Topics: Is there a major shift underway?
Is there a major shift underway?
There is a lot of talk about LLC converters, and other narrow-range topologies like the Vicor and Synqor. Preceded by regulating PFC circuits, they are impressive.
Is there a major change underway moving away from the conventional families of topologies into the esoteric?
Or is this just a small segment of the industry for those who need performance and can afford to pay for it?
09-10-2013 03:25 AM
Well, I think there are a few factors that influence your perception.
Basically, hard-switched "conventional" converters are barely mentioned in the trade press. It's old-tech. Everything is "resonant-this, resonant-that", and for good reason, I think. There is the promise of easier EMI control, lower transistor stress, and higher efficiency, all good things in today's regulatory environment. Whether that promise is realized is, of course, up to the designer.
In the case of LLC, there are many new control chips, some containing output MOSFET's, along with lots of good design and modeling software to aid in the design process. Of course, it still comes down to what your prototype does on the bench, but the chances of an experienced designer getting close from the start improve significantly. It's true that you can't beat something like a 3524 for a low-cost design, but by the time you've attached a transformer and some EMI components, the cost of a "better" controller can be absorbed.
I'd also like to think that we engineers are always on the look-out to try new things. Personally, I find these more-modern QR topologies quite interesting, and I'm using more of them if they fit an application. And the more attractive the chip vendors make these things, the more they'll sell and the less they'll cost. (He said hopefully....)
At least that's my perspective. What do you think?
09-10-2013 03:25 AM
Its interesting to me to see the adoption going on now of new technologies. The LLC and other resonant converter have been around since the 80s, but something always held them back.
Maybe the mandatory PFC has changed the game because these converters always struggled with wide range operation.
But if you've got to go fast and rugged in your design cycle - the old tech still probably has its place.
09-10-2013 03:26 AM
I like old tech! When you need to bang out a product in short order, there's nothing like an old tried and true design!
What you say about resonant converters reminds me of Class-D audio amplifiers. The design theory has been around for decades, but it wasn't until the 90's that you could find power MOSFET's that were up to the task of switching reliably and efficiently at hundred's of kHz. Today, "everybody" has an array of devices with the low Qg, low Rds-on, and fast body diodes needed. And the various half-bridge controllers geared towards Class-D have gotten quite good, too. Class-D is very "in"!
I guess it's one thing to have the theory but another to have the tools to implement that theory. Today, we have an array of dedicated LLC controllers and predictable power devices. And it doesn't hurt to have some complementary technology to better enable these new/old techniques, like PFC front ends, as well as a favorable regulatory environment.
09-10-2013 03:26 AM
That's really a very good point. Actually in the field which i have been working for more than 4 years, i witnessed the transition from flyback to LLC year by year. Firstly we tried to get rid of classical flyback using Quasi-resonant but only with valley switching benefits because of the mosfet breakdown voltage limitations. And then the manufacturers made a marketing with their Asymmetric HB solutions still keeping the PWM tradition. That didnt last so long and it all ended up with using PFM and mostly half bridge LLC. Actually we are using LLC even at 70-80W power levels because of its thermal and EMC performances. Surely the mosfet technology helped and the mosfets with lower-trr internal diodes appeared in the market which made the half-bridge mosfets more reliable during the switching. In conclusion when our requirements are around 50W flyback is still very practical, easy to deal with and surely cost effective for us. But if we need a slim design with a limited pcb space and good cross-regulation in dual outputs we are definitely utilizing LLC. Is there something wrong in what we are doing?
09-10-2013 03:27 AM
Well 70-80 W is definitely pushing a flyback hard. I don't know how many people would immediately go to an LLC from the flyback rather that some other topology.
However there is nothing "wrong" if you are comfortable building these and they work over your voltage range, Lots of people complain the LLC struggles with wide range input but if you don't have that problem then it's good.
Perhaps its a good thing that higher tech wins out rather than the usual push in our industry for the dirt-cheap solution. That way, the engineering is still needed.
09-10-2013 03:27 AM
Ray's right. There is no "wrong" way to do things. To paraphrase an old saying, if you ask 10 engineers for an opinion, you'll have 11 opinions.
It's good to see that engineering is still needed. So many aspects of design have become "commoditized". By this I mean that someone has done your job for you. I work in audio, and a lot of designs that were once custom are now commodities, like low- and medium-power amplifiers and DSP.
Another example is the proliferation of evaluation boards from semi manufacturers. They're intended as a general-purpose platform to evaluate a chip or chips, but they frequently get used in products "as is". Just go to ebay! But this will rarely offer the best-performing or most cost-efficient way to do things. We've become a nation of "good enough".
09-10-2013 03:28 AM
Hi guys, LLC picked up in the merchant market (high-volume stuff) perhaps 10-15 years ago when flat screens where getting thiner and thiner. LLC was good because ZVS on the half-bridge helped to push frequency higher with reduced-size magnetics. However, most of the designs I've seen still operate at 70 kHz for EMI reasons. PFC-wise, an interleaved borderline type is associated for thickness reasons too but tends to disappear now. The LLC converter is not an easy guy to deal with, despite the abundance of literature. It's small-signal analysis requires complex tools and fault tolerance is difficult to deal with. You have to know that power supply engineers spent 90% of their time coping with fault management (open or short elements) rather than on design itself. Now that flat panels are going down in power, the flyback comes back again and we see less LLC converters in TVs. In notebook adapters, it is not uncommon to find QR flybacks up to 180 W - yes, I know :-) and single-stage flyback up to 90 W. Low-cost, ease of design and low standby power, something the LLC is still not able to compete with!