• State Resolved
  • Locked Locked
  • Replies 20 replies
  • Subscribers 64 subscribers
  • Views 1151 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

Symmetric 5v from symmetric 30v supply

David Lake
Intellectual 300 points
Other Parts Discussed in Thread: TL2575-05, LM3102, LM2675

I need positive and negative 5v from a symmetric 30vDC supply on an amplifier project, its got to be simple and efficient with easy to solder packages, the TL2575-05 looks suitable but its not easy finding a matching negative equivalent.

  • 0
    TI__Genius 16025 points

    Hi David,

    Based on your requirement of a negative output, I've attached the following schematic which is both simple and efficient.

    0636.E2E Simple Switcher - David Lake.pdf.

    While the TL2575 is suitable, you might run it a little warm for your application. I would suggest using the LM3102 to increase the efficiency of your project.

    Let me know if you have any further questions.

    Thanks,

    Anston

  • 0 in reply to Anston Lobo
    Intellectual 300 points

    Thanks but I dont like the packages the LM3102 comes in I have limited experience with SMT and I am at risk of messing it up, an SOIC or TO-220 package would be easier for me.

    The drawing is lovely but theres a lot of components, is theres a simpler solution even if its slightly less efficient?

  • 0 in reply to David Lake
    TI__Genius 16025 points

    Your welcome, As far as negative bucks are concerned, no one makes a device that works in that quadrant. Hence the need for a current mirror to set Vout on the existing layout.

    We will send you an updated schematic using the same 30V input to drive both positive and negative rails symmetrically shortly. I recon our updated solution will overall reduce the number of components on your power supply stage and we will keep in mind that you'd prefer SOIC or TO-220.

    -Anston

  • 0 in reply to Anston Lobo
    TI__Genius 16025 points

    Hi David,

    Attached is an updated schematic which should suffice based on your application. 5241.E2E Simple Switcher - David Lake (Revised).pdf

    Let me know if you have any further questions.

    Thanks,

    Anston

  • 0 in reply to Anston Lobo
    Intellectual 300 points

    SOIC and TO-220 are just examples of packages with a wider pin spacing which I prefer.

  • 0 in reply to David Lake
    TI__Genius 16025 points

    In the revised schematic (attached in my previous post) we provide an extension of your existing design to provide the negative rail. Take a look and give us your feedback. 

    We'll work with you until we come to an ideal solution. Let me know if you have any questions.

    Thanks,

    Anston

  • 0 in reply to Anston Lobo
    Intellectual 300 points

    On the revised circuit theres an oscillation on the positive output and the diode on the negative output causes the voltage to drop by the forward voltage of the diode but I have solved this by putting another diode after the inductor on the positive side.

    Could you provide component values for the revised circuit please?

    You said the TL2575 would run hot, the load current is unlikely to go above 0.1A does that make a difference? Would the TL2575HV be any better?

  • 0 in reply to David Lake
    TI__Genius 16025 points

    Hi David,

    We filled in a few unlabeled component values on the previous schematic. 3122.E2E Simple Switcher - David Lake (Revision 2).pdf

    This topology is a buck with an over-winding. Additionally we've placed a SEPIC cap between the two winding's which greatly improves cross regulation. The negative output is not truly regulated but is very close to the main output, a difference based on forward voltage characteristics and for symmetrical load current should be quite close. 

    If you need a truly regulated negative output then you will need a second switcher to generate that output. For a discussion on this please refer to http://e2e.ti.com/support/power_management/simple_switcher/f/858/p/260859/914085.aspx

    I hope this helps. Let me know if you have further questions.

    Thanks,

    Anston

  • 0 in reply to Anston Lobo
    Intellectual 300 points

    My idea of putting a diode after the inductor on the positive side is not only much simpler but also evens out the voltage on the negative side by applying the same v drop on the positive side allowing the regulator to compensate for the v drop and completely eliminates the oscillation when its powered on.

    Your revision 2 reduces the oscillation using the over winding but does not eliminate it, the 10uF caps do nothing and it does not fix the >0.5v drop on the negative side.

    Are there any advantages to your solution and is there a problem with my solution?

  • 0 in reply to David Lake
    TI__Genius 16025 points

    Hi David,

    Our stance, the T1 primary has 5.5V across it when its associated diode is conducting and the secondary will therefore have 5.5V across it as well. The error in the negative output, will be dependent on the difference in the load current across the two rails. If your 150uF capacitors are ideal, the 10uF capacitors will have no effect, however it is good practice to bypass Aluminium Electrolytics with 1-10uF ceramic. If the 150uF capacitors are tantalum or polymer, then the 10uF ceramics can be avoided.

    Thanks,

    Anston

  • 0 in reply to Anston Lobo
    Intellectual 300 points

    What diode associated with the T1 primary? There is no need for a transformer/overwinding the voltage drop on the negative output is caused by the capacitor being charged via the diode so if the positive output capacitor is charged via a diode also positive will get the same volt drop as negative causing the switcher to compensate and bring them both up to 5v, I can make this more efficient by using a diode with a 340mv forward voltage.

  • 0 in reply to David Lake
    TI__Genius 16025 points

    Hi David,

    I was referring to the B260B Diode present to the immediate right of the IC between ground and the SW node. In either case, it looks like you've achieved the ideal solution for your circuit. 

    If you have any further questions I can help you with please don't hesitate to ask.

    Thanks,

    Anston

  • 0 in reply to Anston Lobo
    Intellectual 300 points

    You are right I didn't realise that's what that diode is for.

    I found the LM2675 5.0, it seems ideal, how important is the input capacitor if I have a 33mF cap with an ESR of 0.042 ohm at about 8cm from the switcher?

    The Webench designer that I have just started using suggests diodes with a VRRM of over 44v, I found diodes with forward voltage of 340mv but the VRRM is 40v, would these be OK?

    Also the Webench designer shows the LM2675 will run at about 43C if that's without a heatsink that's great!

  • 0 in reply to David Lake
    TI__Genius 16025 points

    Hi David,

    If the 30V is well regulated (or well limited) then the 40V diodes are probably okay. However, if the +/- 30V rails are from an AC line powered transformer then you need to consider what +/- 30V will be under high line conditions (130VAC?) and under light output loading. Ideally you should use 50 or 60V rated diodes. They will have slightly higher Vdrop all other things being equal.

    If you really mean 33milifarads or 33,000uF then you will still need to add a 4.7uF ceramic in parallel and very close to the IC. Aluminium input capacitors are inductive and need a parallel ceramic in switching topologies/applications.

    I hope this helps. Let me know if you have any additional questions.

    Thanks,

    Anston

  • 0 in reply to Anston Lobo
    Intellectual 300 points

    We have 240v here and its very stable, the actual voltage will be a symmetrical 30.65v, but the difference between 340mv and 500mv is only about 0.02w so I wont bother with the lower forward voltage diodes.

    Yes I do mean 33,000uF it is a multichannel class A amplifier so the more the merrier, if I use a PP film cap for Cin would it still need to be 4.7uF or different?

  • 0 in reply to David Lake
    TI__Genius 16025 points

    Hi David,

    Can you specify the make and model of the PP Film capacitor? Anything that's wound will have an inductance associated with it. PP Film caps fall in that category though other properties about their manufacture render them efficient in limited scenarios such as decoupling or bypass caps. 

    Thanks,

    Anston

  • 0 in reply to Anston Lobo
    Intellectual 300 points

    Wima is my preferred brand of film capacitor.

  • 0 in reply to David Lake
    TI__Genius 16025 points

    Hi David,

    Based on the information presented here [http://www.wima.com/EN/mkp4.htm]. This type of capacitor is non inductive since the construction is layered and not wound.

    Thanks,

    Anston

  • 0 in reply to Anston Lobo
    Intellectual 300 points

    Thats great! So would it still need to be 4.7uF?