IR2153 350 watt RMS Audio SMPS with current limiter --based on Borysgo2 design--

Silvio

Member
Hi everyone, here I am presenting a tutorial and explaining my hard work in building a successful working SMPS. I being a newbie myself in this field of work had to learn things the hard way and the way to it all is explained in the PDF attached. All is found in it so one can replicate it. I hope the information included will be of help to new comers.




Introduction


Dear all, I am a newbie in smps and here I am going to explain the hard journey in building a successful smps based on Borysgo2 design. I am a hobbyist and I am also a licensed radio armature. I have a basic knowledge in electronics and I am not an expert in anyway in this field. Building these types of power supplies always fascinated my taught and I tried for the first time to build one from scratch. First of all I made some research and read a lot in the forums on DIYSMPS.COM which was a great help to me in achieving a lot of information. I have been looking a lot in the forums and what other people share on this site. I also asked questions to the gurus here, and some of them where very helpful especially MicrosiM and also Wally. Success took a bit of a long way and burning about 26 mosfets and also quite a few chips, but I will explain why later on. I had no oscilloscope and could not see what was actually happening. I must stress that this is a must here otherwise disappointment will ensue for sure. Care in every way has to be taken here especially dealing with mains. HIGH voltages are present ranging in the 300 volt region and are LEATHAL and can easily KILL YOU. Some form of isolating transformer will give you good protection although care should be taken here as well as SMPS do not joke when they blow and at times they are dangerous not to mention spectacular. The forum page presented by diysmps.com regarding Necessary Equipment's needed for SMPS development. The information here is very helpful and easily understood. So please follow these rules.

IR2153 Chip


The Circuit is based around the IR2153 chip which is a mosfet half bridge driver. It also has a built in oscillator which can be set with a capacitor and resistor to relatively a wide range of frequencies. Unfortunately it does not have a soft start and this has to be catered for otherwise problems will occur when charging the main secondary caps or perhaps started on load. The internal mosfet drivers can sink a current up to around 150-200ma and can drive medium power fets, however driving large fets in the 20 amp range the chip needs to be enhanced with a current amplifier on the HI and LO outputs.

Circuit design requirements for audio

The requirements needed are as follows:

1) Main secondary output for powering an amplifier with single or double voltage to suit.
2) An auxiliary winding for powering a preamp stage with a double output voltage of around 0.5 to 1 amp.
3) Another auxiliary winding for powering the power supply itself of around 500ma and an optional auxiliary for powering a cooling fan. This can be omitted if not needed.
4) A soft start to limit inrush current on switch on.
5) Some kind of delay to wait for the bulk caps to charge before oscillation of the IC commences.
6) A current limiter, or better with a short circuit protection.
Circuit design requirements for audio
The requirements needed are as follows:
1) Main secondary output for powering an amplifier with single or double voltage to suit.
2) An auxiliary winding for powering a preamp stage with a double output voltage of around 0.5 to 1 amp.
3) Another auxiliary winding for powering the power supply itself of around 500ma and an optional auxiliary for powering a cooling fan. This can be omitted if not needed.
4) A soft start to limit inrush current on switch on.
5) Some kind of delay to wait for the bulk caps to charge before oscillation of the IC commences.
6) A current limiter, or better with a short circuit protection.

Borisgo 2 design


Well the author of this circuit included a constant current source feeding the input of the chip at start up this makes things stable and eliminates the need of the large 27K resistor usually used. He did it in a very clever way such that firstly the constant current source will charge up slowly C3 giving time for the bulk caps to charge up as the threshold voltage is reached (9 volts) the chip starts oscillation and the mosfets start giving power to the transformer. A single turn winding around the trafo energises enough voltage to the gate of a triac TR6 which is now shorting the 15 ohm thermistor (this limits inrush current to bulk caps) thus reducing voltage drop and un-necessary heat. This way a common resistor can be placed instead of the thermistor as this is only active for 1 or 2 seconds. A current limiter is also included in a clever way as it senses the ripple voltage on the dc blocking capacitor C11 feeding the mid point on the mosfets. This is coupled by C10 and R10 amplified by a small transistor TR3 which in turn pulls the supply feeding the chip via an LED hence switching it off and re-start when the excessive load is removed. This also takes care of the fets when starting the supply on load and charging also the secondary capacitor bank.

Re-design


The PCB was re-designed by myself as I had to modify it to suit my needs. I do not know how to use eagle software or any other form of pcb design tool but I know how to use Microsoft publisher quite well and made my PCB with it, well it takes a lot of time to do as every line, circle, oval etc had to be placed in the right position and taking care of the measurement of every component. It is rather a lot of hard work and time involved but where there is a will there is a way.



Modification from original


Well modification had to take place regarding the Trafo footprint, capacitve coupling instead of the current transformer for overload protection and also the diode bridge as I used 3 double common anode fast switching diodes. I bought 50 of these so I got plenty of them. I modified the PCB to take these instead of 4 single diodes. I will show later how they are hooked up. I also had to change the size of the resistor footprint as these where of 1/8 watt instead of 1/4 watt. Which I have a good store of them in my workshop. 1/8 watt resistors are hard to find locally as well

<--- The rest of the information's in in the PDF file below---->

Please comment and share your work if you benefit from this hard work

View attachment SMPS 350 watts RMS.pdf
 

Attachments

  • IMG_0212.JPG
    IMG_0212.JPG
    1.8 MB · Views: 700
  • SMPS_Schematics.JPG
    SMPS_Schematics.JPG
    105.7 KB · Views: 1,301
Last edited by a moderator:

CobraNet

Member
Hi Silvo

This is very nice project, I would like to build this circuit. i didn't see that you made test for short circuit?

Or i missed it?
 

Silvio

Member
Hi CobraNet, I did not make a short circuit protection in this circuit but a current limiter, The reason is that it will protect both the smps and also the amplifier. However I made a little modification in the resistor value (R10) and changed it from 8K to 4.7K this way I got more current to the base of TR3 and works better. I was having some problems with a complete short circuit though. I also was using IRF840 fets in my test as I burnt all the IRF740 till I got the values correct. However the response is better now and it could take a direct short circuit. I also changed the output transistors to IRF740 and the smps can peak 450watts easily and starts to limit current beyond that. I am also posting some wave forms I got with no load and full load (450w). I advice you download the article and you will find the whole story of this first build of mine.

Silvio.

IMG_0443[1].JPGIMG_0441[1].JPG
 

lucas-nld

New member
Very nicely build Silvio, although did you switch over to a proper PCB designing program by now? Because there are many free and very good PCB designers, some have a moderate learning curve, such as Eagle and KiCad (though they are also a lot nice and have more functions), and although Eagle is limited to a small board size for the free version, KiCad is not. If you want a program is has practically no learning curve and which is free for all board sizes and amount of board layers, go with ExpressPCB. Although I have all 3 programs, I found ExpressPCB so easy to use that I've stuck to that, untill I get myself over the learning curve of KiCad or Eagle, and switch over to one of them.
 

Silvio

Member
Hi Lucas, I tried to download eagle but it was rather difficult for me to understand. I guess the reason was it was too complicated. I don't know anyone who could teach me how to use it. I will try to download PCB express and try to figure out how it works. The other thing is then when selecting components you have to be specific and at times its only some kind of part number or reference is displayed and this will get me guessing again if I took out the right component or not. The other task is to draw the schematic with the software provided first. All this will get on my nerves and puts me off. Don't get me wrong I am a patient person in a lot of ways but if I could find someone to give me some instructions how to operate it maybe it will be much easier for me to understand
.
What I do to produce my pcbs.
I configure out the tracks on a piece of paper and re draw them in Microsoft publisher making rectangles ovals triangles etc. I will graph the page in 0.1 inch pitch firstly so with a piece of perforated board and a vernier caliper by my side it would be easy counting the number of holes needed according to components used and giving necessary clearance.
I will draft it in the mirror image so that I am seeing the tracks from the component side. This will help me see every point where it should go and avoiding reversing mistakes. I will then print it on a piece of paper and check it according to schematic. This will help me see the real thing and I can put components on it and take a general view of what I done.
After that I can draw the components on it and re-color the tracks to a lighter shade and this will produce the component layout. I can also produce a silkscreen for the top layer. IF the need comes to produce the pcb in actual layout instead of mirror image I would simply reverse it. Schematics are also drawn with publisher and are really nice and colorful. (see 1000w smps)
If you took a look at my PDF files posted here they are all made this way. They are accurate in size as everything is measured accordingly.
I guess I take a longer way to it, but at least I found a remedy.

regards,

Silvio
 
Is the value of the bootstrap capacitor (C4) 470n? In other schematics it's either 100u, 4.7u, 10u...
From my experience, more than 1u slows the switching of the high side mosfet.
 

Silvio

Member
Yes it is 470nf (0.5uf) well to switch an IRf740 is really enough. The wave forms taken are across the primary winding, and as you can see it is symmetrical at no load and also at full load. The larger the bootstrap cap the longer it takes to charge up, again having a film capacitor instead of an electrolytic is more reliable. It all depend on the type of fet used and gate charge needed at the operating frequency.

If you take a look at the pictures in this tread you can see what I got 700watts out of irf740 and again I used 470nf as bootstrap cap.

Link http://www.diysmps.com/forums/showthread.php?894-Uni-SMPS&p=14144#post14144

Silvio.
 

Silvio

Member
hi Silvio,
Have you powered you amp with this SMPS? How is the sound?

No I have not tried it yet but a radio works well beside it. With high load more EMI is generated and I suggest that you screen the smps side in the box to block it out as much as possible. As you can see from commercial smps they all come in a screened box. Mine is naked :)

Silvio
 

jazzclassics

New member
Hi Silvio,

thank you very much for your reply. For me, it is not hard to find a metal box to house the PCBA. So, those SMPS you develope are not for audio ampsl Are you playing radio transmiters?
 

Silvio

Member
Hello Jazz,
Those smps I develop are for audio purpose. I myself being a newbie in this field I am trying to learn as much as possible. I make a lot of reserch so I learn the subject as much as possible. Seeing what other people do and try to learn from them. I try my hand and experiment a lot. I do this purely for hobby. I will buy a ready made amplifier board and fit one of my smps to it. I am also a radio amature but I don't practice anymore the internet killed it all.

Regarding the ripple on the output depend on the input and output capacitance used. An unregulated smps has less emi than a regulated one. The reason for this is that it works near full duty cycle and produces less noise on the output. One other factor is that there may be always some 100Hz ripple present. The more smoothing on the output the better. Having this reserve makes the power supply work more smoothly and power to the amp is always present when needed in the capacitor bank. The power is mostly consumed by the base notes. Smps has an advantage over a normal trafo as it tends to recover the power to the caps more quicker having 60Khz pulses per second compared to a normal mains transformer having only 50 or 60Hz per second. For this reason the output capacitance is not so large in smps.

One drawback thus in smps is that it tends to generate EMI (Electro magnetic interference) due to the square wave which is more subject to harmonics. Well the noise is generally out of the audio spectrum and the human ear can only hear up to 20Khz of which for myself I don't hear more than 14Khz more than that is inaudible for me. A young boy or girl may hear up to 20Khz but as you get older the highest audible frequency tend to go down.

Consider only this that when you operate an amplifier and is working on low volume the noise is hardly or not audible at all. The reason is that even the smps is at light load and and emi is at its lowest. Increasing the volume tend to hide it and it will not be noticeable. For all the reasons explained it is best to have an smps that can cover the load continuously and not needing to peak its power frequently.

Regarding the amplifier box you can use one side for the smps and screen it with a metal divider so that a compartment will suit the smps and the other the amp itself. This will screen it from the amplifier. A band pass filter in the input of the amplifier will block any high frequencies entering the preamp stage.

One last thing I like to mention is that you follow the blog post about winding transformers for smps. There is a guide to help you do it properly and safely to avoid the risk of the smps killing you. I also posted a video on youtube to help you understand how to do it.

I hope this will help understand more about smps

Regards

Silvio
 
Last edited:

jazzclassics

New member
the human ear can only hear up to 20Khz of which for myself I don't hear more than 14Khz more than that is inaudible for me. A young boy or girl may hear up to 20Khz but as you get older the highest audible frequency tend to go down.

Regards

Silvio

Exactly! I remember when I was a teenager ,I could hear the sound of a 20Khz sine. It was pretty annoying to me because most of the sound at that band was sort of noise. I could be annoyed by the sound of the transformer in a CRT TV receiver especially when it was switched it on. Regretably, I had no signal source and replay system to make pure signal above 20K, otherwise I would have no doubt whether a 20KHz upper limit apples.

My main purpose for making SMPS is not for audio, but for chargers. I think in the future decades, what I called as Peraonal Vihecle will be prevalent all over the world. A PV could be a bike, a car, a boat, or an airplane, the common character of them is they all electrically powered. A super capacitor or a battery provides the energy while a motor converts the energe to power for motion. Before nuclear generator could be made small enough, and man has found better way to deal with the waste, it is excluded as an option. And a charger will remains to be needed for all such PVs. Therefore, the ideal SMPS in my mind is something capable of taking 3phase AC and supplying up to 750V, 100A DC.

I am also an audio hobist, so it is easy to start with a SMPS for audio. EMI must be stricly dealt with. Although we cannot heat the Switching noise, there is no scientific proof that such signals will not interfare the audio band by transducers such as transformers, inductors and speakers.
 

Silvio

Member
Hello Jazz,

Well if you need the smps for a charger I was wondering what output voltage where you looking for. If I am not mistaken you do not need more than 200 watts. well if you do not need more than 24-26 volts then you can easily convert an ATX psu from an old computer which is driven by a TL494. You can make it with variable voltage and also current limit. I converted one myself and works very nicely. I can vary the voltage from 1 to 30 volts. At the peak current 8-10 amps however the output will drop to about 26 volts. The current can be set from around 0.1 to 10 amps. Provided that you leave the existing trafo as it is just cutting the center tap to make use of the full voltage available.

About EMI if its purely for a charger then why does it need to be so clean?

About the strict need of low EMI needed then you should opt for an ZVS and resonant soft switching smps. Here comes more complications in design as it may need a PFC as well. I cannot help you here as its beyond my knowledge.

Regards,

Silvio
 

Silvio

Member
Hi Jazz sorry about my last tread I mixed you up with someone else. silly me mentioning 200 watts.
Well I guess you are thinking big with 750v and 100 amps well that will be 75 Kilowatts. That I guess will be a small sub station serving your street lol.

What exactly are you looking for?

Silvio
 

kmiecikse

New member
Hi Silvio,
I want to enter into SMPS world and I have question for you.
On your device picture there is one loop (brown wire).
My question is:
What is the purpose of this loop? Is it current limiter feature?

Sebastian
 
Last edited:

Silvio

Member
@ Kmiecikse

The purpose of the brown wire forming a single loop around the trafo.

This produces around 6 volts ac which then this voltage is rectified by a diode and feeds the gate of the triac which has its terminals MT1 and MT2 tied across the input limiting resistor. The triac is used instead of a relay to short out the current limiting resistor in the input after charging up the main bulk capacitors to minimize losses.

Silvio
 

naminator200

New member
Dear Silvio,

I have built the SMPS, but the current limit circuit seems to be working incorrectly in my case. I have placed R10=4.7K and still it does not resist to a short circuit. Transistors get burned. Also, at about 2A load current, a few volts of peak-to-peak ripple appear on the output rail.
Are there any considerations about C10 capacitor? Maximum voltage, type?

Thank you and best regards,
Mihai
 
Top