1kW LLC converter

Sch3matic
A source for a good heatexchanger for free might be an a old X-ray tube from CT scanner,
they work from 110/220 and are a closed system with nice pumps, I used them a couple of times if you like you can use oil instead of water.Should be able to find one at a clinic or from a service engineer that service scanner's in your area.
 
Two things are bothering me. My eyes are not what they use to be, and most of what I am seeing here seems to be in the direction of using Surface mount devices. But I will adapt as best I can with some visual aids. The other is that how to use a secondary inductor on the output side of the transformer for the full bridge output, that is a plus and a minus. Maybe I am over complicating this. Chas1 or anyone, do you have an example of a FULL bridge for the LLC? Which direction are you headed in the second L in series or in Parallel ? It is complex, but the main reason I am going here is the 110 and 220 input to meet the universality of the power amp input needs...

The MCU challenge is doable, I can just see a program now. AND you can lock it up.

The Arduino chip is cheap and PC programmable, both great.

Let get it started, this is the way power supplies will be designed in the future, that is for sure. I worked on a printer for Xerox a few years back and we literally programmed the driver motors h bridge to produce power when different accessories were hung on it. Various loads with the same H bridge....

I found this too. http://www.onsemi.com/PowerSolutions/supportDoc.do?type=tools&rpn=NCP1395

Thanks
 
Last edited:
Basically you can monitor voltage and or current. You can sense temperature. You can then make decisions based on these variants. Then you can switch different external components such as resistors or capacitors into or out of the external circuitry of the controller. The logic diagram of the controller is fixed.

For instance you could have the controller act differently for 110 volt input vs 220v.
You would have one set of resistors, capacitors and such for one and parallel or serially connect others. with MCU controlled solid state switches. You could switch in/out loads electronically. Lots of stuff. Programs are sets of logic that turn on or off switches. Just as the controller does, but it is fixed, MCU are flexible.
 
I break the ice and i attach the schematic of the SMPS500R-24, the one which i have made last week and is available here: Connexelectronic

how to use a secondary inductor on the output side of the transformer for the full bridge output, that is a plus and a minus. Maybe I am over complicating this. Chas1 or anyone, do you have an example of a FULL bridge for the LLC? Which direction are you headed in the second L in series or in Parallel ? It is complex, but the main reason I am going here is the 110 and 220 input to meet the universality of the power amp input needs...

The LLC converter does not need an output inductor, either HB or FB topology. The full-bridge converter which i made uses 4 switches, driven by a gate drive transformer, UCC25600 controller (i chosed this one bcz can conveniently adjust the dead-time) and the PFC side i made with UCC28070 interleaved PFC controller. at that time, about 4 months ago i have asked for some support on the TI page but.... they might have missed the message 😡 UCC25600 Based 3KW LLC DC-DC converter - Isolated Controllers / Motor Controllers Forum - Isolated Controllers / Motor Controllers - TI E2E Community
The MCU challenge is doable, I can just see a program now. AND you can lock it up.
The Arduino chip is cheap and PC programmable, both great.

is not enough just a simple MCU, for general purpose, must use a dedicated one, which have hardware PWM, to generate the frequency and fast high resolution ADC for feedback. also the control loop must be implemented inside so it should have enough resources for that too. there are already available such mcu's 😀
 

Attachments

I used EER4220 core, 6mm spacing between primary and secondary, i use litzwire, 32x0.19 for primary, 16 turns, 32-40-50-64x0.2mm for secondary, depending on the output voltage and number of turns. for 24V 2x2 turns, 48V 2x4 turns, 72V 2x6 turns. Lm=150uH Lp=30uH. the centre leg of the core is grinded to about 0.4mm. after assembly, the transformer is impregnated with special transformer resin. the core which i used has the best area/volume ratio from all this types of cores, which allows to make a very compact smps.
 
first few pcs i grinded myself (together with part of nail 😀) then i send them to factory to do this. the spacer is made of what is that called ? wrapping cords ? is used for cables and some others, to keep them together. any other suitable spacer can be used, but the distance between windings should be at least 4-6mm for safety reasons, but also, this distance dictates the leakeage inductance value and coupling.
 
haha
I have etd(the same as eer?)44 and I have 0.2mm gap, i don't know if that would be enough. And I don't want to grind it myself, coz it wouldn't be flat no more and probably under some angle. Wrapping cord you say...hmm, will have to find something.
What about resin, is it really any special or normal like for boats or car repair?
 
ETD44 has much lower centre pole area, even ETD49 has smaller area than EER4220. that's why i said that this core is the best. with ETD44 you can get 400-500W if the frequency does not go below 100KHz. the number of turns must be higher, since the area is lower.
the resin is a special laque used for transformers impregnation, any similar can be used. the idea is to keep the windings together, to not change the Lp and better electrical insulation.
 
Nice supply,but I am trying for +/-50VDC@5A(500W) along with +/- 15VDC aux,supply will use small transformer for housekeeping and controller power,controller will be on secondary side.You are right about (Up) for controllers been there done that,a digital controller would be nice but develop cost is keeping major players away however students are trying very hard and maybe a cheap breakthrough is close.It is obvious you can contribute a lot and I realize you have a commerical venture but all your advice and help is welcome.
 
Nice supply,but I am trying for +/-50VDC@5A(500W) along with +/- 15VDC aux
then you are close to the specs of this one Connexelectronic which has almost the same schematic as the 24V one, the difference is in the output rectification and the voltage feedback which is tied between V- and V+ and optocoupler is biased from GND.

i see no reason to use another transformer for the aux. supply or to move all the control in secondary for this output power level. for example the L6599 has start-up current under 0.5mA and operating current between 5-10mA depending on the switching freq. and transistor used. this supply current can be easy get from a two windings turn on the main transformer, or for simplicity, from the AC+rectifier+10-22K/3W resistor.
 
Startup and design

Cristi

You are right about startup, A throw back to my hard switcher (transformer not needed for this design)Also the goal is 1 kW I am starting out with 500watts since one of my amps will be the test fixture for this supply(classAB) and again thanks for the input.
 
why do you think only 500w? with hard switching in not a problem >1kw. So is this because of this type of converter?


LLC converters operate in variable frequency domain to achieve ZCS & ZVS at various loads, thereby the transformer core needs to be bigger in size and also gap at center pole is needed to stay away from saturation. Seldom these converters are used beyond 1-2kw range.I don't find their advantage in audio smps.

A simple quasi-resonant smps[QSC patented it under the name "discontinuous resonant mode" but virtually used by every pro-audio company and chinese copy cats] can pump out more than 1.5KW from ETD44 if things were done correctly. Just put an inductor and cap in series with transformer primary [calculated values of L & C to achieve resonance at specific F] with a normal fixed frequency half bridge smps and you get current thru trafo following sinusoidal pattern enabling ZCS at high loads.

For more stringent need switch to Phase Shifted Full bridge Resonant converter to get ultimate in robustness and efficiency at multi-kilowatt levels.

Kanwar
 
Last edited: