PSN-L Email List Message

Subject: UPS power systems
From: sean@...........
Date: Tue, 8 May 2001 22:28:59 -0500 (CDT)


In line with the current discussions  about backup power,
here is a repeat of some info on UPS systems and inverters:

I have consistently used inverters to power seismic systems, and
several are currently in use at SLU to operate both the analog
and the digital systems, including several SUN workstations. 
These are of the totally uninterruptible design, without any switching
to the inverter in the event of an outage: the loads operate continuously 
from the 1 kilowatt precision inverters, which operate on either
12 or 24 Volt, 200 or 100 ampere DC supplies, which are connected in
parallel with 660 ampere hour battery systems, which give a full load
run time of about 8 hours. The "12" volts is actually adjusted to the 
critical cell voltages of the Lead-Calcium cells, which are warranted 
for 20 years. These Exide batteries are used in critical communications 
systems and as backup hoist power for mines.

I used similar systems at Adak, AK, in the Aleutians. The AC power from
the navy was from diesel generators, about the reliability of Angel's.
Our systems could operate for at least 6 hours from the batteries, but we 
also had our own 15 kW diesel backup generator with enough fuel for weeks.
The frequency control of the AC power was so poor that ferroresonant (CVT)
regulators (constant voltage transformers, or CVTs) were useless, so we 
used 1.5kw servo controlled variacs to control the line voltage. For 
frequency dependent mag tape and continuous microfilm recorders, the Topaz 
inverters had crystal controlled stability.  For efficiency, these large 
inverters synthesized a sine wave output.

For the IRIS stations, we use switched-mode, microprocessor controlled
inverters by BEST. They provide only line conditioning in normal mode, but
sense a brownout within 1 cycle of the AC line and switch in the inverter,
starting it with perfect synchronism; the return of the load to the AC
line is similarly synchronized with the zero crossing of the sine wave.
(Their main weakness is that they sense a "low battery" condition and
shut down the inverter to protect the battery, but then continue to run
the microprocessor from the battery, which eventually kills it AND the
program/memory backup lithium cells if the outage lasts longer than a 
day. I had to install Hg relays to disconnect the batteries if neither
the line or the inverter voltages were present). Their charger program
can be adjusted for the capacity of additional external batteries.

Similar technology is used in their better (transformer based) PC level 
UPS boxes of a few hundred watts, such as I have at home, but without 
access to the microprocessor program. They are not, however designed for 
long term continuous use, mainly because of inadequate cooling.

But a less expensive "sort-of" square-wave output inverter can be a problem. 
The better ones still use a large transformer, and will run continuously
at about 50% of their rating, But some really lightweights use a totally 
switched mode system to get 110VAC with lots of noise. I have a large 
transformer coupled one in my field van, and the square wave even messes 
with the electronic speed control of some VSR drills. The frequency is 
stabilized, though, so for some electronic applications I use a CVT which 
cleans up the square wave into a sine wave. 

As always, the good, the bad, and the ugly of it is that you get what 
you pay for, so a good initial investment is worth considering.
And as with any power switching device, a good design does include proper
transient suppression and shielding/grounding. Surprisingly, a significant
radiator of noise are the DC input cables, which often are not by-passed
with capacitors and isolated with torroid inductors.

An important note about battery life: unless you invest in the expensive
"stationery service" cells like Exide's, the choice becomes the 12 V marine
or deep discharge batteries, or the Delco photovoltaic of about the same
price. But in float service, even with Best's equalizing program, these
only last about 3 years. So for the IRIS station CCM, I need to come up
with $800 every three years to replace the 8 batteries in the systems.

This limited life is also true of the PC system UPS boxes like Tripp-Lite
and APC. They don't tell you this on the box, but they do have instructions
for replacing the battery, which can be hard to find.

Regards,
Sean-Thomas
__________________________________________________________

Public Seismic Network Mailing List (PSN-L)


[ Top ] [ Back ] [ Home Page ]

Larry Cochrane <cochrane@..............>