Bests & Worsts Reviews from Amazon

Good PSU. I decided to open this Evga 450BT and check is circuited, I was surprised by what I found. The unit uses common forward topology (check this out in order to understand this review better). Starting on the mains, the unit has a Ceramic fuse right behind the power switch, I can't read its rating, but hey nice to see a ceramic fuse instead of a glass one. Altough it is very badly placed and can't be replaced easily, you will need to desolder the smaller board behind the power switch in order to change the fuse. Moving on we have 4 Class X2 caps and 7 class Y caps, a bit overkill, we also have 2 common mode inductors at the input, this is for EMI suppression, we also have a varistor after the fuse, (can't read its rating as it is inside a heatshrink, although it is very likely 275V AC), what this does it that if the mains voltage gets too high, the varistor shorts the mains and opens the fuse, Now we get to the main bridge rectifier, it has no printing on front, this is really odd (it might be facing the heatsink), however it isn't much to worry about, the high voltage side of these PSU's operates with low currents. After the rectifier we have a 0.47uF polypropylene cap (again more EMI protection), along with the main filter cap, which is a 400V 270 uF teapo cap rated for 105 degrees, the main cap comes after a relay and a thermisistor, the thermisistor is there in order to reduce the high start current of the main capacitor, there's also a huge inductor used for the active PFC. The main switching transistors and the PFC transistor are 5R280CE, these are also with the bridge rectifier heatsink and are rated for 11 Amps are 100 degree centigrade, The secondary uses two S30M60C in parallel for the 12V, each are rated for 30A at 100 degrees (giving 60A total at 100 degrees), again more overkill for a unit rated at 35A for 12V. All secondary caps are also teapo 105 degrees. 12V: The previously mentioned diodes -> Common mode inductor (shared with 5V) -> 2200 uF capacitor -> Another inductor -> 1000 uF capacitor. 5V: One S30M45C -> 1500 uF cap -> Inductor -> 1500 uF cap. 3.3V: Mag amp -> One S30M45C -> 3300 uF cap -> Inductor -> 1500 uF cap. -12V: An SMD diode under the PCB (Can't read the model) -> 220 uF cap -> Inductor -> 330 uF cap. 5VSB: One S30M60C -> Inductor -> 1500 uF cap. So almost all rails have their own pi filter. This is essential to keep ripple as low as possible. The controller chips are a FAN4800C and a EM8564A, on the secondary we have a GR8323. The output wires are all black, 18 gauge and rated for 80 degrees. Overall the unit is very good ans should be perfectly capable of delivering its rated wattage, and even a bit more depending of how the over current protection is tweaked. The level of overkill of this unit is something you would expect of a unit worth several times the value of this one. Edit: The main bridge rectifier is a GBU806 (had to remove the heatsink), nice. Edit2: The varistor is a 10D561K, not the right part for this, it's clamping voltage is well beyond of what the GBU806 can take. They should have used a 10D431K instead. Edit3: I had one unit die. Almost all diodes at the secondary failed short, and there was damage to the relay circuit at the primary side. Really odd. Edit4: Fixed it. I still don't know what caused the diodes to die, I suspect that since this unit lacks OTP, was in a really hot place (ambient temp gets as high as 37 degrees centigrade very often here) with not much airflow for its fan, was running a PC with one RX480 and one GTX1060 mining 24/7, and that the unit controls the fan speed (likely uses the voltage drop across the shunts as reference for the fan speed). The diodes likely degraded because they were too hot and the fan was not maxed out. I also had replace a transistor at the primary side that turns the unit on, Apparently the unit suffers a hysteresis issue, when it failed I could hear the relay switching rapidly. The transistor is an NPN one with a Y1 mark on top of it. The transistor had a 16V zener diode next to it that also failed. Edit5: There's an NTC next to the secondary heatsink (was hidden by some black glue) that controls de fan speed, and it isn't working, I hit it wit the soldering iron and the fan speed doesn't change in the sightless. That explains it all, that fan always ran at minimum speed.