IR2153 SMPS high frequency noise issue on touchscreen device.

sbdada09

New member
Hello everyone. I've just build the Audio smps 700w (IR2153) posted by Mr. Silvio.
The original File posted by Mr. Silvio is Here.

I've used smd version of the chip instead of DIP version. Also, i've used 4 U1520 diode instead of 3 RURP820C as the output rectifier. In 12V section, I've done it with TIP122+TIP127 and zener combo, while the original uses 7812 and 7912 So I've modified the PCB according to it. Except these, everything including parts value and pcb is exactly same as the original.

My output is +-30V and the overload protection is set to 300W. Everything works fine with some nichrome heating element. But when I attached the SMPS to an amplifier, it creates a 100hz humming noise. Not much loud, but it was there. After doing some test with oscilloscope, I found that there is a visible 100hz noise on the negative rail, while positive rail is ok.

But the main problem started, when I've connected a mobile phone via an audio jack. My phone screen completely gone mad. The touch screen does not work properly, random touches occurs automatically. It seems, there are some high frequency noise presents on the power rail.
How can I solve this problem? What did I missed? Thanks in advance.
Here is my Board-
 

MicrosiM

Administrator
Staff member
I dont know the exact reason for the 100Hz noise, but I am sure that the noise is related to your PCB, or maybe to your transformer.
 

Silvio

Well-known member
I made this smps and published it here. From what I can see your input filter is poorly made. I cannot see the X capacitors to make up the filter I cannot see also the Y capacitors. Without these the input filter is not complete and does not function as it should. I guess you also omitted the safety capacitors from the negative high voltage rail to earth ground. With all these parts missing noise will surely find its way to the output. The output inductors need not be more than 3uH for the reason that these will create voltage drop in your output especially with the output voltage rails being only 60v. Your current draw at 500w will be around 8A or so. At this current the output voltage will suffer with higher inductance than I suggested.

Keep in mind that the overload limiter will cater also for the charging of the secondary capacitors in fact you will notice the current led flash on start up as this would be limiting the current while charging the secondary capacitors. The output inductors are really there to filter out any residual noise.

If you want to filter out more the main output you can insert or solder them from underneath on the copper side a couple of 1uF film capacitors rated for the output voltage in your case 50v caps will be be adequate. You solder these between + and center ground and - and center ground. Insert also 0.1uf ceramic capacitors on the output rails. The selection of various values will help to filter noise at various frequencies.

Here is a video showing the smps tests for noise. The coupling is AC on the scope and the the setting is 200mV per division. with no load the noise is hardly noticeable.

link https://www.youtube.com/watch?v=KKjBy1pqhMM


I hope that helps

Regards Silvio
 

sbdada09

New member
Hi everyone, Sorry for late reply. The problem is solved after attaching the Y capacitor between high side and low side ground. Actually this one is my second board. I missed the Y capacitors and the ground rail while I was recreating the PCB layout. I noticed it while I was making my first board. It was a low quality general purpose CCB and got some damaged during soldering. But I tested the board and these noise problems was not there. As the board got damage, I made the second board with a good quality fiber CCB. So, now I have two working boards. Surprisingly I forgot the capacitor issue again and this time I faced the noise issue. Adding those now solved my problem.
As the noise was not present in the first board (even there was no capacitor) and shows up in the second, I'm guessing that it came from the transformer. Both transformer is winded by hand and I'm not an expert on that, so it can happens.

Anyway, I've another question to Mr. Silvio (and others also). In my locality the main voltage fluctuate too much. Sometimes it drops to 180 and sometimes raises up to 250 VAC. Most of the time it stays 225, but the fluctuation happens almost everyday. So it will be good if the board has some kind of output regulation. The regulation does not needs to be supper tight, +-1 or 2 volts is acceptable. All I need is that the output should not be more than 32 volts.
I've tried to regulate the voltage by pulling down the IC supply with TL431 and optocoupler, just like the conventional PWM controller. You can see them on my board. The output winding was such that it can cover all the input range from 180 to 250. The result was terrible, transformer made loud audible noise. Adding compensation network to TL431 just changed the pitch of the sound. Then I reduced the voltage to 34/35 volts at 225 VAC input and changed the feedback with a simple zener+opto network. This time the noise was low enough with no or light load and no noise with higher load. But another problem appears. With no or light load, the voltage on the +-12 Volt rail and the auxiliary rail of the IC was too low, around 2/3 volts. So, I'm currently running it without any regulation.
Is there anyway to regulate the voltage with this chip?
Thank you all again :) .
 

Silvio

Well-known member
The IR2153 is not meant to be regulated. All you can do is that you add some more voltage to the secondary and use a buck convertor to regulate your output.
If you want a regulated supply then another type of chip must be used that can regulate pulse width such as the SG3525. However the circuit will be more complex and you have to add also either a pulse transformer or a mosfet driver chip to it. The SG alone cannot drive mosfets in half bridge mode and some kind of isolation is needed between the low and high side fets. You have to change the pin header board completely. You have to make arrangements for feedback and also you must use a larger output inductors. The power supply will become more noisy as while chopping the pulse width it tends to generate more noise. As it is it is the PSU is working at around 48% duty with only the dead time minimising the maximum duty of 50%.

Your 3rd option is to add a PFC circuit to the input and this will regulate the input voltage instead of the output. However you will need to make a second board which carries the PFC circuitry and the inductor.

I hope this gives you a better view of the situation.

Regard, Silvio
 

sbdada09

New member
Thank you Mr. Silvio, I got your point. My board is now complete and running well. I've again attached the TL431 feedback circuitry, but this time not to regulate the output, but to shut down the PSU during over voltage. I'm using the "schmitt trigger" circuit you showed in the document. Now in both situation (over load or over voltage) the PSU should shut down, and this thing is enough for my purpose.
And yes, you were right about the inductor, it was more than 3uH. At about 100W load, the voltage drop on each rail was more than 5 volts. I changed the inductor, don't know the value though. Now only 2 volts drops on 300W load.
Thank you again for this awesome project :) .
 

opacheco

New member
Thank you Mr. Silvio, I got your point. My board is now complete and running well. I've again attached the TL431 feedback circuitry, but this time not to regulate the output, but to shut down the PSU during over voltage. I'm using the "schmitt trigger" circuit you showed in the document. Now in both situation (over load or over voltage) the PSU should shut down, and this thing is enough for my purpose.
And yes, you were right about the inductor, it was more than 3uH. At about 100W load, the voltage drop on each rail was more than 5 volts. I changed the inductor, don't know the value though. Now only 2 volts drops on 300W load.
Thank you again for this awesome project :) .

sbdada09,

I would like to see your design if you could share we us; please could you upload a schematic for?

opacheco
 

sbdada09

New member

Silvio

Well-known member
@sbdada09

I am sorry to say that I find the spacing between the mains and also the high voltage DC is not adequate. The spacing for the mains traces should be 3mm minimum and for the 320v traces this should be a minimum of 4mm. Take a look at my PCB and notice. This is done for your own safety DC voltage will easily gap with some dust and humidity. I also see that you again omitted the Y capacitors in the input filter (4.7nF x1000v) these are important to complete the filter. You also omitted the safety capacitor between Primary and secondary windings. In my schematic you will notice that the high voltage ground is grounded via 2.2nf to chassis at 2 different points. The secondary center grounds are also grounded to chassis via the earthing point and the stand off nuts that tie the pcb with the metal chassis of the amplifier box. We sometimes not ground directly but use a low value resistor such as 10 ohms or so to lift the ground point a little and this will eliminate any ground loops especially when using long leads to microphone circuits.

I also see that you did not leave enough space between the high voltage traces and the secondary voltages traces.

I see also a second common mode choke which in my opinion is not needed. If you are still using the IR2153 this works at nearly 50% duty and minimum noise is generated. The smps is not regulated and hence an output choke is not really needed. The 5uH inductors will be enough. I would rather include a 0.1uf ceramic capacitors between the secondary voltage rails to center ground and also will include space for 1uf film capacitors in parallel with them. This will smooth out any noise generated.

As a rule of thumb keep high current traces away from small signal traces. Avoid passing traces under the transformer. The transformer generates a lot of magnetic fields and try to stay away from it as much as possible. Keep your paths as short and thick as possible and do not run with them all over the PCB this is not good practice they serve as antennas to noise and pick it up.

I hope this will help you out with a better pcb design

Regards Silvio
 
Last edited:

sbdada09

New member
Hi Mr. Silvio, this is the board I posted earlier, and there I mentioned about the capacitors. Currently they are soldered underneath the board. I know it is risky, but don't worry, it's just for test purpose. I'm not going to use this psu regularly.
The second common mode choke is not used. I kept that space because I thought it might help. They are now shorted with jumpers.
And all about the traces? I know it's a complete disaster. I'm still learning, hope you will understand :)
 
Top