IR2153 HO doesn't switch

Deltor

New member
Silvio, I can't understand your concerns about the high-side buffer (bootstrap?) cap.
Bogdan uses - as I can see from the photographs - a perfectly suited multi-layer ceramic capacitor.

The only thing which is wrong is the way he wired his high-side gate-discharging resistor.

I also have to correct a statement which I gave.
It is of course necessary to use a high-voltage ultrafast diode and not a simple schottky diode.

The gate-discharge resistors should be >10kOhm.

There may be designs from stone-age, where multi-layer ceramic caps didn't exist.
So your commercial smps-thingies you've seen, may be from designers from this age.
 
I fixed the high side gate discharge resistor, replaced the diode with 1N5819 and tested again with the dummy light bulb instead of the transformer. This bulb does not light up but the current limiting one does. There is switching activity but the waveform gets severely distorted once I put in the high voltage. I think I'm giving up on this, there's too much time invested in what was supposed to be a simple circuit.
 

Silvio

Member
Well this is up to you if you are to abandon the simple circuit. All i can tell you be aware of fake chips, my experiance with this is hell when i was building my first smps. If you like to know what i have been through just go to the blog posts and download the file a 350w smps by me.
 

Deltor

New member
Bogdan, you didn't read my errata.
A 1N5819 has a maximum reverse voltage of only 40V, which is far too low.
Use the diode you have used before!
 

Deltor

New member
Norazmi, you are telling complete bull****!
Ok, you aren't able to write in English, so it's horrible to to read your writings at all!
But I don't know what you are trying to tell us about some 4148.
Bogdan's bootstrap diode will see over 300 volts reverse voltage - so I was wrong when telling him to use a simple schottky diode!
The Schottky idea was just as I've never needed to go beyond 30 or 40 volts.
But the reverse voltage in Bogdan's case is much higher.
The wrong resistor wiring was probably the only thing which was causing the error.
 

Deltor

New member
Haeh?
What's this?
Some stupid script changed some of my spelling 'S' 'H' 'I' 'T' to '*' '*' '*' '*'.
WTF ...?
 

Deltor

New member
Fancy dumb bot!
But - to return to our problem - I'd like to know how does the circuit do with its original bootstrap diode but with correctly connected discharging resistors.
:)
 
Last edited:
Fixed schematic is below. I powered the chip separately with a 12V supply and increased the input voltage. The rest of the circuit is just a bulb instead of the transformer and the input circuit without the bootstrap resistor.
First waveform is the high side mosfet without the input voltage. As soon as I increase that voltage (even a tiny bit), the waveform becomes the one in the second picture.
I will try one more thing, I have a board with SG3525 and IR2110, and I'll put it in place and see how it's performing.
 

Attachments

  • IR2153_Board.pdf
    12.8 KB · Views: 6
  • DS1Z_QuickPrint3.jpg
    DS1Z_QuickPrint3.jpg
    19.7 KB · Views: 9
  • DS1Z_QuickPrint4.jpg
    DS1Z_QuickPrint4.jpg
    19.9 KB · Views: 8

norazmi

diysmps Senior Member
Norazmi, you are telling complete bull****!
Ok, you aren't able to write in English, so it's horrible to to read your writings at all!
But I don't know what you are trying to tell us about some 4148.
Bogdan's bootstrap diode will see over 300 volts reverse voltage - so I was wrong when telling him to use a simple schottky diode!
The Schottky idea was just as I've never needed to go beyond 30 or 40 volts.
But the reverse voltage in Bogdan's case is much higher.
The wrong resistor wiring was probably the only thing which was causing the error.

wow, nice talking, show to us then your working schema here, you have?? dont make thing too complicated while thousands of people without full knowledge of switching easy to make this ir2153 chip working and never blown, you are completely have no respect to other people and you claim your english is too good?? so this is your attitude? before blame me on anything, see all my post and thread here what i have been contribute into this forum, you do?? with analyze data and spectrum? you are talking in theory but never do in real life. and please respect each other. NO ONE is PERFECT in this world no matter who are they.
 

norazmi

diysmps Senior Member
the small signal 1n4148 i mean is parallel with gate resistor from ir2153 chip at LO and HO to mosfet gate.
 

norazmi

diysmps Senior Member
Fixed schematic is below. I powered the chip separately with a 12V supply and increased the input voltage. The rest of the circuit is just a bulb instead of the transformer and the input circuit without the bootstrap resistor.
First waveform is the high side mosfet without the input voltage. As soon as I increase that voltage (even a tiny bit), the waveform becomes the one in the second picture.
I will try one more thing, I have a board with SG3525 and IR2110, and I'll put it in place and see how it's performing.

from the controller schema as i can see you have set up about 61Khz switching frequency , and i like to suggest you to remove 12k resistor from the gate LO and HO side of each mosfet gate and again parallel 1N4148 small signal diode with R2 and R2, and please measure you transformer core, Np = Vin x 10^8 / 4 x Fsw x 2000 x Section area. For example im using ER35 from computer supply, because here im using 220-240 Vac, i got 310-320 vdc average after rectified. my primary winding will be Np = 310 x 10^8 / 4 61000 x 2000 x 1.32cm2. So i got arround 43-44 winding for primary. in my case with fsw 60-70 khz my gate resistor 22 ohm parallel with 1n4148 eash side lo and ho, for the testing to see if the smps is working, enough if you are winding at primary first, the secondary you can winding later after all the primary is working good. im using irf740 mosfet. from the wave detail you put here, first picture showing that switching is working properly and second one is the failure, ir2153 has built it internal zener diode 15v, if you want to use external supply for this chip do not exceeded 15vdc or better you clamp with resistor from main supply after rectified.
 

Silvio

Member
@ norazmi
I think there is a little error regarding the transformer details. An ER35 has a center core diameter of 1.1cm^2 and not 1.3cm. Regarding the formula itself I guess the voltage at the half bridge is 160v which is half the rail voltage. The flux density will be more appropriate at 1600 guass @ 61 Khz. Please correct me if I am wrong but we do not want to mislead our friend.

My results according to Excellent IT 4000 (newer versions go crazy on windows 10)


Core ER35/20/11 N87 Total transformer power 530W Total output power 350W
Frequency Switching 61 Khz
RDS on for IRF740 mosfets 0.5 Ohms
Winding secondary single rail with a center tap winding. Wire diameter 0.5mm current density @ 6A/mm^2
Secondary voltage 35v @ 10A regulated
Output choke inductance 33 uH for minimum output current of 1 amp up to Max. For lower minimum current 0.2A choke inductance 100uH

Primary winding turns 32 turns (3 wires 0.5mm side by side or twisted) see what fits best in a single layer
Current amplitude in primary winding is 3.3A ( at the half bridge.)

Secondary winding turns 10 + 10 (6 wires 0.5mm side by side or twisted) see what fits best

If you need an auxiliary winding to supply the chip (15v) then the turns will be 5 with a single wire of 0.3 to 0.5mm which ever you have. I suggest that you make a series resistor like a 100 ohms or so to limit the current to the zener within the chip. Rectification can be with a single fast diode and a small smoothing capacitor like 47uF x 25v.

Regards Silvio
 

Silvio

Member
I just realized that our friend is using an IR2153 chip however the chip cannot be regulated. The output inductor is not really needed and maybe only a small one is necessary such as 6uH to block out any residual noise. The turns in the primary are still the same (32 turns) for the switching frequency of 61 kHz. The wire thickness also stays the same for the primary winding for 350w output at he secondary. Now we are to focus how to calculate the peak voltage at the output.

This will be calculated as follows-:

Primary voltage 155v / turns primary 32 = 4.84v per turn.

So secondary voltage peak will be approx 4.84 X number of turns secondary.

The current density is found by finding the cross sectional area (CSA) of the wire used (preferred 0.5mm for the frequency used) This will be multiplied according to the wanted current at the output.

Ex. CSA = pie x Raduis squared so 3.142 x 0.25mm x 0.25mm = 0.196mm^2 x 6A/mm^2 = 1.178A for every wire used.

So for example we need 6A current at the output then we divide 6A by the number of amps each wire can handle
which is equal to 6/1.178 which is equal to 5 wires. so our secondary wires are 5 in parallel or twisted together.

Regards Silvio
 

norazmi

diysmps Senior Member
well for the core because im measure by myself here not using excellent IT, the core shape and dimension maybe different according to the company production, remember.. its not static forever, the production sometimes change with new datasheet, so we need to always measure it. when we are using software to calculate but in real life the value is different so we can say that software will ASSUME as per data collected at that time, so there will be no issue regarding calculation because we have formula that can calculate any core, the the core type is only different is what frequency it can be running some ferrite can go up mhz some hundreds khz and etc.. so select the correct core to operate what frequency you wanted, as far as im concern, this chip ir2153 maybe not good damage or else, because of ir2153 need only 5ma to operate and yes you need to limit the aux current or your ir2153 will get hot fast without current limit with resistor and diode. because this halfbridge ir2153 there is no feedback, what i do is i will wind 1 round at secondary and i will calculate how many turn i need for the output, its not hard. if your primary is too short the pwm will have hard switch and that chip also will stress, core will heat more than usual too, too many winding at primary will cause your trafo will produce sound and not strong output, this chip has been design for fixed duty cycle 50% all the time which i think thats perfect, you can make kilowatt smps with ir2153 without any problem, 4 pair of mosfet even more with totem pole. always use light bulb for safety and save ur component from burning. i will put video how i success with this chip even i put 20,000 uf 63v each side secondary +v and -v. and for the mosfet im only use small to220 heatsink without any problem at all.
 
For my core I got 1.33cm2 so 30 primary turns for 66kHz and 1500 gauss (60 turns / 2 since it's a half bridge).
But that's irrelevant for now, since I've tested the circuit with a dummy light bulb instead of the transformer, as suggested here. Even with the SG3525-IR2110 circuit the waveform gets distorted. I'll be testing everything on a breadboard circuit with genuine transistors.
 
Top