IR2153 HO doesn't switch

I'm building an IR2153 SMPS and I have a problem with the controller. HO doesn't switch, it stays high and blows the transistors. LO switches fine. I measured it both ways just to be sure:
- with HO referenced to GND I get HO high and no switching
- with HO referenced to VS (with an isolated scope, no diff probe, single channel) I get nothing
I also found the bootstrap diode was short circuited after the first power-up and replaced it. The weird thing is that I tested a handful of chips and they're all the same, so either they're fake or I'm doing something wrong.
 

sbdada09

New member
Hi, I faced the same problem earlier. But that was my fault. I forgot to connect the VS pin (pin#6) to the center point of the two FET's. The bootstrap diode and capacitor was connected to pin 6, but not to the MOSFET. It was floating, and this problem occurred.
I don't know about your circuit, I'm just sharing my experience.
Regards,
Sukdeb.
 
Last edited:

Silvio

Member
I'm building an IR2153 SMPS and I have a problem with the controller. HO doesn't switch, it stays high and blows the transistors. LO switches fine. I measured it both ways just to be sure:
- with HO referenced to GND I get HO high and no switching
- with HO referenced to VS (with an isolated scope, no diff probe, single channel) I get nothing
I also found the bootstrap diode was short circuited after the first power-up and replaced it. The weird thing is that I tested a handful of chips and they're all the same, so either they're fake or I'm doing something wrong.

Keep in mind that the IR2153 has two totem poles with there output being the low side at pin 5 which should be referenced to ground and the high side being at pin 7. The high side is referenced at pin 6 which is the VS pin.

If you want to try it on a bread board before connecting it to the circuit itself this is how you do it.

1) connect a 1K resistor to the output pin 5 and the other end to ground (pin 4) The supply for this output is connected internally to pin 1 which is the supply pin of the IC

2) Connect another 1K resistor to pin 7 which is the high side output pin. The supply input for this pin is pin 8 so we put the diode from the supply pin 1 and pin 8. Put it the bootstrap capacitor between pin6 and pin 8.
The VS pin 6 is referenced to nothing so The other end of the resistor coming from pin 7 should be tied to pin 6 . You must also ground this pin 6 via another 1K resistor otherwise you will see no output from the IC on the high side. (This you do only to test the IC outside the circuit. I am assuming you are to use 12vdc to check out the IC)

3 Connect the timing resistor and capacitor to pin 2 and pin 3 in the normal way. Put a small bulb in series with the positive supply from the power supply to limit the current in the event you hooked up something wrong.

Connect the scope probe firstly to pin 5 and ground and after from pin 7 and ground. You should see a nice square wave coming out of both channels. If the high side does not show and signs of oscillation it means that the chip is bad. Most probably the totem pole fets on the high side are shorted.

Lastly IF you find that the chip is good then search in your smps what you done wrong. Please note that when the IC is in the smps a bootstrap capacitor of no more than 1uf is inserted between pin 6 and pin 8. The VS pin 6 should go between the mosfets. During oscillation the VS pin is grounded via the low side fet when it switches on. This will charge the bootstrap capacitor. The capacitor will discharge when the high side signal is on and hence charges the gate of the high side fet.

I hope this explanation is clear enough for you.
 
Last edited:
Thank you Silvio, that worked very well. Only one of the chips was actually bad, the rest are good. So now I'll check the rest of the SMPS.
This would actually be a nice test circuit to have.
 
For some reason the mosfets blow up. So either there's a problem with the turnoff or a problem with the mosfets themselves. I'll probe them to see what's happening.
 

norazmi

diysmps Senior Member
you should use bulp lamp 40 watts for testing ... not directly to ac.. that will prevent your mosfet blow up...
 
Ok I added a bulb and I managed to see the high side waveform which looks more like a half rectified sine wave. The high side mosfet warms up and the light bulb flickers.
 

Silvio

Member

This is a very basic circuit that I see. What is not appropriate is that the gate resistors should not be less than 22 ohms I guess 27 ohms will be better not to load unnecessarily the IR2153. You should also insert a couple of Gate to source resistors with a value of 10K ohms. This will help with switching off of the mosfets between switching cycles. Make sure that your traces leading to the gates are as short as possible.

I am not sure about this but it seems that you have an auxiliary winding from the main trafo that feeds pin one of the IR2153. This voltage is very critical and should be no more than 15v, however if this voltage is higher than a resistor according to the voltage present from the transformer should be inserted in series otherwise it would burn the internal zener within the IC

One last note. Measure carefully with the scope to see the waveform at the gates. The low side is measured between gate and ground. The high side is measured between gate and VS. However this is rather critical because if your scope or smps is not isolated (isolating transformer) then there is a chance you blow the scope or smps. Do not measure the high side if you have no means of galvanic isolation from earth ground.

Is it possible to post a couple of pictures of your setup?
 
I've added 33ohm gate resistors and 10k discharge resistors. Now the high side looks better but I still get a large current flowing. I've also added a 12V zener diode at VCC of the controller, just in case.
The SMPS is powered from an isolated autotransformer, and the scope is not earthed. The measurement in the video is for the high side. I don't have a differential probe and I can't measure the low side at the same time, so I assume the switching is complementary (at least that's how it was with the test circuit).

https://youtu.be/rRqB7iMdFTI
 

Attachments

  • IMG_20200406_084928.jpg
    IMG_20200406_084928.jpg
    1.2 MB · Views: 14
  • IMG_20200406_084912.jpg
    IMG_20200406_084912.jpg
    1.3 MB · Views: 14
  • IMG_20200406_080126.jpg
    IMG_20200406_080126.jpg
    717.2 KB · Views: 13

Silvio

Member
@Bogdan 2011,

With the video that I saw and also the changes you have made it seems that there could be something wrong somewhere. It also could be the case that there may be a shorted turn in the transformer.
The IR2153 usually present a very nice square wave and the spike in the rise time may be also a reflection of the output inductor.
For an unregulated supply the output inductor is not really needed and only maybe a small one of say 3 to 6uH is needed here.

Please note that the duty cycle is near 50% in the IR2153 and hence the wave is not chopped as in a regulated smps. This is why you do not need the output inductor. An output inductor will induce losses and voltage drop at the output. It will also generate spikes across the primary winding. A small one such as 3 to 6uH will help block out switching noise at the output. Try also fitting a 1uf capacitor at the output it will help very much for a clean output.

Some suggestions:-
1) If you have one try to hook up a used transformer from a computer power supply (the frequency for these is usually around 20 to 30KHz)

2) If you do not have one disconnect the transformer from the pcb and hook a a small lamp of say 15w-25W instead at the primary of the trafo terminals. You can check again the wave form. If the problem still persists then it could be that the chip is not good. I do not know when you tried the pin header on the breadboard how it performed and how the waveform looked. If on the other hand the waveform is nice then it could be the transformer at fault.

3) You may also need a snubber across the primary of the trafo. This will help eliminate any spikes generated between each cycle. Usually a 470pF X 1kv and say 40 to 100 ohm resistor in series will suffice. The resistor for the snubber should be at least 2watts.

4) Try to disconnect the double diode at the output and check again. It could be the secondary winding orientation may be in the wrong direction.

5) You are doing the right thing measuring the high and low side ONE AT A TIME AND NEVER TOGETHER.

I hope this helps to solve your problem.
 
Last edited:
These are the waveforms of the test circuit. For some reason, the high output is lower in amplitude. I've measured it both referenced to ground and to VS. The LO amplitude is around 11V and the HO around 6V, and that might not be enough to turn the high side transistor properly.

IMG_20200407_115139.jpg
 

norazmi

diysmps Senior Member
im sorry late reply wait i see your circuit, this smps ir2153 is very simple and easy to make, but the condition for test run must use light bulb, then i would like to notice you this cjip does NOT include in rush current which when you make secondary more than 24vdc it will spike at high side primary winding which will create huge spike high in rush current from the main supply which this chip ir2153 cannot handle it, so what you need is to create relay plus resistor to prevent this, we call it SOFT START. its recommended soft start for high voltage and high current at secondary when you make output 30vdc and above. or you put big caps bank at secondary also will cause this in rush current at primary. to protect this you need to make SOFT START from main supply, easy to make, for me i use 470 ohm resistor for 10 second then releay release it, i use from main to make soft start, 470nf + 10 ohm, then bridge rectified go through 13v zener then i have simple circuit for simple soft start. LET ME SEE YOUR CIRCUIT.
 

norazmi

diysmps Senior Member
first, on control board put zener diode, gate resistor must be 22ohm and above for irf740, zener need each gate hi and lo, and bootstrap caps change from 1u to 100nf, change rt and ct to 10k and 1nf, AVOID big CT 10nf it will increase deadtime, you should lower ct and and calculate with RT is better. like i said above, this chip dont have soft start for high current, so use soft start for it, i never blown my ir2153 so far, and none broken yet even i push it to the limit, just voltage drop, its because of small trafo cannot produce alot power, also mosfet. ir2153 consider easy to make and simple , with step i tell you so i hope you ccan make it. i have many ir2153 circuit include protection at start. irf740 gate must be 22 and above.
 

norazmi

diysmps Senior 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
 

Silvio

Member
These are the waveforms of the test circuit. For some reason, the high output is lower in amplitude. I've measured it both referenced to ground and to VS. The LO amplitude is around 11V and the HO around 6V, and that might not be enough to turn the high side transistor properly.

View attachment 7225

That is quite normal and the reason is that the high side is charged from the capacitor and not from direct dc. In circuit it would behave differently and do not worry the voltage at the high side gate will be the rail voltage (12v) less the diode voltage drop around 0.6v. One other detail to drive an IRF740 a 470nF film capacitor is a better choice. This is more than enough for it. I seen a lot of commercial amplifiers using smps and even when driving large IGBTs and mosfets the bootstrap capacitor is not more than this value. What I usually notice is that they put a film type rather than an electrolytic capacitor. The reason for this is that here this capacitor takes a lot of beating and must be of good quality and low esr.

One last thing I want to mention. Put a 100w bulb in the input for current limiter. Do not be fooled by the input current read on the DMM. This will be seen larger than normal due to that at idle the smps has a power factor of 0.5 and the current drawn is due to this rather than that of what really the smps is consuming.

The flickering of the input current limiter may be the case that it is lowering a lot your input voltage. (this is why I suggested a 100w lamp) This in turn the auxiliary winding feeding the smps may not have enough voltage to sustain the expected 12v to supply the rail voltage. The start up resistor however will be facing the same problem and hence the high side gate may not have enough voltage to be driven as it should.
 
Last edited:
The flickering of the input current limiter may be the case that it is lowering a lot your input voltage.
This is true, my test bench variac can't hold the current drawn by the power supply, so the voltage drops quite significantly. But then without the bulb I have burned 2 pairs of transistors and a few fuses (hence the litz wire in place of the fuse holder). In the video, I was gradually increasing the voltage until the chip started switching, then the bulb starts to light up, so the current draw must be high. The reading on the DMM was the output voltage.
 
Top