Hello everyone!
I am building a switching power supply to be used in the CAR AUDIO field, feeding it from the wall network, obtaining an adjustable output voltage from 12.4v to 14.4v (and beyond).
I made a project from scratch.
The power supply uses the SG3525 PWM controller which drives an IR2113S driver which in turn drives a pair of very powerful IGBTs.
The power supply is regulated, and keeps its output stable even when we are absorbing almost all the deliverable current (or at least it should) and through a potentiometer that adjusts the reference of a TL431 which in turn drives the LED of an optocoupler, I get voltage regulation and setting.
I made the PCB and installed the main components to start testing it, currently I have not installed the protections (short circuit and thermal) because I was interested in knowing if at least it did its primary job (produce enough voltage and current as I thought it would).
After doing everything, instead of immediately using the direct electrical network, I used a transformer to isolate myself from the electrical network and obtain a lower voltage to avoid disastrous failures, in fact, with 78Vdc on the large capacitors, everything was fine, the power supply produced the voltage desired, but obviously little current, in fact by loading the power supply with a 40w car bulb, the voltage dropped to 7 volts.
So, I used a bigger isolation transformer, and with 110Vdc on the big capacitors, still all right, and the voltage was also more stable, the current delivered, higher, and the voltage dropped a lot less (down to 10v).
Everything seemed ok to me, there was no abnormal heating anywhere, so I switched to direct mains, first with a 220V - 40w incandescent bulb installed in place of the primary fuse and I got a lot more stability, but obviously when trying to absorb more current, the bulb started to light up and choked everything, then I replaced the bulb with a 47ohm 10w resistor, and I still got some improvements, then finally I decided to remove everything and install the fuse, now the problems started, because the power supply turned on without problems, but as soon as I connected the car bulb in output, one of the two internal diodes (of one of the 2 double power diodes) spontaneously shorted out suddenly, and having no protection systems I broke the two IGBTs and the driver accordingly.
Cabbage! everything was going well, I would have also installed the protections and I could have fully tested my power supply, but something went wrong. Maybe the output filter? Perhaps the absence of Snubber on the output power diodes? It wasn't just a coincidence, because I fixed everything, and it happened again. I will post the schematic and some drawings of the waveforms that I saw on the oscilloscope before the failure, I have drawn them by hand, because it all happened yesterday, and I did not have the thought of taking pictures of the screen.
And of course, this is the videotest:
video
thanks to all,
Aids are appreciated!
Mario
I am building a switching power supply to be used in the CAR AUDIO field, feeding it from the wall network, obtaining an adjustable output voltage from 12.4v to 14.4v (and beyond).
I made a project from scratch.
The power supply uses the SG3525 PWM controller which drives an IR2113S driver which in turn drives a pair of very powerful IGBTs.
The power supply is regulated, and keeps its output stable even when we are absorbing almost all the deliverable current (or at least it should) and through a potentiometer that adjusts the reference of a TL431 which in turn drives the LED of an optocoupler, I get voltage regulation and setting.
I made the PCB and installed the main components to start testing it, currently I have not installed the protections (short circuit and thermal) because I was interested in knowing if at least it did its primary job (produce enough voltage and current as I thought it would).
After doing everything, instead of immediately using the direct electrical network, I used a transformer to isolate myself from the electrical network and obtain a lower voltage to avoid disastrous failures, in fact, with 78Vdc on the large capacitors, everything was fine, the power supply produced the voltage desired, but obviously little current, in fact by loading the power supply with a 40w car bulb, the voltage dropped to 7 volts.
So, I used a bigger isolation transformer, and with 110Vdc on the big capacitors, still all right, and the voltage was also more stable, the current delivered, higher, and the voltage dropped a lot less (down to 10v).
Everything seemed ok to me, there was no abnormal heating anywhere, so I switched to direct mains, first with a 220V - 40w incandescent bulb installed in place of the primary fuse and I got a lot more stability, but obviously when trying to absorb more current, the bulb started to light up and choked everything, then I replaced the bulb with a 47ohm 10w resistor, and I still got some improvements, then finally I decided to remove everything and install the fuse, now the problems started, because the power supply turned on without problems, but as soon as I connected the car bulb in output, one of the two internal diodes (of one of the 2 double power diodes) spontaneously shorted out suddenly, and having no protection systems I broke the two IGBTs and the driver accordingly.
Cabbage! everything was going well, I would have also installed the protections and I could have fully tested my power supply, but something went wrong. Maybe the output filter? Perhaps the absence of Snubber on the output power diodes? It wasn't just a coincidence, because I fixed everything, and it happened again. I will post the schematic and some drawings of the waveforms that I saw on the oscilloscope before the failure, I have drawn them by hand, because it all happened yesterday, and I did not have the thought of taking pictures of the screen.
And of course, this is the videotest:
video
thanks to all,
Aids are appreciated!
Mario
Attachments
Did you try a different diode like those used in car amps?
Insert the lamp in series between the transformer and diodes to allow you to see current draw into the rectifiers.
Also try the lamp between the diodes and the inductor.
Insert the lamp in series between the transformer and diodes to allow you to see current draw into the rectifiers.
Also try the lamp between the diodes and the inductor.
Unfortunately not, Perry!
I had selected these diodes because the power supply must deliver a lot of current (more than 250A) and keep the output voltage stable, almost like some Brazilian power supplies do, in fact these diodes, branded IXYS mod. DSA120C150QB have characteristics (which seemed to me) impeccable, fast, voltage range much higher than useful (to rest assured), allowable current of 60Ampere for each diode, so each device can handle 120A.
I also thought that the diodes could be susceptible, also because I do not understand why at low rail voltage, they worked well, then with the direct electrical network they short out as soon as a load (even low) is connected, but I do not I was able to understand why.
Do you advise me to use "common" fast diodes like MBR 3045 or similar?
I had selected these diodes because the power supply must deliver a lot of current (more than 250A) and keep the output voltage stable, almost like some Brazilian power supplies do, in fact these diodes, branded IXYS mod. DSA120C150QB have characteristics (which seemed to me) impeccable, fast, voltage range much higher than useful (to rest assured), allowable current of 60Ampere for each diode, so each device can handle 120A.
I also thought that the diodes could be susceptible, also because I do not understand why at low rail voltage, they worked well, then with the direct electrical network they short out as soon as a load (even low) is connected, but I do not I was able to understand why.
Do you advise me to use "common" fast diodes like MBR 3045 or similar?
This is just to try to narrow down the reliability issue, not as a permanent solution. At this point in testing, you don't seem to be anywhere near the limits of the diodes used in amps.
I will replace the diodes with something more "human" also used in car audio amplifiers, and in addition, while there are, I will use snubber networks, both on the diodes and on the secondary windings, both before and after the output inductor, to avoid any water hammer in the diodes caused by the inductor itself and by the load.
What do you think about it?
What do you think about it?
Attachments
I haven't built any high-voltage switching supplies so I won't be much help. You might try posting this on the PS forum if you haven't already:
https://www.diyaudio.com/forums/power-supplies/
https://www.diyaudio.com/forums/power-supplies/
Thanks for the advice Perry, but I've already tried asking for information or opinions in that section.
They are all very active in other people's threads, except mine, in more than 48 hours I have not received even half a response, to the point that I thought I had the wrong section and I asked some administrator to move that thread here in the section " car audio ", where at least you, you managed to tell me something.
Maybe my thread is too long to read so people get bored and don't calculate me at all, or maybe there's no one really expert in power supplies that could help me, but the reality is that nobody calculated me.
Sometimes I wonder if it is I who venture into too many complicated things and the others do not follow me (or do not waste time following me) or they who do not keep up with me.
I want to learn, to understand how things work.
Having made that power supply (and its PCB) as well as I have made other working boards, for me it is a great satisfaction, even if they do not work very well, they make me understand that I have to go forward, improve and maybe in the future I can become someone.
But obviously not everyone thinks like me. Patience.
They are all very active in other people's threads, except mine, in more than 48 hours I have not received even half a response, to the point that I thought I had the wrong section and I asked some administrator to move that thread here in the section " car audio ", where at least you, you managed to tell me something.
Maybe my thread is too long to read so people get bored and don't calculate me at all, or maybe there's no one really expert in power supplies that could help me, but the reality is that nobody calculated me.
Sometimes I wonder if it is I who venture into too many complicated things and the others do not follow me (or do not waste time following me) or they who do not keep up with me.
I want to learn, to understand how things work.
Having made that power supply (and its PCB) as well as I have made other working boards, for me it is a great satisfaction, even if they do not work very well, they make me understand that I have to go forward, improve and maybe in the future I can become someone.
But obviously not everyone thinks like me. Patience.
Last edited:
I think the 1μF/400V capacitor that feeds the trafo needs to be connected to +310V, not to"float".
You are wrong! this is a half bridge power supply, the capacitor is there where it should be.