The PSU follows a somewhat nonstandard variety of the dual-transformer design. It is a circuit with synchronized transformers. There is nothing new in the idea of using two transformers, though. With a high-wattage PSU, a single transformer may not fit within the required dimensions, so it is logical to split it in two transformers of half its capacity. The problem is to distribute the load in such a way as to avoid situations when one transformer is overloaded while the other is idle. This could be observed with Enermax Galaxy DXX power supplies. For them to be stable, the load had to be connected in such a way that each of the two transformer worked at a load of a few dozen watts.
So, this is where the transformer synchronization fits in.
The design shows the circuit in a simplified way, both in the high-voltage (in fact, two transistors control each transformer, allowing Enermax to talk about a “quad converter”) and low-voltage (the Revolution 85+ uses transistors instead of diodes as I will discuss shortly) sections, but that’s unimportant for understanding the point.
So, we’ve got one PWM controller that controls two forward converters Q1-T1 and Q2-T2 in such a way that the transistors Q1 and Q2 are opening up alternately.
Each transformer has a dedicated rectifier and a choke in which power is accumulated, but after the chokes the two circuits join into one circuit that has ordinary smoothing capacitors. Since the transistors Q1 and Q2 open and close alternately, the impulses arrive to the chokes L1 and L2 in antiphase.
As a result, the circuit’s operation recorded in an oscillogram looks like that:
The impulses that come alternately from both converters sum up but do not cross each other in time. As a result, the oscillogram at point 3 (i.e. at the PSU output) looks exactly as if we had one converter working at a double frequency. Thus, the problem of load balancing between the transformers is solved: the circuit is built in such a way that each of them delivers half the total output power and, from the load’s point of view, the PSU is no different from a single-transformer one. This solution also doubles the frequency at the output filter. When the frequency is higher, smaller capacitors and chokes can be used to additionally smooth out the voltage ripple.
Increasing the frequency of a single converter is not easy: it requires expensive high-frequency transistors and expensive materials for the transformer’s core. Here, Enermax’s engineers have solved a few problems with one solution: the transformers fit into the PSU housing, the load balancing between them is ideal, the job of the output filter is easier now.
NB: You can learn more about synchronous transformers in the article Interleaving power stages – not just for buck converters any more.
Is it a new technology? Yes, it is new in power supplies but it was developed not by Enermax and not today. This is indicated by the mentioned article which is dated the year 2004.
Two identical chokes can be seen at the PSU’s output: one choke for each transformer.
A UCC28220 chip located on a small daughter card is the controller of the synchronous converter.
This is just the beginning of the circuit design peculiarities of the Revolution 85+, though. Taking a look at the heatsink that usually carries diode packs of the output rectifier (marked as D1 and D2 in the schematic above), you can find no diode packs! Instead, there are IRFB3307 field-effect transistors:
The fact is the Revolution 85+ employs so-called synchronous rectifiers in which diodes are replaced with transistors. Why?
Let’s take a look at the specifications of a typical Shottky diode which is often used in power supplies: STMicro S60L40C (a 55KB PDF file). Look at Figure 1 on the second page that shows the correlation between the power dissipated on the diode and the current. At a direct current of 20A there will be more than 8 watts of power wasted – dissipated as heat – on the diode. This is due to the fact that there is a small voltage drop occurring when the current passes through the diode – about a few tenths of a volt. If you multiply that by tens of amperes, you get a few watts.
What does the diode do in the rectifier? It opens in one direction of current flow and closes in the other. It can be replaced with a transistor that is controlled in such a way as to imitate the diode’s operation. It can be the above-mentioned IRFB3307 (a 357KB PDF). In the open state the resistance of its channel is a mere 5 milliohms. So, at a current of 20A, it will dissipate P=I²R = 20²×0.005 = 2 watts. This is only one fourth of the dissipation of an ordinary diode! Of course, that’s an ideal example, but it provides a general notion of how much power can be saved.
It is no problem to make the transistors switch at necessary moments. In the simplest case their gates are connected right to the transformer’s windings.
If a higher efficiency of control over the transistors is necessary to minimize power loss, a synchronous rectifier controller is introduced into the circuit.
NB: You can learn more about the use of synchronous rectifiers in the article The Implication of Synchronous Rectifiers to the Design of Isolated Single-Ended Forward Converters (a 433KB PDF file).
Is the synchronous rectifier circuit new for computer PSUs? Yes. We have not tested such models before in our labs. Is it Enermax’s invention? I can quote myself: “I can predict, for example, that synchronous rectifiers will be sooner or later used in the secondary circuits of PC power supplies. There’s nothing new in that technology, but it is too expensive as yet and its advantages don’t cover its cost”. That was written in 2006.
Taking yet another look at the internals of the Enermax Revolution 85+, you can note a number of small components on the card with the output connectors.
These are DC-DC converters you have already seen in the Antec Signature section above. They are used to transform +12V into +5V and +3.3V. Enermax’s engineers have used these converters to the full, moving them from the PSU’s main card to near the connectors in order to make use of the space at the back panel.
Anpec APW7073 chips are used as the converters’ controllers. Next to them there are power transistors which are cool and do not require a heatsink. The function of a heatsink is performed by the copper foil of the card the transistors are soldered to.
On the reverse side of the card there are chokes (one for each converter) and smoothing capacitors. The connectors for detachable cables are nearby: it is to them that the voltages produced by the converters are applied, besides everything else.
There are other tricks you can find in the Enermax Revolution 85+. For example, the bridge in the photo connects two points of the same interconnect and reduces its overall resistance and power loss. These things are not fundamental and are far less interesting, though.