UPS Battery question

[Erik Dienberg]

What are your thoughts on paralleling strings of dissimilar brands of
batteries to serve as the DC source to a UPS system. Currently we have 2
strings of 30 PRC-1290 batteries, these batteries are series to provide
power to a 408Vdc buss. We have and would like to add 4 additional strings
of 30 UPS12-270 batteries to the dc bus. Will these two types of batteries
work at all this configuration? The characteristics of the batteries
appear to be almost identical Watts/cell and impedance, ect... The one
person I have talked with so far, says that this will not work but he is
also trying to sell me all new batteries.

Erik

 

[John Gilmer]

What are your thoughts on paralleling strings of dissimilar brands of
batteries to serve as the DC source to a UPS system. Currently we have 2
strings of 30 PRC-1290 batteries, these batteries are series to provide
power to a 408Vdc buss. We have and would like to add 4 additional strings
of 30 UPS12-270 batteries to the dc bus. Will these two types of batteries
work at all this configuration? The characteristics of the batteries
appear to be almost identical Watts/cell and impedance, ect... The one
person I have talked with so far, says that this will not work but he is
also trying to sell me all new batteries.

Wow! 408 Vdc. Be careful out there.

There are other groups (e.g.: homepower) that worry more about batteries but
because of manufacturing variations even cells purchased from the same make
and the same run may have differences.

What that means it that you ALWAYS have to pay attention to problems that
variations between cells can cause. The standard technique was to provide
an "equilization chage" whereby some cells are fully charged and a few cells
end up being over charged and lose a little water.

Likewise when you draw power from the battery to have to ensure that none of
the cells is comletely discharged as those cells will be reversed and suffer
great damage.

With over 200 individual cells you really, really need methods of automating
the checking of the conditions of the cells. Namely you should consider a
system that can read the potential of each cell (or at least every group in
a physical battery) and do some spreadsheet type analysis to see if some are
acting a little TOO different.

Of course, in most larger UPS applications the batteries are only to keep
things running until you can start the generator out back. That's the way
the telephone companies always did it.

 

[SGT]

Have you considered adding a DC output flywheel instead of battery
banks?
Since hours of emergency operation can be provided only by a generator
anyway, large battery bank is a very expensive way (not to mention
being lethal voltage at 408VDC and lead/battery acid being very
dangerous to health & environment) to provide a few hours of emergency
operation. A DC flywheel can replace a battery bank by providing
about 20 to 30 seconds of ride-through for the generator to come upto
speed & take over within 10 to 15 seconds, and low speed flywheel is a
proven solution for the past 20+ years. Two companies that make DC
flywheels (for large UPS systems) are Power Systems & Controls
(www.pscpower.com)in Richmond, VA and Pentadyne (don't know website,
but a little googling will get you the website, I am sure) somewhere
in the western USA.

Paralleling two different strings of batteries not recommended at all.
For the economy of using existing batteries you're risking a complete
failure:

- The charging is different, risk of over / undercharge
- When called for one string may charge feedback the second one, and it
might give an undervoltage signal to the UPS which in turn jumps to the
bypass for the last ditch, that means "unprotected" if the bypass is
available and "dead" if it is not.

I wouldn't do it.

SGT

 

[John Gilmer]

pretty fascinating anyway

Do you have any "back of the envelope" numbers on a typical installation.

Stuff like: A flywheen that's 3' in diameter that weighs _____ lbs in
slowing down from 3600 rpm to 2000 rpm provides ____ kW of power for ______
minutes.

I realize that we could "work this out" for ourselves but I am lazy!

 

[Ben Miller]

[email protected] (Nam Paik) wrote in message

Perhaps I've missed something, but I haven't seen a flywheel based
system in wide-spread use since the IBM power line stabilization
systems marketed during the early 1960s. Even IBM discontinued them
when battery-backup UPS became available shortly after that time. Then
too, it was the goal of the IBM product to stabilize computer
operation during only the loss of a few cycles of commercial power.

Then too, a flywheeel system providing anything approaching 20-30
seconds of emergency back-up would be incredibly expensive, bulky, and
heavy compared to the cost of a battery-backup UPS, let alone
providing 5-30 minites of backup power.

The classic problem with flywheel systems (beyond weight and size) is
that the flywheel's angular velocity begins to decellerate immediately
on the imposition of a load, plus its energy storage capacity is
severely limited by both its mass and initial angular velocity.

Just my observation.

Harry C.

The new flywheel systems run very high speed inside vacuum chambers on
magnetic bearings. The
output is converted to DC, then to utility-grade AC for delivery to the
load. This allows a large portion of the energy to be used. There are
trade-offs between power and ride-through time, but minutes, and in some
cases hours, are possible. Most of the problems that you mentioned do not
exist in today's systems.

Ben Miller

 

[Rob]

whats the q?

The new flywheel systems run very high speed inside vacuum chambers on
magnetic bearings. The
output is converted to DC, then to utility-grade AC for delivery to the
load. This allows a large portion of the energy to be used. There are
trade-offs between power and ride-through time, but minutes, and in some
cases hours, are possible. Most of the problems that you mentioned do not
exist in today's systems.

Ben Miller

 

[Cameron Dorrough]

The new flywheel systems run very high speed inside vacuum chambers on
magnetic bearings. The
output is converted to DC, then to utility-grade AC for delivery to the
load. This allows a large portion of the energy to be used. There are
trade-offs between power and ride-through time, but minutes, and in some
cases hours, are possible. Most of the problems that you mentioned do not
exist in today's systems.

I'll second that. Rotary UPS's are in wide-spread use in financial markets
(stock exchanges, major banking facilities) world-wide. The reason? They
don't fail..

The ANZ Bank HQ in Australia uses a single Piller 400kVA rotary UPS to run
the whole building.

Cameron