A Look Inside & Component Analysis

Before reading this page, we strongly suggest a look at this article, which will help you understand a PSU's internal components much better. Our main tool for the disassembly of the PSU is a Thermaltronics TMT-9000S soldering and rework station. It is of extreme quality and is equipped with a matching de-soldering gun. With such equipment in hand, breaking apart every PSU is like a walk in the park!


The PCB is sparsely populated. Especially the lack of large filtering electrolytic capacitors on the secondary side caught our attention. SAMA obviously mostly relied on polymer caps for the filtering of the rails, and we also noticed an increased number of monolithic ceramic capacitors that act as bypass caps, there to filter ripple.


We like the neat design and lack of power-transfer cables, which increase power losses and affect the PSU's efficiency, especially at higher loads. There is only one heatsink, and it houses the bridge rectifiers, the APFC converter's active components, and the primary FETs. The main transformer's design is interesting as well, and we will take a look at its design in another paragraph on this page. There are no heatsinks in the secondary side since the +12V FETs are installed on the mainboard's solder side and are cooled down by the chassis.


The AC receptacle hosts the first part of the EMI filter consisting of two Y caps and a single X cap. The second part of the EMI filter consisting of two Y caps, one X cap, two CM chokes, and an MOV after the bridge rectifiers is on the mainboard. All in all, the EMI filter is complete, and we will see how it performs in our EMC pre-compliance tests.


Two powerful bridge rectifiers are used. These are by Shindengen, and their model number is LL25XB60. These bridge rectifiers can handle up to 50 A current combined, which is clearly way more than an 800 W PSU needs.


The APFC converter uses two Infineon IPW50R140CP FETs and a SCS210AG boost diode. The bulk caps are two Chemi-Cons (400V, 2x 270uF or 540uF combined, 105°C, KMQ series, 2000h @ 105°C) with a low combined capacity, so we don't expect this unit to meet the ATX specification's hold-up time requirements.


The APFC controller is on the mainboard's solder side. It is a Champion CM6502S, a controller that is more efficient than the most commonly used APFC controllers around.


The primary switching FETs are two Infineon IPW50R140CPs SAMA arranged into a half-bridge topology. They are backed by an LLC resonant converter for reduced energy losses.


The resonant controller is on a small vertical board. It is a Champion CM6901 IC that is commonly used in very efficient units.


The main transformer uses an interesting design. For starters, it isn't soldered to the mainboard through pins but through much larger blades, which decreases impedance and, as such, reduces energy losses. The transformer is also surrounded by a metal shield; it acts as an EMI shield that also removes heat. A thermistor is attached to this heatsink, and it provides information to the fan-control circuit. This transformer's density should be higher since it is kind of small for a unit with a capacity of 800 W. Right next to the main transformer is the LLC resonant tank.


The +12V FETs are installed on the solder side of the mainboard and transfer their heat output onto the PSU's enclosure for dissipation through a pair of thermal pads. In total, eight Infineon BSC010N04LS FETs are used in such a way.


Ripple filtering in the secondary side is mostly handled by polymer caps. However, each PSU has to use a small number of electrolytic caps - in this case six Chemi-Cons KY series caps. One solid reason behind doing so are their increased ESR values as compared to polymer caps, which helps in avoiding unwanted oscillations that can lead to instability issues. While low ESR is also important for ripple filtering and less heat output, minimizing ESR in PSUs isn't a good thing either. Here is a very informative paper we found on this subject, written by Dr. Erik Reed.


Several monolithic ceramic capacitors filter high frequency ripple.


Two vertical boards house the DC-DC converters. Each of these uses four Ubiq QM3006D FETs and an Anpec APW7073 controller.


The 5VSB circuit is regulated by a PFR20V45CT SBR, and the standby PWM controller is a EM8569A IC.


The modular board is directly soldered onto the mainboard, which avoids the use of power-transfer cables as such cables would increase EMI noise and suffer from significant voltage drops and power losses. There are only two polymer caps at the front of the modular PCB for some additional ripple filtering.


The supervisor IC is a Weltrend WT7502 with only the very basic protections. It is supported by an LM393 op-amp. We also spotted an Si8233BB isolated driver on this side; it most likely acts as a high-side driver.


We spotted a MC34063B buck/boost switching regulator at the PCB's backside.


Soldering quality is pretty good overall. We didn't find any long component leads, and even the enhanced PCB traces are nicely soldered, so we have no complaints to make.


The cooling fan is by Sanyo Denki, so it is of very high quality. Its model number is 9S1212L402 ( 12 V, 120 mm, 1500 RPM, 48.1 CFM), and its dual ball-bearings will allow it to last for longer than others in tough environments (40,000h @ 60°C). Our only objection here is that a larger fan could have been used for the same amount of airflow at lower RPMs, but this fan's high efficiency and low maximum speed still allow it to operate quietly overall under normal conditions. Its fan profile is quite aggressive once operating temperatures exceed 30°C, though.


The plastic baffle on the fan that directs airflow toward the rear of the PSU is dangerously close to the fan's blades. This is apparently a quality control problem that most likely only applies to our sample.