Connect with us

Help needed for 12V lead-acid switched mode battery charger.

Discussion in 'Power Electronics' started by OhMy, May 3, 2014.

  1. OhMy

    OhMy

    63
    0
    Apr 5, 2014
    Hello!

    There is a need for a quite powerful battery charger, which wouldn't mind helping at the motor starting moment too. That would mean a regulated output voltage of 14.4V and max current of 30A. It would be very nice, if the output voltage could be isolated from all the rest of the circuit- voltage regulation feedback through an optocoupler perhaps, and the current sense via the primary winding- when the max current is reached, the output voltage may drop to maintain the current. I'm sure, it can be done, but not with my own knowledge :(. I have searched for a proper schematic for almost a week now, but with no real success.
    What I have (or will have in a 3 weeks time) from the main parts, which could be used: UC3843; TL494; SG3525; IR2153; IR2110; TL431A. And, of course, an unknown type of ferrite core EE type 42x42x19.5mm outside and 12x19.5mm core. Would it be sufficient for this much required power? According to some SMPS calculators, it should be, but what about the real life?
    Calculated windings for a 200-240VAC supply and a half bridge configuration at 50kHz should be Primary-23(0.6x3 wires) + Secondary-3+3(0.6x14 wires). Not sure, if they would fit, though.

    Please, share your thoughts about this.

    Edit: Some of the possible circuits I found for modification:
     

    Attached Files:

    • 3.png
      3.png
      File size:
      193 KB
      Views:
      6,751
    • 1.jpg
      1.jpg
      File size:
      47.2 KB
      Views:
      3,370
    • 2.gif
      2.gif
      File size:
      30.1 KB
      Views:
      3,579
    Last edited: May 3, 2014
  2. OhMy

    OhMy

    63
    0
    Apr 5, 2014
    Monologue continues... As they say- it's nice to speak to someone clever :D.

    This is what I came up with combining these three schematics. The result can be seen in "Ladetajs.gif".
    If only someone could take a look at it for big mistakes...
     

    Attached Files:

  3. (*steve*)

    (*steve*) ¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd Moderator

    25,192
    2,694
    Jan 21, 2010
    I was going to suggest a modified PC power supply. I'm not sure if that's an option for you as the output won't be floating.
     
  4. OhMy

    OhMy

    63
    0
    Apr 5, 2014
    Sadly, but I don't have any spare PC power supply. And I guess, that it's easier to build something new, that it would be to modify a complete device without actually knowing its real circuit.
    So- any obvious flaws in my schematic?

    Edit: Just noticed one very obvious error in the voltage feedback section- there definitely something is missing, but not sure what. That part of the circuit was taken from some German web page, so it's their mistake.
     
    Last edited: May 4, 2014
  5. OhMy

    OhMy

    63
    0
    Apr 5, 2014
    Perhaps this solution would be better. It's from here.
     

    Attached Files:

  6. KrisBlueNZ

    KrisBlueNZ Sadly passed away in 2015

    8,393
    1,266
    Nov 28, 2011
    You say you have an "unknown type" of ferrite core? The chances of it being suitable are not that high.

    I agree with Steve. You would be best to start with a PC power supply. If you don't have one, you should be able to find one at a dump or from a friend... You can control the output voltage by modifying the output-side circuit that drives the optocoupler, to add smooth current limiting to the voltage regulation...
     
  7. OhMy

    OhMy

    63
    0
    Apr 5, 2014
    The cores I have are mostly from old TVs. They had windings with multiple strand wire, so I hope, that they were used in high frequency applications and might might suit my need. If one doesn't satisfy, I can try a different one.

    I really don't think, that I stand any chance on getting a powerful enough and not totally broken PC power supply for modification, but they are quite pricey to buy a new one. A day in a dump yard would be a blessing :rolleyes:

    For now my best bet would be to build something from scratch, if only I got approval on the workability of the composed circuit.
     
  8. KrisBlueNZ

    KrisBlueNZ Sadly passed away in 2015

    8,393
    1,266
    Nov 28, 2011
    I think you're probably wasting your time using ferrites out of TV sets. You want an output of 14.4V at 30A which is 432W. No transformer in a standard TV will carry that kind of power. There are many factors that will need to be correct just for the circuit to deliver a few watts. And if anything is wrong, the most likely problem will be core saturation, which will probably cause the driving MOSFET(s) or transistor(s) to lose their magic smoke very quickly. You can calculate the AL value by winding a known number of turns around the core and measuring the impedance, but you would have to guess at the saturation current. In other words, if you try to use your junk box cores, I think you would end up cutting your losses anyway, and buying a known core, or preferably, a complete PC power supply.

    Here are my (slightly informed) opinions of the schematics you've posted.

    The design called 3.png, named "50 KHZ SMPS 900 WATT OUT" by Veysel Arslan, Rev 1.0, 30.10.2007, looks reasonable. At least, I can't see anything definitely wrong with it, but I'm no expert. You would have to change the transformer winding to reduce the output voltage from ±70V to 14.4V which would be tricky given the very small number of turns (which in itself is suspicious). I would use more turns, on a core with a lower AL value.

    The diagram called 1.jpg is very similar to 3.png but it has a single output. Many component values are missing, as is the important details of the transformer.

    The design called Ladetajs.gif also looks like a modified version of the Veysel Arslan design, but it is not isolated! At least, not the way it's drawn. The person who drew it up made a mistake and connected the primary and secondary 0V rails together. It's obviously supposed to be isolated, because the optocoupler is still present. I take that as an indication of an inexperienced designer, and for that reason alone, I wouldn't use that design. It is also missing winding information for the magnetics. As it is drawn, its output is at half mains potential. Unsafe and illegal.

    The diagram called 2.gif has nothing obviously wrong with it (as far as I can see) but it is poorly drawn which I take as a sign of an inexperienced designer.

    4.gif is only part of a power supply.

    sg3525-smps-Aizsardz.gif is not complete, and uses two centre-tapped primaries on the transformer. I would need to see a circuit description to comment further.

    It would be better for you to link to the pages where you got these designs from, since they will include a circuit description which will help us understand why certain design decisions were made nd how the power supply is intended to perform.
     
  9. OhMy

    OhMy

    63
    0
    Apr 5, 2014
    Thank you, KrisBlueNZ, for such a long post, I really appreciate that.

    From design called 3.png "50 KHZ SMPS 900 WATT OUT" I took the IR2110 driver part and both power supplies (320V and 12V).
    From sg3525-smps-Aizsardz.gif I used only the circuit around the SG3525 (some values are changed) and the way the optocoupler's transistor is connected.
    From 4.gif (can be seen here) I took the overcurrent protection section.
    2.gif is not used for now, but it should be working, it came from some Russian site.
    The TL431 part is from here.

    It's my "masterpiece" :D, combined from different sources, so no wonder about mistakes. The common ground was caused by inattention.

    The data for the transformer was calculated in ExcellentIT (supposed to be a very good program- Transformer.JPG).
    Corrected circuit- Ladetajs.gif.

    You're welcome to leave more comments on this.
     

    Attached Files:

  10. KrisBlueNZ

    KrisBlueNZ Sadly passed away in 2015

    8,393
    1,266
    Nov 28, 2011
    OK, I really can't comment authoritatively because I have very little experience with forward converters - my experience has been with flyback converters.

    Do you need external diodes across the drive MOSFETs? Or are the internal ones adequate?

    Are you planning to use 0.6 mm diameter wire for the transformer secondary? Or did I misread that screen capture?

    Other than those points, and the general untidiness of the schematic, I can't identify any possible problems.

    I think Adam (Arouse1973) has experience with forward converters. I'll PM him and ask him to comment.
     
    Last edited: May 5, 2014
  11. OhMy

    OhMy

    63
    0
    Apr 5, 2014
    Not sure about the snubber diodes, I don't remember seeing them in this type of converters, probably the 47R + 470pF across the primary is meant to be enough.
    According to that calculator windings should consist of 3 strands of 0.6mm wire for the primary, and 13 strands of 0.6mm wire for the secondary winding, 0.59mm diameter being an optimal choice for the 50kHz frequency. I can share that calculator, in case if anybody is interested.
     

    Attached Files:

  12. Arouse1973

    Arouse1973 Adam

    5,164
    1,078
    Dec 18, 2013
    Hi Kris can we just recap what the issues are here, there is a lot going on in this thread and I am struggling to pick through it.
    Adam
     
  13. OhMy

    OhMy

    63
    0
    Apr 5, 2014
    Hello, Arouse1973. If you could just take a look at the circuit in my post #9 and tell your prognosis about its chances of working and delivering 14.4V @30A (not for extended periods), also if the voltage feedback and current limiter is correct?

    Thank you.
     
  14. Arouse1973

    Arouse1973 Adam

    5,164
    1,078
    Dec 18, 2013
    Ok give me sometime and I'll get back to you.
     
  15. KrisBlueNZ

    KrisBlueNZ Sadly passed away in 2015

    8,393
    1,266
    Nov 28, 2011
    Adam: OhMy wants to make a battery charger powered from 200~240VAC that can supply a fixed 14.4V DC at up to 30A with smooth current limiting.

    He is operating on a budget. We've suggested that he starts with a PC power supply and modifies it, but he says he can't get one. He wants to build one from scratch. He has some ferrite cores (I think they're all salvaged, with no details on any of them) and he has ordered UC3843, TL494, SG3525, IR2153, IR2110, TL431A.

    He has posted several designs he has found for forward converters, probably half bridge using two MOSFETs, and combined several of them into his own prospective design, which he calls Ladetajs. The latest version so far is in post #9.

    OhMy: Is the 1:100+100 transformer between the MOSFETs and the main transformer primary for current limiting? I'm not sure, but I think current limiting would be better done at the output, where it would cause feedback through the optocoupler just as the voltage monitoring circuit does.
     
  16. OhMy

    OhMy

    63
    0
    Apr 5, 2014
    Yes, it is for current limiting, and I guess, that it could be put from the secondary winding's center tap to ground as well. Only the turn ratio probably should be 1:10 not 1:100 or so, and no optocoupler would be needed.
     
  17. Arouse1973

    Arouse1973 Adam

    5,164
    1,078
    Dec 18, 2013
    Hi Kris
    I think its quite dangerous using transformers from something else as you know saturation current can be very dangerous at that high current. I would have gone for a poly phase converter. I'll have a look but I am a bit concerned about these salvaged transformers.
    Cheers
    Adam
     
    Last edited: May 5, 2014
  18. Arouse1973

    Arouse1973 Adam

    5,164
    1,078
    Dec 18, 2013
    Ohmy Ohmy sorry had to say it. What PWM are you running on.
    Thanks Adam
     
  19. OhMy

    OhMy

    63
    0
    Apr 5, 2014
    Hi, Adam.
    o_O Sorry, not sure, what you mean? Maybe SG3525?
    Me too, but I get a heart attack every time I see the prices for new ones :eek:.
    The only ferrite core with some kind of marking on it ("N27" stamped on both E halves) is broken and has a quite large air gap in the middle. A bit of glue could fix one issue, but what about the air gap?
    The biggest core I have is 58mm in length with a cross section of 12x22.5mm, but it looks noticeably different from the others, it's like made of tiny flakes (could it be iron powder?).
     

    Attached Files:

  20. Arouse1973

    Arouse1973 Adam

    5,164
    1,078
    Dec 18, 2013
    Ok looks like it's good to go. Personally I would re-draw the switching part to look something like this.
    New Picture (3).png
    It's a bit easier to follow as is the conventional way to draw this. The turns ratio for the output is spot on only thing to make sure is you get the dot convention correct see drawing below.

    New Picture (4).png
    I am not sure about the current sense transformer. I have only seen a 1:50 single primary and secondary positively rectified. It's in the right place protecting the primary in case it does not start up correctly. I would put some diodes across the Mosfets as shown and not use the internal diodes, they are not defined on the data sheets for a reason. Fast Shockley diodes are probably your best bet. You also have to make sure the capacitors you use are suitable low ESR, ESL and high ripple current types with good temperature range. Polymer types seem to be a popular choice. The voltage regulation looks like the standard shunt regulator type. I can't see anything wrong here but I have not worked out the values.

    I still have concerns about the cores, N27 is a high frequency core material but there are several types that all have different characteristics. The calculations can be quite time consuming to do so really it's up to you if you want to continue using them not knowing how they are going to respond in this circuit. I am not a transformer designer so I am not going say either way.
    Hope this helps

    Thanks
    Adam
     
Ask a Question
Want to reply to this thread or ask your own question?
You'll need to choose a username for the site, which only take a couple of moments (here). After that, you can post your question and our members will help you out.
Electronics Point Logo
Continue to site
Quote of the day

-