Electronics > Beginners
LLC vs LCC Converters for High Output Voltage
<< < (6/6)
MagicSmoker:

--- Quote from: T3sl4co1l on May 16, 2019, 12:49:46 pm ---Oh neat, haven't seen a GDT drive like that before.  Makes sense.  Think I'd put a diode across Q4 too, but it's not required.


--- End quote ---

Feel free to steal it (I suspect its been reinvented plenty of times, anyway)! It's what I consider a "cheap and cheerful" approach but which does a good job of clamping the MOSFET gates in the face of huge dV/dt when they are supposed to be off without requiring the usual mess of parts on the GDT secondaries. It's pulse width fidelity is a bit sloppy, however, so not recommended for bridges and the like.

EDIT - I forgot the best part about it: you can use a more common 1:1:1 GDT, rather than a 1:2:2
MagicSmoker:

--- Quote from: state_of_flux on May 16, 2019, 03:43:14 pm ---So, from re-reading this thread, that only really leaves me with the buck current-fed full bridge, or the LCC running at fixed frequency, fixed duty, with a buck pre-regulator to achieve control?
--- End quote ---

The LLC isn't very accommodating of a wide range of input or output voltage, so it's really off the table completely. In fact, the only times I've ever seen it used commercially is when there is a PFC stage in front of it. Also, it's running neck-and-neck with the PS-FB as the most difficult converter to design. The buck current-fed full bridge is a good choice mainly because: a) the bridge switches have a relatively easy life; b) you can control the output damn near all the way down to 0V; c) it is highly tolerant of short circuits and, of course, doesn't care about open loads; d) the buck converter is the easiest of them all to stabilize. It's main downside - and it is a doozy - is that it's still a hard-switched converter so stray capacitances, in particular, are going to be an immense source of losses. Also keep in mind that stray capacitance on the secondary of a step up transformer will be multiplied by the square of the step-up ratio. E.g., a 30x step-up transformer will reflect the capacitance at its secondary back to primary as 900x higher. Conversely, stray inductance on the secondary will be divided by the square of the step-up ratio, hence it is less of a concern with step up transformers (at least from the perspective of the primary).

Consequently, the LCC would be the much better choice between LLC and LCC if for no other reason than it automatically incorporates the stray capacitances into the shunt capacitor that goes across the transformer primary, but it will still be very difficult to get working, especially since the LCC converter has "voltage gain" and it will be destroyed if the switching frequency drops down too close to the resonance point of the series LC network (the shunt capacitance across the transformer is typically smaller than the series capacitance, though see above in which the reflected capacitance might very well be much bigger than expected). Also, the switching frequency of the LCC spans a wider range, typically, than the LLC (especially if you also need to vary the output voltage over a wide range).


--- Quote from: state_of_flux on May 16, 2019, 03:43:14 pm --- This is a project that I plan on running into next year so I do have time to attempt a more complex solution if that is the case and it doesn't seem as though I have many more options available.
--- End quote ---

Ah, well, you have plenty of time to blow things up repeatedly; so long as you have the budget!


--- Quote from: state_of_flux on May 16, 2019, 03:43:14 pm ---MagicSmoker, I have one point in regards to a previous point you made:


--- Quote --- It's not so much a matter of lower efficiency if you open-circuit a series resonant converter or short-circuit a parallel resonant one, it's more a case of instant switch destruction, though it is typically easier to protect the former from abuse than the latter.
--- End quote ---

Isn't it the case though, with the LCC converter, that it is naturally protected in overloading and short-circuit conditions due to the parallel capacitor? In my application, short-circuits can and do occur - so again this could point towards the use of the LCC.
--- End quote ---

No, the shunt capacitor actually makes the LCC more prone to failure from a shorted load, though as mentioned above, this capacitor is typically smaller than the series capacitor and so doesn't dominate the converter's behavior. In fact, a properly designed LCC can tolerate a fairly wide load range (I hesitate to say from open to short, but on paper it is capable of such). Look up a paper from TI titled slup376 for a better explanation and comparison of the LCC and LLC.



--- Quote from: state_of_flux on May 16, 2019, 03:43:14 pm ---One issue I do have with the LCC/LLC resonant converters is their controllability - it is to my understanding that both these multi-resonant converters have limited usefulness due gain variations and chaotically moving poles and zeroes in their dynamic power transfer function.
--- End quote ---

Sorta - it is probably more correct to say that the transfer function of resonant and multi-resonant converters is non-linear and often has no formal solution at certain combinations of line, load, transformer ratio, etc. And by that I mean the converter will fail to regulate correctly, latched itself at zero output, or blow itself up.


--- Quote from: state_of_flux on May 16, 2019, 03:43:14 pm ---I've seen Texas Instruments use what's called Hybrid Hysteretic Control to overcome this, https://www.mouser.co.uk/new/Texas-Instruments/ti-ucc256303-controller/. It says that the system is always stable with proper frequency compensation, but again in my case where this is most likely not okay - would it make these devices inapplicable and thus meaning a terrible control characteristic regardless of choosing LLC and LCC? Or could this be alleviated through the use of a pre-regulating buck converter again?
--- End quote ---

I didn't read the link, but understand that the usual definition of hysteretic (or the more colorful, "bang-bang") control is to turn the switch on when the output drops a certain amount below the setpoint and turn the switch off when it rises a certain amount above it; the width of this error band being the hysteresis. Hysteretic control is woefully underappreciated, in my opinion, because it needs no frequency compensation and has the fastest transient response possible. It's downsides are variable frequency operation and a need to balance the upper limit of such against the tightness of the setpoint regulation (whether the setpoint is for output voltage or current or power).


--- Quote from: state_of_flux on May 16, 2019, 03:43:14 pm ---The load of my converter is typically pulsed on/off with a pre-determined duty cycle, and the regulation is quite stringent - so I believe this to be an important point I neglected to mention.
--- End quote ---

*groan* So you need 0-10kV output and up to 50mA, tight voltage regulation, tolerance of shorted and open loads, fixed frequency, compact size, high efficiency... anything else I forgot?
T3sl4co1l:

--- Quote from: state_of_flux on May 16, 2019, 03:43:14 pm ---Also, T3sl4co1l, I have dropped you a quick message in regards to consulting, in case I am still banging my head against the wall in late Summer.  :palm:

--- End quote ---

Got it, will check back in later :)

Tim
Navigation
Message Index
Previous page
There was an error while thanking
Thanking...

Go to full version
Powered by SMFPacks Advanced Attachments Uploader Mod