are or have been associated with the company and get rid of stock
thats been sitting around for too long. This unit was sold as is,
but was tested by them before they brought it to the swap. Given
that I've bought stuff like this before with good luck I had no
problems in buying this unit. They were very upfront about things,
and while this wouldn't be a good thing to have in a doctors office,
it will suit us just fine. If we have to get new batteries I expect
that we'll pay about $80 for them. I believe we have four 11AH
batteries in the unit. At swaps I can usually get them for $1/AH or
thereabouts and think that new commercial cost should be around
$2/AH.
Since these are not deep-cycle batteries I agree that the UPS
should be shut down before the end of the power from the batteries
appens. I now have the web page for software on this. Will
report back when I know specifics about the software.
|
drew
|
|
response 57 of 83:
|
May 21 01:31 UTC 1999 |
STeve, is this UPS square wave, modified sine wave, pure sine wave, or
something else? And what is the efficiency? Just curious, as whatever it is,
Grex has gotten the deal of the century on it, compared to the Whistler 1.5KW
inverter that I got for $330, which includes NO provision for automatic
switching for computer equipment. (Grex can be thankful that I wasn't at that
show; I would have been tempted to grab the thing for myself!)
As for batteries, the Auto Zone sells deep cycle 95 amp-hour marine batteries
for about $60 including the core charge; and it shouldn't be too difficult
to do the necessary surgery to attach one or two of them. Just use at least
#4 guage wire, or better yet 1/0 guage, the connection. (1.5 kW is a *LOT*
of current at 12 volts.)
|
i
|
|
response 58 of 83:
|
May 21 02:24 UTC 1999 |
I doubt that a UPS that big is 12V on the battery side, though it might
use a few 12V batteries in series. Thou shalt only replace with a matched
set...
|
drew
|
|
response 59 of 83:
|
May 21 21:09 UTC 1999 |
So much the better, so long as it's an integral multiple of 12 volts. "Matched
set" can be identical deep cycle marine batteries, it seems to me.
|
rtg
|
|
response 60 of 83:
|
May 23 00:49 UTC 1999 |
drew, you shoulda' been there. There were two other very similar units sitting
right beside it. Another 1.5kva and a 1.2kva, I believe.
WHile what you say about replacing the battery string with an external string
of equal
voltage rating, but higher capacity, is generally possible, I would not take it
to extremes. I'm not sure if the charging characteristics of lead-acid are the
same as the gelled-electrolyte. Might we get into a mode where the trickle
charge doesn't quite keep the big batteries topped off, or the high inrush
current when we begin the charge cycle immediately after a long outage
overloads the charger's capacity. Also, the liquid-electrolyte batteries tend
to give off hydrogen gas during charging, and we'd have to make arrangements to
vent this safely.
|
steve
|
|
response 61 of 83:
|
May 25 18:22 UTC 1999 |
Grex is now running on the UPS. Currently, only Grex itself and the
external disks are connected. We'll need to move over the console and
probably router too.
With Grex & disks as a load, the UPS supplied power to both for seven
minutes before I got nervous and plugged the UPS back into the wall. It
may be the case that the batteries are deficient--we won't really know
this 'till we've power cycled them a few times. Right now, based on the
test I did yesterday I will estimate that we have about three minutes of
life if we run the UPS with 1.5KVa of stuff on it, which is less than
a completely good set of batteries would provide. But again, I need to
stress that we don't have enough data yet. Even with three minutes of
life on the batteries, that is more than enough to power Grex down
safely in the event of a power outage.
|
i
|
|
response 62 of 83:
|
May 25 23:39 UTC 1999 |
And more than enough time to get through many power outages. Thanks, steve!
|
steve
|
|
response 63 of 83:
|
May 26 01:20 UTC 1999 |
Heh. We need to get the console on the UPS before the thunderseason
starts for real--it would be embarressing, wouldn't it, to have a healthy
Grex and disks which was sitting in the ROM monitor because the console
lost it. ;-)
|
remmers
|
|
response 64 of 83:
|
May 27 14:43 UTC 1999 |
Yes - many thanks for the efforts. A UPS is something we've been wanting
for a long time.
|
arthurp
|
|
response 65 of 83:
|
May 28 23:09 UTC 1999 |
Thank you greatly. I feel much better.
|
scott
|
|
response 66 of 83:
|
Jun 6 17:43 UTC 1999 |
Update:
All of Grex now lives on the UPS, vith only 3 of the 5 "load" LEDs lit. So
plenty of headroom, it looks like.
On a related note, all of the power is now going through a watt-hour meter,
so we will be able to get a very good total on electricity used each month
(and no shutting Grex down to measure power, either).
|
aruba
|
|
response 67 of 83:
|
Jun 6 21:35 UTC 1999 |
THat's great, Scott. I suggest posting a piece of paper on the wall next to
the meter, and then whenever anyone's in there they can write down the date,
time, and reading. Using that data I can do a very simple linear regression
to get an average power reading and thence a payment.
Or, staff people could just mail me when they take a reading, or store the
info in a publicly readable file on Grex. Now that I think of it, that's
probably the best solution.
|
mdw
|
|
response 68 of 83:
|
Jun 7 03:41 UTC 1999 |
If it's a watt-hour meter, you could easily get the actual power used
per-month just by reading it once a month, and ignore averages
altogether.
|
scott
|
|
response 69 of 83:
|
Jun 7 10:54 UTC 1999 |
Yes. But first we need to know if the readings are anywhere near accurate
before committing to them. So more readings early, to compare to the ammeter
readings.
|
aruba
|
|
response 70 of 83:
|
Jun 7 15:37 UTC 1999 |
Re #68: If someone is willing to commit to going to the dungeon at the same
time each month, that will work fine.
|
steve
|
|
response 71 of 83:
|
Jun 7 18:01 UTC 1999 |
Given the drop in current consumption from 8.9A to 7.5, we've saved roughly
161 watts of electricty. Given that a watt is juat about $1 per year, this
means we'll save about $161 a year. Since the UPS cost us $175, it will pay
for itself in just over one year. Its definitely cooler in the Pumpkin now,
which is another benefit.
|
janc
|
|
response 72 of 83:
|
Jun 7 19:36 UTC 1999 |
Actually, a slight irregularity in when we take the readings wouldn't
matter much, so long as we pay whatever it was since the last reading.
We might want to clear this procedure with Flying Dutchman before
changing the way we bill, just to avoid painful confusions.
|
aruba
|
|
response 73 of 83:
|
Jun 7 19:50 UTC 1999 |
Yeah, there's the rub. I'd rather the payments not be erratic, becuse the
last thing we want to do is annoy them.
|
rcurl
|
|
response 74 of 83:
|
Jun 8 04:40 UTC 1999 |
Does this include the room lights?
|