IR2153 Soft Start


New member

I have a problem with my diy smps. Sometimes a transistor explodes when turned ON. Unfortunately, it also takes circuit 2153 with it. Apparently missing soft start circuit. Runs absolutely reliably when it survives the start.
I need to implement a soft start circuit. I found two similar schematics. Version 1 - via LED to pin1 (Red line in schematic). Version 2 - to pin3 (Blue line in schematic) Please advise me. Which is better?

Thank You.



Active member
Seeing this schematic I see that the green part of the circuit is not really a soft start but rather some kind of protection. Usually this circuit will trigger an SCR rather than a transistor. To pull pin 3 to ground is either done instantly and latches rater than going on and off with a transistor. The reason is that with a transistor the frequency will go low and then the transformer will saturate and blow the transistors.
However pulling down pin 1 to ground via the LED would be a better choice.

Another issue with this circuit is there is those two diodes coupled with a capacitor on pin 6. This part is to enhance some more current after the smps starts oscillation. You also need to monitor the voltage on pin 1 as if this voltage exceed 15v it will destroy the internal zener within the chip. Because of this the coupling capacitor from pin 6 must be carefully selected as if it is larger than necessary it will produce more current than needed and destroys the zener diode within the chip. I cannot see any values on the circuit so its difficult to calculate what is really going on.


Staff member
IR2153 Can be a painful design if not designed properly.

Look at Silvio SMPS with IR2153, its a well-designed one



New member
That's not my idea. I found two almost identical designs.

First by Silvio:

Second from web:

There is an interesting comment

Quote Salvu:
I experimented a bit and trimmed some resistance values. I'm done with the current limiter and it works very well.
The benefits I found where: -
1) Limits current when charging secondary covers. (IR2153 does not have a soft start)
2) You can run smps at almost full load (600 W) The maximum power I got was 700 watts
3) No need to reset in case of overload.
4) If the overload is continuous, the current drops to half maximum, in my case about 4.5 amps.

5) The limiter can be set so as not to overload the amplifier it controls (from 1.5 to 10 amps)

That's why I was intrigued. It looks very believable.
Can I do it like this?


  • sft1.png
    112.4 KB · Views: 46
  • sft2.png
    357.4 KB · Views: 46
Last edited:


Active member
It is because they where both drawn by me, On the eda board I was just trying to get an additional idea to get also a short circuit protection, I discarded the current limiter because during current limit if held for a long time my mosfets where also blowing. After that I modified the design and went for short circuit protection. However the soft start is included in the input and here instead of using a relay I used a triac to bridge the soft start resistor. The soft start resistor can be a 47R 5w type which gives a very good soft start.
If you want I can send you the Version 2 pdf of this power supply. You can also see the load test on youtube on this link


Active member
Here is the PDF file version 2. In the file you can find several protection circuits as these where developed during the making of this SMPS. The main pcb was adopted so that the main components are there and this could be modified accordingly. The oscillator and protection however was made on a separate pin header which is plugged in and this could easily be changed according to what one prefer to use. On the last drawing of the main pcb there is a slightly different main pcb which another 2 output capacitors are added, this time having a total of 6 on the output. Another change was with the output diodes which this time only use 2 double diodes being one with a common anode and another with a common cathode. However at times the mur2020CTR is not readily available but one could still make the main pcb using 3 double diodes instead and the smps still works. The 3 choices regarding the pin header are all illustrated in the pdf. However the last one with High frequency soft start and short circuit protection was used.
The transformer details can be found on my Youtube channel which is very detailed. The video is called winding small transformers for smps.


  • Audio smps 700w V2 (2).pdf
    1.3 MB · Views: 38


Active member
What is your opinion on the attachment?
Well the attachment is related to snubbers. This will depend on 1) the leakage inductance of the trafo, 2) if the secondary winding has an inductor after the diode. 3) pcb design. These are the main causes of spike generation and also the need for a snubber.
An smps running with a 50% pwm duty cycle does not generate much spikes and usually a light snubber will be enough to suppress them. However at times the snubber may not be needed at all. At times snubbers need to be more intensive to suppress high voltage spikes.
To sum it up it all depends on the design and need to that particular case.
Regarding the attachment we see 1) No snubber 2) a snubber across the main winding and 3) we see a further 2 snubbers added across each mosfet.
As I said you cannot tell which is best as every particular case is different from one another. One should choose the one that works best for that particular smps. Please note that snubbers waste power so these should be carefully selected and use the one that is just enough.