You are here: Greyghost > Product Reviews > APC Back-UPS Connect BGE70 Product Review
APC Back-UPS Connect BGE70 Product Review
Sometimes you wonder why anyone would buy a certain product.
Sometimes you wonder how it is a product works as well as it does despite severe corner cutting in its design.
Sometimes you wonder why a company designs something in the way that they did.
Here's a product that begs all three questions, and maybe a few more besides.
BGE70 Back-UPS Connect is an ultra tiny uninterruptible power supply
made for use with low power loads, such as wireless routers and VoIP
gateway devices. It may also come in handy for use with other types of
networking equipment. Two other models besides the BGE70 exist in this
particular product category, a BGE50 and a BGE90. The primary
difference is in their capacity just as the model numbers would
suggest. The BGE50 uses a lithium ion battery, while its siblings use
conventional lead acid batteries. For the two largest models, APC
claims that when running on battery, that the output waveform is a
square wave. Going by the specifications, BGE50 is again the outlier,
with a claimed modified sine wave output and, rather surprisingly,
three pin grounded outlets!
Manufacturer suggested retail price on these units starts at a
downright astonishing $140 for the BGE50 and $50 for the
midrange BGE70. They ought to turn up cheaper in other places with some
looking, and I definitely suggest that you do
some looking around for the best price, if you really want one of
these. Retail price for my example was around $30, and much less after
stacking up some coupons and other discount offers.
All of these products are marketed as a way to keep your Internet
connectivity equipment, and a few other "essential" low power draw
devices online throughout power outages. Even the most miserly of
computer systems are likely to far exceed any of these units'
capability, and that's assuming you could even plug them in. APC
includes only two pin outlets on models other than the BGE50ML, and I
can only guess that their intention was to stop people from majorly
overloading these units by expecting a computer of any kind to operate
from one when the lights go out. Unfortunately, this decision also
limits their utility. Some networking devices that would fit
comfortably within the power available from these units have three pin
grounded electrical plugs. And that's that--without resorting to
floating the safety ground by misuse of an "adapter" or by simply cutting it off, you can't plug
any such device into these units.
Floating the safety ground of any electrical device, outside of some very isolated and controlled scenarios, is a bad
idea. No, really. It is. It doesn't matter that people do so "all the
time". People do lots of things that really aren't good ideas and a
surprising number get away with it for a very long time.
It can happen, and is more
common than you might think, for a device to develop a short circuit to
ground. If ground is properly connected as it should be, a circuit
breaker will trip and a fuse will blow almost instantly. Without safety
ground being hooked up, it's possible for a malfunctioning electrical
device to develop a live cabinet. Consider for a moment that anything
standing on a more or less direct connection to the Earth is in fact
grounded. It may not be a very good ground, but what is there is a
ground nevertheless. It doesn't have to be a good ground for someone to
come along, touch a malfunctioning or miswired device with a live
cabinet and end up dying from the experience. It only takes a few
milliamps across the human heart to stand a pretty good chance of stopping it.
Another question brought to mind by the lack of a ground connection at
either the outlets or line cord on the 70 and 90 models is that
of surge protection. APC claims these units have surge
protection capability. An inspection reveals a lonely single MOV in
line with the power input. MOVs, or Metallic Oxide Varistors, are
common, cheap and passably effective surge suppression devices. They
benefit of fast surge clamping time at a very low cost. MOVs work by
turning into a short circuit under surge energy conditions and divert
as much surge energy as they can to ground. Whatever's left over is
clamped to a level still high above normal line voltage, with the hope
that an attached device can withstand voltage in excess of its normal
rating for a moment or two.
Only there's no ground here, at least not a proper Earth
ground. Here in the United States, homes and small businesses with
"split phase" power coming into the premises have a neutral that is
connected to the center tap of a pole mounted utility transformer. At
the same time, that center tap is bound to Earth ground--both at the
power pole and at the location where electrical service enters a
building's primary electrical panel. Or, again, it's supposed to be.
Despite the presence of a nationwide electrical code, there is no
restriction here in the United States (at least not on a federal
government level) as to who can buy or install electrical equipment for
a residential setting. A lot of particularly old but otherwise
sufficient electrical wiring remains in service that may not have been
done to current standards. And things do degrade over time.
This is also why devices sold as surge suppressors sometimes come with
a notice stating that they must be connected to a grounded
outlet. Otherwise they aren't very effective.
I can only suppose that APC intends the MOV inside these units to shunt
as much of an incoming surge as possible through the neutral side of
the power line, hoping that it'll ultimately get to ground. There's
really only one problem with this: electricity always takes the path of least resistance!
Chances are very good that your means of connection to the outside
world IS properly grounded somewhere along the way, and even if the energy from a surge had to
jump a few circuit traces over to reach that ground, it would be more
likely to do so than to go through a much longer pathway all the way
back to your building's grounding rod or that of the utility
(December 2015 update: Since writing
the above, I researched the matter by looking at some MOV data sheets.
This revealed that it is possible to place an MOV across the hot and
neutral conductors of the AC power line and that some protection will
be afforded by doing so.)
MOVs do have the unfortunate habit of decaying over time to the point
where they are no longer useful. As they are by far and away the most
common means of surge protection in almost any surge suppressor offered
for sale, APC gets a pass on using them here. As with many power
protection devices, you will not receive a warning when the MOVs are no
Rather curiously, there is no
provision on the BGE70 for network, coaxial or telephone line surge
protection. Of course, without a safety ground, there would no place
for surges coming in via these cables to go.
Click the above picture to see the 2.3 MB high resolution (2288x1712) original.
The primary (in actuality, only) circuit board is held in only by
friction and one molded guide channel within the case. Rather
surprisingly, APC included a disconnect for the built in outlets.
Release it, and the circuit board is more or less free to come out for
inspection. Be careful of the attached circuit breaker, which is also
held in only by the relation of the top and bottom case panels.
There's not really a whole lot to say here. What's here looks like your
average inexpensive low output capacity standby-type uninterruptible
power supply. There is no voltage buck (reduction) or boost capability,
not that you could expect such things at this kind of price point.
Every electrolytic capacitor on the board is made by one of three
companies: OST, CapXon or Chang.
All of these companies are well known for having produced poor quality
capacitors that vent electrolyte and subsequently fail. It is
questionable as to whether or not their quality has improved since
those times. At least some of them are 105°C rated parts...
The battery charging circuit has a transformer isolated output and is based around a Power Integrations TNY274P
highly integrated low cost regulator. Other portions of this circuit,
particularly the controller IC, are directly connected the AC line. Its
peak charging current is somewhere around 300mA, which drops to nearly
zero when the battery is fully charged. Perhaps a state of charge test,
a small amount of current is periodically drawn from or sent to the
The circuit board is marked as part number 640-4219A-Z and as revision
1. It bears a code name of Mt,Whistler [sic]. Exactly what this
references is unknown. (Author's note: Since writing this I have come to know that some APC products
were code named prior to release after various mountains. Other code
names that I found were McKinley (CS350/500), Cerro Torre (BR1500), and
A little bit of hot glue does wonders to quiet the onboard beeper down.
If you really don't care in the slightest about having a warranty, and
don't want to know when this unit goes on battery or reaches a
critically low level of battery charge, you could simply remove the
beeper. Be sure you disconnect all sources of power before tinkering inside the unit, and don't do anything if you're not sure or lack experience in electronics repair.
Setup and Installation
There's very little involved in placing the BGE70 into active service.
APC provides fairly decent basic instructions to help you connect the
battery, connect whatever equipment you wish to protect and ultimately
get everything connected to power. While nothing is foolproof, those
who can't get this unit set up might want to consider a different
I was curious as to what kind of runtime this unit might offer under a
nearly fully loaded condition, and if it would remain in a safe
operating situation under high load. APC claims these units will shut
down when overloaded. The rating given for the BGE70 is 75 watts. I
found a 65 watt plant light bulb in a portable fixture, plugged that
into the unit and pulled the BGE70's plug. Runtime was approximately 21
minutes before a low battery alarm sounded, and 28 minutes before a
shut down due to an exhausted battery took place. These figures will
decline over time as the battery wears out.
Having connected an oscilloscope
meter to the BGE70 while it was
running on battery, I discovered that its output waveform is actually
the usual "modified sine wave" type, at least until the battery starts
to get weak. When that happens, the inverter's output gradually reaches
a point where it does become a (somewhat sloppy) square wave.
Click either picture to see it larger.
the BGE70's battery drained, I noticed that its output voltage kept
going up (tested with a true RMS meter). What started out as a
perfectly reasonable 114 volt output soon rose to 121 odd volts AC. As
the battery became critically low, however, behavior of the inverter
became much more unstable. Output voltage rose to 130 volts AC and my
Kill-A-Watt meter's display began to flicker in a rather unnverving
way, suggesting a (very) poorly filtered output. This could cause
damage or erratic operation of an attached device -- exactly the kind
of thing that a power protection device is supposed to prevent!
All this was present while driving a simple resistive load. Just how
bad does an inverter have to be that it can't maintain stability under
such a straightforward situation? What might it do in the face of
highly inductive loads that are much more difficult to drive?
The BGE70 uses an APC RBC153
replacement battery. Searching around online did not reveal a generic
source for this 12 volt, 4.2 amp-hour battery, as it has an unusual
form factor. At best this leaves you paying a highly inflated cost for
a new battery. At worst it could mean that a time will come when you
can't actually buy a replacement battery. I choose to be optimistic
here by thinking that third party battery suppliers will produce a
compatible replacement at a fair price.
feel as though the limitations of this product outweigh its benefits.
It lasted surprisingly well under heavy load and only uses its alarm
beeper sparingly. Two pin outlets preclude its usefulness with some
very common pieces of network equipment. A battery of unusual shape and
size locks you buying a replacement from APC at their very inflated
prices, at least as of this writing in late 2015. Instability of the
inverter as the battery level dropped has the potential to cause
problems with some devices. I'm not sure the surge protection
capability will actually function as intended, unless there is some
aspect to its design that I am not aware of. There's also the matter of
a power LED that's brighter than the sun.
Every product has compromises in its
design. I feel the BGE70 has far too many compromises of a serious
nature to recommend buying it.
If cost is a concern, my first recommendation would be to pick up a
used uninterruptible power supply. Perfectly serviceable examples
(possibly with worn MOVs, but that's a minor nit in my view) are being
tossed out all the time for want of a replacement battery. Buy the
battery yourself on the secondhand market (there isn't any significant
difference in the brand), install it and enjoy many years of reliable
service. Most of the time you can get the UPS for nothing and the
batteries for less than $20. Sometimes you can even modify a unit to
work from a cheap lawn tractor or car battery, so long as you know how to do so safely and don't massively overload the unit!
bringing a rejected UPS back to life with a replacement battery
isn't what you want to do, or doesn't fit into your lifestyle for some
other reason, save up a little money and buy something bettter. A 500
volt-amp or ideally better UPS will run small loads for a VERY long
time, much longer than these small purpose built units could ever
manage. They'll have proper three pin grounded outlets and many models
offer supplementary surge protection for communication lines and even
the capability to communicate with a computer or other device so that
an intelligent, graceful shutdown may be performed. And when the time
comes, they'll take a standard, inexpensive replacement battery.
Go Back >
Review Copyright © 2015-2016
William R. Walsh. Some rights reserved. Please view the terms and
conditions, available from the top level page of this server, to see
the terms governing the use of this material. Initial draft written
around September 25th, 2015. Final review edited and published November
25th, 2015. Updated November 30th, 2015 with pictures of output
waveforms and the internal primary circuit board. Updated again on
December 10th, 2015 with additional MOV surge suppression related
information. Charging circuit technical details updated on January 3rd,
2016. Possible explanation of code names added January 4th, 2016.