sg3525/ir2110 half bridge

yoosefheidar

New member
hi everyone
i use this schematic to build a half bridge power supply.
but i have some problems and need some helps.
first problem is feedback.i think this feedback is not good beacause whit this feedback transformer produce a noisy sound like"tik-tik-tik"about twice every second.
i read some where this feed back dosn't control duty cycle and it just turn pwm on or off but i need a feedback that control duty cycle
without feedback there is no noisy sound.
help me to change feedback.what should i do?

----
is there any problem or worng parts or circuits in this schematic?which part?and what is the correct part?
----
i change this schematic a little bit.
my primary winding turns is 19 and secondary is 8*2
i use a etd 39 core
target output voltage is 45*2
i change frequncy to 80 KHz
Schematic.jpg
 

Silvio

Well-known member
@ Yoosef

Welcome to DIY smps.

Yoosef. I see some things that are not right in what you just said.
1) Where did you conclude that the core you are using ETD39 needs 19 turns for the primary?
2) What is your input voltage?
3) Does your smps work with the feedback disconnected?

Give me some answers and I may help you.
 

yoosefheidar

New member
@ Yoosef

Welcome to DIY smps.

Yoosef. I see some things that are not right in what you just said.
1) Where did you conclude that the core you are using ETD39 needs 19 turns for the primary?
2) What is your input voltage?
3) Does your smps work with the feedback disconnected?

Give me some answers and I may help you.
hi
1-I calculate primary by "V=4NFBA" formula
2-my input voltage is 220VAC
3-my smps works fine without feedback and i connect a 300 watt load and voltage droped about 20 volt which is normal.
but now by this feedback transformer has a noise(like turning on/off) and output voltage change continuously beetwen 40-45 each rail
 

Silvio

Well-known member
hi
1-I calculate primary by "V=4NFBA" formula
2-my input voltage is 220VAC
3-my smps works fine without feedback and i connect a 300 watt load and voltage droped about 20 volt which is normal.
but now by this feedback transformer has a noise(like turning on/off) and output voltage change continuously beetwen 40-45 each rail

1) I think you made a mistake with the number of turns for the primary winding. Usually an ETD39 core takes around 28 to 30 turns for the primary winding. Are you sure it is an ETD39 not something else? DO you know the core material?

2) if you have few turns the core may saturate and heat up quickly. At 80Khz we do not usually go more than 1600 guass (0.16 tesla)

That circuit I tried it also and I experienced the same thing.

1) I suggest you put a 10K multi turn preset in line with the zener diodes to find the right value of the resistor. The balance here is not good and it is opening and shutting suddenly the pulse width

2) Put a small capacitor across like 0.47uf entering the diode for the opto coupler for start and see if it tames down a little bit.

3) You must have a small load at the output.

4) You should use an output inductor with each rail I guess 20 to 30uH each. I hope you fitted these.

Try these for now and come back with your findings
 

yoosefheidar

New member
1) I think you made a mistake with the number of turns for the primary winding. Usually an ETD39 core takes around 28 to 30 turns for the primary winding. Are you sure it is an ETD39 not something else? DO you know the core material?

2) if you have few turns the core may saturate and heat up quickly. At 80Khz we do not usually go more than 1600 guass (0.16 tesla)

That circuit I tried it also and I experienced the same thing.

1) I suggest you put a 10K multi turn preset in line with the zener diodes to find the right value of the resistor. The balance here is not good and it is opening and shutting suddenly the pulse width

2) Put a small capacitor across like 0.47uf entering the diode for the opto coupler for start and see if it tames down a little bit.

3) You must have a small load at the output.

4) You should use an output inductor with each rail I guess 20 to 30uH each. I hope you fitted these.

Try these for now and come back with your findings

1-there is no problem with core saturation or primary winding.the core works at 2000 guass and even can work at 5000 guass without saturation(pc40 material)core didn't get warm and transfer power without problem
the only problem is feedback and i should redesign it.
this feed back just work at positive output(above target voltage)and i need a feedback to control output voltage continuously by decrease/increase duty cycle(for example using TL431 but how??)
2-i have a small load in output / i use a 28uh inductor in output for each rail

just should change this crude feedback and need help for new feedback design
 

Silvio

Well-known member
1-there is no problem with core saturation or primary winding.the core works at 2000 guass and even can work at 5000 guass without saturation(pc40 material)core didn't get warm and transfer power without problem
the only problem is feedback and i should redesign it.
this feed back just work at positive output(above target voltage)and i need a feedback to control output voltage continuously by decrease/increase duty cycle(for example using TL431 but how??)
2-i have a small load in output / i use a 28uh inductor in output for each rail

just should change this crude feedback and need help for new feedback design

I am posting a pdf file for you to take an idea for the feed back circuit. This I tried and works. The schematic is from a DIYsmps member Carlos Bettecher.

View attachment Bettecher_1Kw_Smps correction.pdf

The only mistake I can see here is that R14 should be tied to pin 6 of the IR2110 instead of ground

Regards Silvio
 

yoosefheidar

New member
I am posting a pdf file for you to take an idea for the feed back circuit. This I tried and works. The schematic is from a DIYsmps member Carlos Bettecher.

View attachment 6872

The only mistake I can see here is that R14 should be tied to pin 6 of the IR2110 instead of ground

Regards Silvio
thank you very much Silvio
this pdf helps me very much and now i should calculate feedback resistors and capacitors and compensation parts based my circuit parts

about R14 i think that is correct such as R12
 

Silvio

Well-known member
thank you very much Silvio
this pdf helps me very much and now i should calculate feedback resistors and capacitors and compensation parts based my circuit parts

about R14 i think that is correct such as R12

About R14 that is the gate source dumping resistor. This is fitted to the high side gate. the source of the high side gate is at pin 6 not ground. On the other hand the low side switch source is to ground so the dumping gate resistor is to ground. I hope this explains it.

The 33v zener is fitted so that the TL431 does not see more voltage than it can handle.

Cheers and good luck
 

yoosefheidar

New member
About R14 that is the gate source dumping resistor. This is fitted to the high side gate. the source of the high side gate is at pin 6 not ground. On the other hand the low side switch source is to ground so the dumping gate resistor is to ground. I hope this explains it.

The 33v zener is fitted so that the TL431 does not see more voltage than it can handle.

Cheers and good luck

i have a question
can you explain how exactly this feed back work when output voltage is under or over target voltage?i need part by part(feedback) explanation at opto diode side and transistor side
 

Silvio

Well-known member
i have a question
can you explain how exactly this feed back work when output voltage is under or over target voltage?i need part by part(feedback) explanation at opto diode side and transistor side

Well when I first tried this feedback circuit I asked myself the same question. What I found out was that the resistor before the zener has to be tailored in such a way to the point where the zener starts to conduct when it hits the target voltage. and feeding the reference, The circuit values shown are for 50v I needed 60v the resistor on the circuit shows a 33K resistor in my case I ended up using a 47K instead and got a 60v output.

In your case I guess you have to go somewhat lower with the resistor. I suggest you insert a preset at first until you get the right resistor value than you can put a fixed one when the value is known. One other thing that I added was a 200nF cap across the zener and it helped a lot with a stable regulation. One other thing that I noticed was when giving the smps an continuous load at full power the output inductor started heating a bit and as the core changes the value due to heat the output voltage did drift a little.

This I did not try but worth considering

You can insert a voltage divider network with the output feeding an opamp input, the TL431 will be used as a reference for the opamp other input. A comparison is made with the reference and the output of the opamp will feed the opto directly. The supply rail for the opamp must be isolated from the one feeding the SG circuitry otherwise there will be no isolation on the ground point from high and low voltages

Regards Silvio
 

MicrosiM

Administrator
Staff member
@ Yoosef

10 Years ago with my first attempt to build a SMPS, I built that circuit, and it worked fine, it was my first regulated SMPS.

But I never completed a regulated SMPS, at least in the last year.

But now I am moving to the new LLC SMPS topology.

======

@ Silvio

I had my ETD39 working fine with 20 Turns at primary. what core temperature you have in idle with your ETD39? at 28 Turns in primary?

Maybe there is a reason?
 

yoosefheidar

New member
@ Yoosef

10 Years ago with my first attempt to build a SMPS, I built that circuit, and it worked fine, it was my first regulated SMPS.

But I never completed a regulated SMPS, at least in the last year.

But now I am moving to the new LLC SMPS topology.

======

@ Silvio

I had my ETD39 working fine with 20 Turns at primary. what core temperature you have in idle with your ETD39? at 28 Turns in primary?

Maybe there is a reason?
designing a good feedback is very complicated.
for a good feedback we need to calculate poles/zeros/bode plots/phase margin/gain margin/transient response and other parameters.
so i should use this present feedback.
why you move to LLC SMPS ?
i think for smps below 500 watt half bridge is the best choice.isn't that?
 

Silvio

Well-known member
@ MicrosiM

Running at 2000 guass at 80Khz is rather a bit rough I guess. According to Excellent IT for ETD 39 it gives 25 turns for PC40 Material at 80Khz Max guass calculated at 1600

My SMPS ETD39 Epcos N87

Idle temperature at 30 turns primary and 240vac input was around 40 deg celcius after 30min operation at ambient temp 25 deg no cooling fan. My switching frequency was 65Khz.

@ Yoosef
I think that feedback arrangement is worth giving a try. If you have a better idea we will be happy to see your outcome
 
Last edited:

yoosefheidar

New member
@ MicrosiM

Running at 2000 guass at 80Khz is rather a bit rough I guess. According to Excellent IT for ETD 39 it gives 25 turns for PC40 Material at 80Khz Max guass calculated at 1600

My SMPS ETD39 Epcos N87

Idle temperature at 30 turns primary and 240vac input was around 40 deg celcius after 30min operation at ambient temp 25 deg no cooling fan. My switching frequency was 65Khz.

@ Yoosef
I think that feedback arrangement is worth giving a try. If you have a better idea we will be happy to see your outcome
best working temp for a pc40 core is about 90 deg because at this temp core losses is minimum
so 60-80 deg is normal for ferrite core.
in my smps with 19 turns in transformer primary core temp is about 50-55 deg without load at 27 deg ambient temp
yeah i will try that feedback as soon as i purchase parts.
 

MicrosiM

Administrator
Staff member
best working temp for a pc40 core is about 90 deg because at this temp core losses is minimum
so 60-80 deg is normal for ferrite core.
in my smps with 19 turns in transformer primary core temp is about 50-55 deg without load at 27 deg ambient temp
yeah i will try that feedback as soon as i purchase parts.

Thats one very good reason why you get 50-55C in idle.
 

yoosefheidar

New member
@ MicrosiM

Running at 2000 guass at 80Khz is rather a bit rough I guess. According to Excellent IT for ETD 39 it gives 25 turns for PC40 Material at 80Khz Max guass calculated at 1600

My SMPS ETD39 Epcos N87

Idle temperature at 30 turns primary and 240vac input was around 40 deg celcius after 30min operation at ambient temp 25 deg no cooling fan. My switching frequency was 65Khz.

@ Yoosef
I think that feedback arrangement is worth giving a try. If you have a better idea we will be happy to see your outcome
hi Silvio
i use that feedback in my smps, but transformer sound become worse!(voltage changes become better and vary beetwen 92 to 94)
listen to transformer sound in this link:
https://vocaroo.com/i/s1mE8HRtrKuB
 
Last edited:

Silvio

Well-known member
So true such ticking sound. Did you try to put a minimum load? say 10 watts or a little more. It could be the case that the pulse width is very narrow. Take a shot across the trafo primary with the scope and see what it looks like. with and without load. Do you have a snubber across the primary or switching transistors. I also found out that the output diode will need a snubber when an output inductor is used, this causes a spike and reflects back in the primary winding

In my case I found a balance between voltage drop from the inductor and regulation at load. My output inductor was 15uH the best compromise. The output capacitor was 4000uf per rail. The output 1000w
 
Last edited:

yoosefheidar

New member
So true such ticking sound. Did you try to put a minimum load? say 10 watts or a little more. It could be the case that the pulse width is very narrow. Take a shot across the trafo primary with the scope and see what it looks like. with and without load. Do you have a snubber across the primary or switching transistors. I also found out that the output diode will need a snubber when an output inductor is used, this causes a spike and reflects back in the primary winding

In my case I found a balance between voltage drop from the inductor and regulation at load. My output inductor was 15uH the best compromise. The output capacitor was 4000uf per rail. The output 1000w
i have two 5K resistor in output.it is about 0/5 watt.i didn't try smps with a load to check trafo sound.
unfortunately i dont't have a scope and just use MATLAB for seeing wave forms.
now duty cycle is about 35% and i don't think pulse width is narrow.
i have a snubber in primary an one in secondary.but in secondary i have a resistor in series with capacitor beetwen positive and negative rail(center not included)

so you think this sound is totally normal?is there any way to remove this sound in idle?whit my previous feedback i never had this sound just two ticking per second and not so loud.
 

Silvio

Well-known member
If output capacitor charges to full regulated voltage it will definitely have a narrow pulse due to it trying to keep the output regulated voltage. 5k resistor does not impose so much load for the pwm to keep the pulse width not so narrow, Why not try a lamp and load the output a bit to see if the ticking sound stops
 

yoosefheidar

New member
If output capacitor charges to full regulated voltage it will definitely have a narrow pulse due to it trying to keep the output regulated voltage. 5k resistor does not impose so much load for the pwm to keep the pulse width not so narrow, Why not try a lamp and load the output a bit to see if the ticking sound stops

i will try this to see what is result.
why without feedback trafo don't have that sound at all?
 
Top