Phase shifted pwm resonant h-bridge supply !

smps_fan

New member
Dear forum members,

I am a new forum member and I would like to say hello first at all :).

Well like my title says I would like to learn more about phase shifted pwm resonant supplies.
I have read a lot of this type of topology and what I guess is that it´s probably one of the best resonant power supplies you can build.

Do I guess right ? First you may read the text what I will write down.

My pros:
-What I expect from this topology is that it´s 100% short circuit proof because I want to build a lab power supply of it
-I read that the regulation aspects are very stable of the circuit and it´s fixed frequency that needs less filtering
-Very high efficiency often read 96 per cent because of zero voltage switching approach
-H bridge can easy deliver lot of kw power

My cons:
-I am new to smps technology even though I read a lot of smps I had less practice

Since I often haven´t see such a topology build diy and this forum here is the best I have ever seen
I wanted to ask here, if someone build such a supply and can say how easy/complicated it is to build
,if we compare with a hard switched h bridge.

I´m not sure if I should even try to build but I really would like to learn more about this topology !

Thank you for listening and if my english is too bad sorry for that.

Best regards your :)

smps_fan
 

smps_fan

New member
Found interesting tutorial guide !

I found a app note for a descrete driver version ! It´s interessting because these special ics cost about 10 euros. There is a nice tutorial about the whole thing too but I don´t know some things and have questions. I will try that with a SG3526 I think because I have it lying arround.

What is meant with transformer capacitance ?
How could I calculate/measure that when I wind a transformer ?
To calculate the primary windings can I use the standard formula for a full bridge and go then with it?

A lot of questions I know but arround these resonant topologies there are no really really good tutorials or guides that every one can easily understand all the things.

However I attached the pdf and hope that I can make a nice power supply without a boom. ;)

Best regards your
smps_fan

View attachment phase_shifted_pwm_app_note_descrete_components.pdf
 

szhighstar

New member
Below these books for reference :
a. Power Supply Cookbook ------ Marty Brown.
b. Practical Switching Power Supply Design------- Marty Brown.
c. Power Sources and Supplies ----- Marty Brown editor.
d. Switching Power Supply Design ------- Abraham I Pressmen.
e. High Frequency Switching Power Supplies-Theory and Design –George Chryssis.
f. Demystifying Switching Power Supplies ---- Raymond A. Mack,Jr.
g. Practical Design of Power Supplies ----- Ron Lenk.
h. Regulated Power Supplies ----- Irving M. Gottleb.
i. Switching Converters Medium and High-power --- Dorin O Neacsu.
j. Principles and Elements of Power electronics ---- Barry w William
k. Switch mode Power Converters Design & Analysis --- Keng C Wu.
l. Switch mode Power Supply Handbook --- Keith H. Billings.
m. Power Electronics Handbook – Industrial Electronics Series --- Edited by Timothy L. Skvarinina of Purdue University.
n. Power Mosfets –Theory and Application -Duncan A Grant and John Gower.
o. Digital Power Electronics and Applications—Fang Lin Luo, Hong Ye, Muhammad Rashid.
p. Switching Power Supplies A to Z --- Sanjaya Maniktala
 

cllow2020

New member
i'm new to this forum too.
interested on PSFB, going to try to build one and test. target output 200V@4A. adj. CVM(constant voltage mode)
attachment draft for driver only, looking for forum member to guide me on the way to build.
1) controller IC's is dsPIC33f504gs MCU.
2) gate driver using IR2110

View attachment PSFB_drive.pdf
 

cllow2020

New member
GDT isolation

very good info. thanks. if i can't make GDT to works, than i would using Photocouplers.

i've IR2110 on-hand and some EE core pickup from scrap.
EE core.JPG

GDT setup:
EE13, Ae=3.1mm x 5.6mm
VDC= 15v (pri-voltage)
Dmax= 0.9 (maximum duty cycle)
Pin= 1.5A*15v =22.5w (power input)
fs= 100khz
Krf=1

==>> L=(VDC*Dmax)^2 / 2*Pin*fs*Krf =40.5uH


Np=pri-turns
L=40.5uH
I-over=4.2A
Bsat=0.35T
Ae=3.1mm x 5.6mm

==>> Np=( (L*I-over)/(Bsat*Ae) )* 10^6 = 27.5 turns

GDT ratio 1:1
winding will be:
pri/2 +Ns +Ns +pri/2
14 + 28 + 28 +14
wire size =0.35mm

if i do anything wrong please guide me . thankyou.


regards
 

cllow2020

New member
have been thinking drive a GDT using four discrete transistors, but MCU output is 3v3 only, driving current about 20mA is there a better way to not using IR2110:D ? cost saving
 
Last edited:

Jagd.Panther

New member
have been thinking drive a GDT using four discrete transistors, but MCU output is 3v3 only, driving current about 20mA is there a better way to not using IR2110:D ? cost saving

You can use low-side drivers, like IXD 604, MC34152, etc (they are cheaper than IR2110).

OR you can build a discrete driver see attached picture or google for "IR current buffer driver" appnote for another example.
 

Attachments

  • discrete-driver.png
    discrete-driver.png
    4.6 KB · Views: 77

Jagd.Panther

New member
Winding fill factor looks to high, are you sure you will be able to stuff that amount of wire? Also B looks high. Be ready to switch to bigger core.

WRT discrete drive, despite it looks OK it won't work in real life, you have to use decoupling caps (like 0.1u-1u low impedance ones, SMD ceramic will do) for each totem pole driver (or at least for each pair you layout it w/o major loops).

Also make a provision for snubbers on the secondary side and ideally a provision to install back-to-back zeners for gate overvoltage protection.
 

cllow2020

New member
thank you for highlighting.
wire size 1.12mm ,winding are pri/2:22:22:pri/2 , core using ==> EE42:42/20 (PC40,TDK, no gaps)
bobbin length 26mm, high at 8mm, i think program miss-leading using EI cores

EE42 code 42_20.JPG

snubber is not necessary but good to have as "option" when needed gate pulse smoothing, when dealing with critical timing(dead-time & ZVS timing).
GDT sec coil peak voltage some-way at 18V, clamp below 20v is not needed, as gate resisters 820 ohm will do the job. GDT pri coil power reduced using 0.47uf cap.

View attachment PSFB_discrete_drive_3.pdf

anyone have inductor leakage fomula for ZVS ?
(refer to L1 schematic)
 
Last edited:

res_smps

Member
thank you for highlighting.
wire size 1.12mm ,winding are pri/2:22:22:pri/2 , core using ==> EE42:42/20 (PC40,TDK, no gaps)
bobbin length 26mm, high at 8mm, i think program miss-leading using EI cores

View attachment 5668

snubber is not necessary but good to have as "option" when needed gate pulse smoothing, when dealing with critical timing(dead-time & ZVS timing).
GDT sec coil peak voltage some-way at 18V, clamp below 20v is not needed, as gate resisters 820 ohm will do the job. GDT pri coil power reduced using 0.47uf cap.

View attachment 5669

anyone have inductor leakage fomula for ZVS ?
(refer to L1 schematic)

leakage inductance depend on the space between primary and secondary winding with all secondary shorted
 

Jagd.Panther

New member
GDT totem-pole drivers still lack decoupling caps.

The snubber you put across the primary is of a wrong kind, just use plain RC snubber (you don't need to solder its component once the PCB is ready).

Whether or not 820R would be enough to damp voltage spikes on GDT's secondaries depends on leakage inductance/current passing through it (di/dt).

Peak turn-on current for mosfets would be ca. 1 if BJT's beta is around 100. Turn off would be much higher (depends on ESR of GDT and BJT's beta droop).
 

cllow2020

New member
decoupling caps and current limit GDT driver.JPG
decoupling caps ? is C23, if i'm wrong please show me. thanks.
GDT pri. di/dt impulse current could be large, R23 10ohm added to limit it, GDT secondaries spikes reduced.
 
Top