Calculation of SMPS resonant parts

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Hi all

Is there software to calculate this on picture? I am from myself a very bad calculator, I am shure not good with numbers, I can but need then help from software or calculator, I have one but seems I get strange results, maybe the explane on the project is not clear.

I have seen in past software who can do this kind of formulas, so maybe someone can tip me.:D

thanks
 

Attachments

  • ScreenHunter_94 May. 03 14.16.jpg
    ScreenHunter_94 May. 03 14.16.jpg
    89.6 KB · Views: 356
  • ScreenHunter_95 May. 03 15.29.jpg
    ScreenHunter_95 May. 03 15.29.jpg
    60.2 KB · Views: 368
Have you tried Excel?

No, because I have never work with it and I think learn it is a big stage, maybe matlab?

I do use LTspice and sim, it works also to find resonance points.

I do not now if a fase shift controller chip can be usedm now it regulate by change frequency witn a vco, a fase shift do only shift one output but keeps duty cycle de pendent of ramp pin.

regards
 
Last edited:
Maybe but for me, a 29 years old guy it is difficult, I am not so good at software.

but maybe there is special software for calculations? sometimes the formulas are so unclear if these are henrys, microhenrys, ohms and hertz or kilohertz.

I do now there are scientificic calculators also on the internet.

But I am not such a mathboy like in stargate universum.

regards
 
Hi

Yes I have found these but seems that for smps resonant calculations things are slighly different, like load independency so formula is different, )except serie resonant LC) but for resonance I did use this, internet is full with them but there are much more calculations for the smps supply, I also did find smps calculator who let calculate on base of output and input voltages and different topologies including resonant type, but still there is much to learn here, so I have now printed 100 pages of info and go into the woods with a pot of beer and read surrounded with nature and singing birds.

Resonance smps has some complicated resonance where also the leaaking inductance of the output transformer is included, however the fase shift topology do different, using a coil who do saturate above some current and so lengthed the soft switch of the igbt or mosfets to en broad level, the normal resonant do put out a sinusoidal signal who go to the output transformer lowering the EMI to a very low level, but also with fase shift this happens very well making a very clean smps supply..

One smps supply is for the hybrid, put out 250 v 200 mA 2 x 65 volts 20 amps, and 14 and 8 volts 5 amps.

Enough projects, maybe even a little to much.

thanks for the links.

I did see your circlophone amp, nice, I have tryed a bias servo in mine allfet circlotron who is a real circlotron but did get trouble, maybe bacause the missing ground potential, who excist only through a 150 ohm resistor, yours is much easier to implement, and it is a very nice looking amp.

I have found something for these calculations. Maybe people here has profit from it, looks nice.

regards
 

Attachments

  • 380konvertor.zip
    177.6 KB · Views: 111
The idea was a half bridge, for both, the supply for the tubes with smaller mosfets and regulation because of filament voltages.

But here is also some interesting reading, a resonant forward converter who is also usable, maybe better for high voltages, it is some kind of flyback.

I have a question about the transformer turnsratio, N1/N2 is this also the case for the inductions? as I need these for simulation, I have now do divide primary by this turn ratio to get the right voltages, this did work, like 70 uH Primary divided by ratio 1.59 = 55uH, as this is the same for the voltages.

some reading about resonance soft switch forward.

http://www.ti.com/lit/an/slua159/slua159.pdf

and about the transformers.

Tahmid's blog: Ferrite Transformer Turns Calculation for Offline SMPS Half-Bridge Converter

Is the use of a pic a interesting idea? can these do well driving a resonance bridge.

Regards
 
Last edited:
Hi There

Thansk for the explaining of transformers.

PFC is not needed here, we have 220 volts grid and the currents are not high, it is not a high power amplifier but high end, maybe for the bass amp later, or maybe I just do it as a practice to learn, making a PFC is not complicated these days..

here is a sim, one is without regulation and one is with.

I go also try the forward resonance bridge, more complicated transformer, maybe need of a reset winding or or some kind of network diode resistor cap parallel on primary.

regards
 

Attachments

  • Unregulated voltages half bridge1.jpg
    Unregulated voltages half bridge1.jpg
    270 KB · Views: 110
  • Regulated voltages half bridge.jpg
    Regulated voltages half bridge.jpg
    210.2 KB · Views: 122
Here I did found also a nice application note.

I have just need the feedback network calculater missing, so I can calculate the optocoupler 431 reference diode parts, this is complicated by trial and error, and I think the optocoupler diode current, better is use a stable voltage, using a zener diode and regulate with the reference diode.

regards
 

Attachments

  • an-1160.zip
    562.9 KB · Views: 55
The voltage source for the feedback mustn't be regulated away from the power output itself, because this reduces the percentage of error translated to the input side. Both TL431 and photocoupler must be powered directly from the main output, or sometimes, from more then one output, to keep regulation. I saw that some authors make a resistive adder to give the led current and keep more than one output closer to the desired value.
 
The voltage source for the feedback mustn't be regulated away from the power output itself, because this reduces the percentage of error translated to the input side. Both TL431 and photocoupler must be powered directly from the main output, or sometimes, from more then one output, to keep regulation. I saw that some authors make a resistive adder to give the led current and keep more than one output closer to the desired value.

Calulating it is a issue, I do not now how to do that properly, I did see that sometimes the opto diode is supply with a constant voltage through a zener connected to a output, and a other one to the tl431 to get regulation.

I have a program who calculate the supply resonant and has such a feedback however there is a error and not outcome, see the attachment.

Did saw also this on internet,

Resonant converter design software - LLC, LCC resonant power supplies

regards
 

Attachments

  • 380konvertor.zip
    177.6 KB · Views: 72
Resonant or hard switching, the feedback through a photo is by far, easiest than the power stage itself.

Given a certain output voltage, you must make a voltage divider (Including a preset most of the times, for fine adjustment) to the reference of the TL431. Then, take a current in the led of the photo, say, 5mA (It must be higher than the minimum current permissible in the 431, Texas say 4mA or so, ST says 1mA), and then with the current plus the diode and the TL's drop, calculate a resistor to drop the difference to the desired voltage (If the supply has only one output, there is no choice), and the power lost in them. Given a photo transfer ratio, take the 5mA multiplied by this TR, then you will get in the photo's transistor, a current that adequately drive the main power stage.

Sometimes, I saw a resistor in parallel to the diode, to increase the current flowing in the TL, but sincerely, I never used nor needed it.

There are lots of ANs about this aspect, which sometimes includes part or all the frequency compensation network of the entire power supply, depending if the main PWM regulator has its own compensation pin or network or not. It is always advisable to add a .001µF from the anode to reference of the TL to reduce its noise and the tendency to oscillate in certain conditions.
 
Resonant or hard switching, the feedback through a photo is by far, easiest than the power stage itself.

Given a certain output voltage, you must make a voltage divider (Including a preset most of the times, for fine adjustment) to the reference of the TL431. Then, take a current in the led of the photo, say, 5mA (It must be higher than the minimum current permissible in the 431, Texas say 4mA or so, ST says 1mA), and then with the current plus the diode and the TL's drop, calculate a resistor to drop the difference to the desired voltage (If the supply has only one output, there is no choice), and the power lost in them. Given a photo transfer ratio, take the 5mA multiplied by this TR, then you will get in the photo's transistor, a current that adequately drive the main power stage.

Sometimes, I saw a resistor in parallel to the diode, to increase the current flowing in the TL, but sincerely, I never used nor needed it.

There are lots of ANs about this aspect, which sometimes includes part or all the frequency compensation network of the entire power supply, depending if the main PWM regulator has its own compensation pin or network or not. It is always advisable to add a .001µF from the anode to reference of the TL to reduce its noise and the tendency to oscillate in certain conditions.

I have done the feedback with esim, that was the most easy.

here some pic of the simulation.

I think now I can go build and try, otherwise I stay simulating. I have 10 NCP 1395ap chips who are resonant drivers, but maybe I do first a openloop to get use on it.

Some advise for the mosfets or maybe IGBT on the output, what kind is best, there are there a lot, In have some but are big ones for welder smps. I think 20amps to220 high speed are oke? or just mosfets, but need to watch out for body diode who needs to be fast, like trench mosfets.,

regards
 

Attachments

  • Opto-feedback.jpg
    Opto-feedback.jpg
    185.6 KB · Views: 101
  • feedback-TL431.jpg
    feedback-TL431.jpg
    149.6 KB · Views: 99
  • hbill1-coptocoupler.asc
    11.2 KB · Views: 37
Last edited:
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.