I expect we all have things in our wardrobe that would appear to be bizarre to just about anybody else.
Being both a journalist -- to some extent or another -- and an engineer (also, arguably, to some extent or another seeing as I am so totally analog) I have a large number of corporate shirts in my closet donated during visits to the offices of the companies whose products I track. But some years ago my spouse gave me a T-shirt that had a circuit on it: which I would probably not allow myself to be publicly seen in.
Nice little circuit and the immediate thought is that, "this is obviously an FM ratio detector!"
Wrong…
The additional secondary on the input transformer (if it is a winding, and not an RF choke -- á la Foster-Seeley discriminator) has no dc path to ground and what's that series resistor all about? There is also no capacitor to provide an IF ac short across the main secondary winding.
The output audio is correctly taken from the common point of the two capacitors after the diodes (which are incorrectly phased) but what does that amplifier think it's doing? A tuned collector circuit? And I guess the collector and emitter-follower pair of outputs gives you balanced audio…
I was reminded of the shirt when a news release from ON Semiconductor was brought to my attention by Lee Goldberg. The release was about a reference design for an 80 Plus power supply to fit in an ATX enclosure. The block diagram provided by the company looks like my shirt…
I've seen some interesting simplified block diagrams in my time, especially with internal IC architecture where the need to confuse competitors a little (or a lot) is considered important. But this block is wonderfully absurd! I'm particularly fond of the transformer sitting in series with the gate of the co-phased upper storage switch (clamp, in ON Semi parlance) and the 3.3 V "post regulator replaces mag amp." I doubt if there is a significant percentage of readers who even know what a magnetic amplifier (saturated amplifier) is, let alone remember what an Oersted is.
I had to go to the supporting documentation to find what the real deal is, but even there there are no full circuits given -- only the pieces given in chunks of separate PCBs. So, I ended up visiting the individual data sheets to find out what was really going on, particularly to find out how the ICs themselves are being powered. The NCP1017 at the bottom of the block diagram, by the way, is there to supply a standby 5 V (up to 15 W) and uses an on-chip 700-V rated MOSFET.
I will leave Lee to point to the reference design green-power virtues -- if he decides to review it -- but I have one other relevant, I think, comment. The NCP1562 prime SMPS controller has a data sheet where the application example uses a 36 V to 72 V dc input. That is not the case here: at some points in this circuitry, with a universal supply input, there will be lethal voltages -- with the highest at about 385 V. Not a piece of work to be approached lightly and skilled use of test equipment is required.
