It's time to get started on the dissection. Our first step is removing the battery door so I can pull out the battery itself. Chunky tabs? Checked. Chunky slides? Checked. Reasonably chunky reinforcement ribs? Checked. Consider this yet another door able to withstand substantial shipping abuse.
What do we find inside the battery compartment, hidden behind another one of those metallic APC stickers? A CSB Battery HR1224W Absorbed Glass Mat (AGM) lead-acid battery with its key specifications for a high-discharge application printed directly on it. This battery can provide 24W per cell (144W total) for a 15-minute discharge. How much is 144W/15 min stretched to 90 minutes? A little over 24W of sustainable power draw, which sounds about right for the 20W output power stated on the packaging.
Just in case you missed it, the squiggles of approval strike again: you can see the stain on the back of the label and a vertical strike through the ‘W’ in “HR 1224W” on the battery.
Mystery Gray Goo And Date
As I suspected, the battery is quite a few months older than the UPS itself: the month dial on the left points to 3 (March), making it seven months older than the UPS’ ATE slip date. By the time this battery got through my door, it was already two years and three months old.
The bottom part of the picture shows the caps on the battery's last two cells. They are the only ones with this whitish residue all around. Did someone splash lead acid between the cells during flooding and forget to clean it after installing the caps? How many months passed between manufacturing and activation/flooding?
Both batteries measured a healthy 12.7V before I connected them, which at least tells me that they did get some degree of maintenance charging between whenever they were flooded and my order.
How does the HR1224 fare compared to other batteries? Much approximation went into putting this table together thanks to various manufacturers skipping different discharge rates. I also had to approximate the HR’s 20h rating based on physical dimensions to answer that question. It is 20% thinner than the GP1272 (51mm vs. 65mm) with all other measurements being equal, so I filed it in as 6Ah. An over-estimation here translates into under-estimations in the table, meaning that actual performance should be better. From the look of it, CSB improved its batteries’ performance by about 10%, pulling 10-20% ahead of the two other power sources represented here.
Yuasa could have won this battle, but its NXP’s specifications only go to 20 minutes. I also had to end the table at 90 minutes since that is how far the HR1224’s specifications go.
Hunting For Screws
Breaking in is as easy as it gets. Only four screws hold the whole thing together, and all four of them are at the bottom of the battery bay. Once they are out, the cover comes right off.
I took this opportunity to take some quick measurements, hoping to answer why the battery was rattling around so much. It appears that APC’s measurements were one millimeter off in every direction. It could have molded some leaf springs (beefier than the puny one you can see at the bottom of the battery bay) to pick up that slack. On the plus side, the battery shouldn’t get stuck inside as it swells with age.
Little space goes to waste here: the USB supply board sits between the battery bay and left side just above the ventilation slots, the main board occupies most of the bottom side with an inch to spare on the right side, and then you have the cord and breaker as the right side’s only citizens. Of course, we can also see the back of the outlet strips that get wedged between the battery bay and front face when the unit is assembled.
Connectors. Connectors Everywhere!
One of the last things I expected to see in such an inexpensive unit was connectors. But we're treated to spade connectors for the power cord, a two-pin connector for the AC output, and four-pin connectors at both ends of the cable between the main board and USB power board (which is driven by the main board’s 12V rail instead of AC, as was the case in the LX1500).
From this angle, it looks like we have one surprisingly busy box. Have you guessed what the BGE90M and the old BX1000 have in common yet? Look at the six tiny heat sinks surrounding that second transformer in the rear.
Instead of having a breaker rated a few amperes higher than the cord it is attached to, we have a 4A breaker on an 8A cord. Soldering on the breaker’s tabs is a little blobular, but as long as the solder wets them and wicks into the wires, a little too much solder is better than too little.
With the power cord, USB power supply, and main board taken out, the last remaining bits left inside are the two outlet contact strips with their attached power leads and in-line common-mode choke. As you can see, the metal strips actually have five outlet positions stamped into them. APC and most other manufacturers reuse the same strip designs and simply omit holes on their enclosures to vary outlet count. Pull gently on the strips from both ends and they come out with minimal effort.
To keep the button’s light pipe in the bottom-left lined up with the main board’s LED, APC put a small piece of unshrunk heat-shrink tubing on its stem to act as an alignment guide.
That’s How It’s Done
As usual with mid-range power products, we get “pinchy” live and neutral contacts. Apart from one rogue strand on the neutral wire, these spot-welds are perfectly executed. Just enough current was applied for just long enough with just enough pressure to fuse the strands into one solid mass and affix them to the brass strips.
MORE: APC BE550G Tear-Down