Recently I developed a MATLAB tool to dimension LLC converters in boost mode. The approach is to dimension the converter such that under given input/output voltage and series resonance frequency (read: switching frequency) requirements the output power can be delivered. The actual steady state power transfer function is calculated with the tool. The results have been verified by means of Spice. When I compare this power transfer function with the results from the commonly applied First Harmonic Approximation (FHA), I notice that the FHA consistently gives too conservative estimates. In other words: an FHA based design would be over-dimensioned in terms of power transfer. Does anybody recognise this? What would be the practical consequences of the conservative FHA dimensioning? (for your reference: I wrote a short technical note on my observations which can be found under my publications or the link below)
http://www.zeonpowertec.com/resources/zinuzz/ZiNuZZ-FHA-analysis.pdf
Hello!! I don't understand clearly the whole text. But it's well written. I think you are doing a good job. But perhaps if you investigate a little more you'll find the answer you are looking for. That is, a better power transfer function. Or not? I don't know.
Hi Alfonso, thanks for taking the time to answer my question. The thing is that in my opinion the FHA model is seriously inaccurate. Most designers use this FHA approach because of its simplicity and insightfulness. And yes, a different approach like used in ZiNuZZ gives a more accurate (power) transfer function. My worry is related to practical consequences. If the FHA prediction is a factor 2 (!!) off then this for sure must have consequences for the chosen power components (a factor 2 to big??)... Am I overlooking something???
The FHA method has its limitations and computational tools are more accurate in design optimization. Thus, the converter can be designed without excess, getting smaller and more efficient to a given power. A computational method will result in a smaller converter, because if you consider the other harmonics and the energy transported by them the result will be more accurate. But you can reach a good accuracy with FHA if you make an iterative design (riding a prototype, test, and optimize, back testing, etc.). What do you have, time to spend in a computational method or time to spend in laboratory?
Hello Adriano, thanks for your contribution to the discussion. I fully agree with the fact that the FHA design must somehow be bigger than needed. In FHA most of the inaccuracy is introduced by the output load transformation into an AC resistor. In my opinion doing iterations in hardware in the lab is far more costly and time consuming than having a first time right estimate by using computational method that is readily available. Thanks again for sharing your insight!
http://www.zeonpowertec.com/Portfolio/ToolShop/Matlab/power-converter-matlab-llc-analysis-design.html
All methods to calculate are a simplification of course. But it's right that the computers today are always used for those kind of things.
your refer this paper it will give more information regarding LLC.
Analysis and implementation of LLC-T series parallel resonant converter using fuzzy logic controller
http://www.ajol.info/index.php/ijest/article/view/64010
- Badges
- Science topic