Behind The Front Cover
Here is the front cover. Along the left and right sides, you can see four hooks that slot into the side panels to hold them together. From this angle, they look like slots themselves. Further inside but still parallel to the sides are two columns of translucent gray-brown plastic holding the glossy face panel in place. The bottom edge features three key holes for the battery door’s two tabs to slot into. In the top half, you find the LCD and control module, which has a simple chip-on-board (CoB) device.
Behind The Rear Panel
It looks like the 1000PFCLCD uses the same coax surge protection as the LX1500. Also similar is the lack of a common-mode choke between the cord and protection-only outlets; the protection is downstream from those outlets instead of upstream.
The phone/network surge protection board is different. It also has the wiring fault indicator circuit on it, along with a header to receive a cable from the motherboard providing access to live and neutral (the thin black and white wires in the bottom-left) required by the wiring fault indicator function.
Downsizing
While the cord uses #14 conductors, everything beyond the “protection-only” outlets is #16, which should be fine as long as the site wiring is correct (no live-neutral inversion), the UPS’ own current limiting on the battery backup outlets is working correctly and there is no fault on the motherboard causing it to draw grossly abnormal current without tripping a breaker.
Strips
We find the same pinchy fingers for live and neutral as the LX1500, and the same torque bar ground connections as well. Soldery-weldy joints between the wires and strips look a little dodgy though.
Crushed
Wires were attached to the strips using spot-welding. An electrode compresses the tinned wires’ strands together, then current gets passed through the strip and strands to fuse them together. Here though, the electrode appears to have had a sharp edge that partially gouged its way through the strands along the upper strip edge. In a high-vibration environment, this could cause premature mechanical failure. It could also cause slight hot-spotting under high electrical load, though the strips should provide enough heat-sinking to prevent it from becoming a real concern.
Network Surge Suppression
The layout may be different, but a quick glance at the number of diodes around the RJ45 connectors quickly points towards a similar if not identical circuit as the LX1500’s. If I wanted to be optimist, I’d guess that the messy design was an attempt to keep signal pairs together. But a quick glimpse at the bottom layer shows that some signal pairs are not even routed on the same layer. The layout change was mainly about reducing board footprint.
Quick And Dirty
Whoever manually soldered the ground wire to the circuit board was in a hurry, slipped, scratched the solder mask, did not apply enough heat for the solder to wick through the strands, and we get a joint that looks like the solder tented over the strands instead of forming a uniform solid joint around the whole wire and pad. The wave-soldered joints, on the other hand, look great.
Not A Fan Of The Fan
Sitting just above the transformer on a small bracket is a 40mm fan that kicks on immediately whenever the UPS switches to battery power and remains on for a few minutes after AC power is restored. With no ducting guiding airflow around the transformer or preventing it from recirculating through the fan, I was really surprised to actually feel a noticeable breeze coming out from the top vents.
And yes, this little screamer is no exception to the “small fans are loud” generalization. Considering that the motherboard has a thermistor going behind the transformer, I was surprised that CP does not at least wait for the transformer to warm up a little before turning on the fan at full speed.
Transformer
Instead of having a center-tapped winding on the low voltage side and operating the inverter by alternately driving one half, the 1000PFCLCD uses a full bridge driver, which allows it to use only one winding there. On the high voltage side, I suspect that the AVR primary is the thin wire connected to black and blue doubling as the inverter’s output winding, which may explain why a fan is needed. CP used the absolute minimum wire gauge necessary to provide the 8A of backup power the UPS is rated for (0.9mm including the enamel coating, which means #20 gauge) and may get warm very quickly under heavy load. The buck/boost winding, on the other hand, is 1.5mm thick including the enamel, which puts it about half-way between #15 and #16.
The Main Course
Can you spot the biggest difference between this board and every other UPS I have opened in the past eight months? It has three heat sinks in its inverter circuit instead of two. The two half-width sinks provide the low-side drivers and double as transformer cable terminal blocks, while the wide one hosts the high-side drivers. A small piece of T-shaped scrap PCB inserted between fins acts as a spacer to prevent short circuits between heat sinks.
The other area of interest is the bottom-right corner where all the relays, surge protection and filtering occur.
MORE: APC BE550G Tear-Down
MORE: Tripp-Lite SMART1000LCD Tear-Down
MORE: CyberPower LX1500GU-FC UPS Tear-Down
