Inrush filtering

Most typically, power supplies have the first stage of inrush filtering mounted right on the AC plug. This is also the case with two ceramic Y capacitors and one foil X capacitor. Heatshrink is used on X capacitor electrodes, ideally, there should also be some on the Y capacitors electrodes and on the end of the AC wires, but everything is soldered tightly in here so these are more of a cosmetic drawbacks. Earth wire is tightly connected to the metal casing.

I will also talk with Huntkey about reinforcing the cover because keeping the output grill as it is without any other improvement (for example by benting piece of the metal on the left side of the grill on picture) is just awfull: you just push the AC cable into plug and the whole back side will bow. I haven't even did it yet and still you see the result so you can imagine how it looks now. Half a point down. But back to inrush filter: there are two common coils on the main board, right after fuse. To spare some room, the coils are „flying“ above one X and two Y capacitors. There is also one varistor under the coils, nicely wrapped with heatshrink.

The X capacitors (between phase and null) and the Y capacitors (between phase and ground/null and ground, nowadays also between grounded cover and ground on PCB to filter noise) are used to filter high-frequency noise from power grid (usually from the devices which lack filtration to save costs, but also from the devices which are difficult to implement filtering in). It also keeps noise produced by the PSU from getting into power grid. Also the coils do this. All these components may also assist in filtering out smaller voltage spikes from power net. Varistor is used to filter greater spikes, thermistor to prevent inrush current spikes.

 Huntkey FX500SE uses a relay to bypass the thermistor: the current only passes through the thermistor for a fraction of second to keep the current down, than the circuit switches to the relay and power flows through it without losses. This also keeps thermistor less stressed, but still, it may be stretched into heatshrink, at least for (my) good feeling :-) There is also a position for IC to intelligently discharge X capacitor(s) but is not occupied.

I have already asked Huntkey for the possibility of occupying it as I have not yet seen power supply with both thermistor bypass and IC for discharging the X capacitor(s). We will probably see this in 80 PLUS Titanium certified power supplies, but anyway, there would be some extra free points from me for the FX500SE if it got both… :-)

Primary side

Primary side begins with a LL25XB60 bridge rectifier which can handle voltage up to 600 V and current of 25 A. That is with a heatsink and up to temperature of 115 °C. With 12,5 A flowing through it, the voltage drop is just 0,92 V (0,86 V typical) what is nice, I have usually seen rectifiers with 1V drop so far. Pretty massive toroid coil is used in the power factor correction circuit. What I really dislike is lack of an extra insulation on it. Here it actually does matter, great voltage is present on the wiring and as long as it is only about a milimeter far from the closest grounded screw, it is not really safe to rely only on lackuiring of the wire for the insulation.

A nice massive hatsink is used to cool the PFC transistor and diode. It is a CoolMOS Infineon IPQ60R190E6 (in TO-247 casing) N-FET; this transistor can handle 20,2 A continuously/59 A pulsed at 650 V and 25 °C with R_DS(On) of 190 mΩ at 9,5 A/25 °C. At 150 °C the resistance grows to 0,44 Ω, that explains the big heatsink - losses would grow too quick with smaller one.

 The diode used is a Cree C3D06060A (TO-220 package), which handles 8 A continuously (up to 135 °C/27 A pulsed (up to 110 °C) at 600 V. The maximum voltage drop is 2,4 V at 6 A and 175 °C, dropping to just 1,8 V at 25 °C. An Infineon ICE3PC501G IC is used as a PFC driver. It has 14 pins (DSO-14 package), yet promises lowest count of other components used and it is true that besides the basic power components for boost type circuit (a coil, a transistor, two diodes and a resistor) there is just a pile of SMD resistors and diodes. The IC also promises high efficiency throughout the whole spectrum of load values and also a great number of protections.

Another interesting fact is even though the datasheet suggest using 220µF bulk capacitor, here is actually a 470µF penguin for 450 V. It is SMQ line from the lower-end of Nippon Chemi-Con's offering with 85 °C maximum temperature and minimum life of (just) 2000 hours, but thanks to great can size (30×45 mm) and high voltage, it won't heat as much and will be better cooled on the other hand. Thanks to that I believe, it will last very long time.

However, the real capacity per watt ratio is not that good as it may seem, the used capacitor has been laying around for eight years and the capacity dropped to 407 µF in the meantime. In theory, capacity may rise slightly with voltage applied, but still… Well, what, I would personally push a car away to get it (I have actually once grabbed Daewoo capacitor on street, really). Having 85 % of rated nominal capacity is still in spec and even than it is 85 % more than what would be enough for this power supply so I will give it half of an extra point…

The switching transistors used are STMicro, STP26NM60N (600V, 20/80 A at 25 °C, 165 mΩ at 10 A and 25 °C) in TO-220 package. All the semiconductors here are greatly oversized and that is very well above what I would expect as oversized. Also high-quality components from branded manufacturers are used, for that reason I add an extra point.