IR2153 HO doesn't switch

Silvio

Member
im enjoying high quality class d amp now lol tl494 hahahaha, running at 212khz, with irfz44n at +52vdc and 0. this is single supply amp. high freq more efficiency means less distort :D

Tell us more about your project maybe a schematic will help us more to understand .
 
Removed choke, same issue. Removed transformer, placed a small bulb in place of the primary, it doesn't light up but the limiting bulb does :confused:
 

norazmi

diysmps Senior Member
ok now lets go deep into your smps, tell me rt ct of your circuit, tell me what core size you are using and tell me primary winding of your core also, let see what u got, we will solve this issue.
 

norazmi

diysmps Senior Member
change 2.2uf 275 v to 1uf 400v , 2.2uf is bad start for your circuit, too big in rush current, i think D3 is refer to aux supply, ok here if you have external supply 12vdc put it there or clamp with big resistor from main after rectified with 40kohm 5watt, the aux supply is overkill without current limit resistor 100 ohm 1 watt at least, 56k 2w clamp that resistor cannot stand for a long time may burn, to power up ir2153 need small current 5ma only, R2 and R3 neex 1n4148 each gate, and from the rt ct ur setting i can see you will have arround 60-62 Khz, if ur using EE35 from PSU, primary winding will be 46 CT, means 23 each side, do not wind secondary winding yet, just start with primary only and add rc snuber to primay pin 1 (start) and pin2 (end) with 100 ohm 2w and 680pf. do what i say, ill draw some circuit for you easy to understand.
 

norazmi

diysmps Senior Member
60 is ok, i make center but i dont connect the center, like psu pc supply i just connect at start and end, my circuit i give u above with short circuit, over current protection, it will never blow your mosfet.
 
Not at the moment. I won't be going back to the lab for a while. But I'm still puzzled at the number of turns in the primary. Is it half of the number of turns you get in a flyback/forward converter?
 

norazmi

diysmps Senior Member
for me i do all primary first winding, then i shield it up with tape, then i continue all secondary, flyback trafo is different it as gapped on middle core while halfbridge is not, ofc flyback has more turn than half bridge. we can parallel the mosfet with totem pole also if necessary , play with high voltage need good care, but when ur success soon you will be good to go, like i said number of turn also depends on fsw ur running, more lower fsw will be more number of turn, thats the principal. dont worry do when you have time later :D.
 

Deltor

New member
Hi!

I've just stumbled over this thread.
My first impressions weren't too bad, although I found it sketchy to feed 1.7W of power into a poor 2W resistor - assuming you feed 230V mains voltage into your circuit.

Secondly I found it strange by Silvio to recommend to replace your 1µF ceramic high-side cap by a 470nF foil capacitor.
This also applies to the 1k resistors.

Now I've seen (from image https://www.diysmps.com/forums/attachment.php?attachmentid=7221&d=1586160033) what you've done wrong:
You put some crappy 12k SMT resistor between LO and GND.
Even worse: you put a THT resistor between HO and GND!
This will draw an extreme amount of current from your high-side driver.
There may be several hundred volts between HO and GND - but this will probably not happen because the high-side capacitor will discharge quickly!
You probably wanted to put the THT (1.2k?) resistor between HO and VS, but this is also nonsense.
Remove these two resistors!
LO and HO are push-pull outputs - there is no need to put some fancy "pull-down" resistors here.
It is also not necessary to use an ultrafast high-voltage diode as a bootstrap diode - use a schottky one!
Probably I've missed something to mention...
 
Last edited:
Indeed the resistors are there to discharge the gates of the transistors, but they make no difference whatsoever. I've seen them in other circuits too.
 

Deltor

New member
Bogdan, you don't understand me.
You've soldered that THT resistor from HO to GND - not to VS, what you've probably intended.
It will quickly discharge your high-side cap soon after the high-side MOSFET turns on, and therefore the high-side MOSFET will turn off.
Remove these crappy pull-downs!
If they make - as you tell - no difference ... you have nothing to lose.

And one last question:
why do you want to use 12 kOhm for the low-side and 120 Ohm for the high-side?
 
Last edited:

Silvio

Member
Indeed the resistors are there to discharge the gates of the transistors, but they make no difference whatsoever. I've seen them in other circuits too.

I guess you did not understand our friend Deltor, If you intend to use a pull down resistor this is there so that any build up charge on the gates will be discharged. This will only help in the initial charge up of the first pulse so that the gates will be held discharged and will not pulse erratically.

Please note that if you intend to use them than they always must hook up from gate to source of each transistor. The mistake you made was that you hooked this resistor from gate to ground on the high side transistor which is WRONG. The high side source is referenced to VS and NOT ground so the pull down resistor of the High side gate must be tied from gate to VS.

I hope I made it clear for you.

Regards Silvio
 

norazmi

diysmps Senior Member
there is no need 12k hi and lo side, you cannot avoid inrush current at first pulse with ir2153 , omg... as i already mention earlier, please read ir2153 datasheet, this chip dont have softstart ( means it cannot avoid inrush current UNLESS you make step up from 12-220 etc.. this does not require it), what ever it is, if you put big caps bank at secondary and big output at secondary for example 50v above, at first start to charge that big cap or high voltage, it will ask the primary to have huge current at first pulse, no matter what, nothing can avoid this orrrrrrr you will be keep burning you fet and ir chip, to solve this issue is by putting softstart at primary with relay, let say 5-6 second before it running fully. there is no need 12k or 10k or what ever at hi and lo and snubber is not necessarily for this circuit half bridge for flyback yes its needed. i have fully experimental this chip ir2153, softstart will save your life and component and you will keep this power supply for long time with it.
 

res_smps

New member
If you intend to use a pull down resistor this is there so that any build up charge on the gates will be discharged. This will only help in the initial charge up of the first pulse so that the gates will be held discharged and will not pulse erratically.

I destroyed some mosfets, and some cheap IR2110 (maybe fake) immediately after switching on just because I didn't install a G-S resistor
 

Silvio

Member
I destroyed some mosfets, and some cheap IR2110 (maybe fake) immediately after switching on just because I didn't install a G-S resistor

I read the same thing while reading Tahmid's blog way back a few years ago when I was learning about smps. This guy used 1K resistors, which I think now its a bit too much and will waste energy going to the gates. I guess 10K or 4.7K will still be enough to discharge any residual charge in the gate given the pulse is rather clean of any stray capacitance or inductance.

One other thing occurs to me about the bootstrap capacitor. Seeing a lot of commercial smps in pro amplifiers, these I find to be rather small like 0.22uf x 2 in parallel (0.44) driving large IGBTs with quite a high input gate capacitance. I think it is also appropriate as its not only seen once but on many occasions with known firms like Harman Cardon (Crown) and also other Chinese and Italian brands. I guess this capacitor has to be of low ESR and also of good quality such as film (CBB) because it takes a lot of beating and cannot fail otherwise boom.
 
Last edited:
Top