• State TI Thinks Resolved
  • Locked Locked
  • Replies 6 replies
  • Answers 1 answer
  • Subscribers 62 subscribers
  • Views 362 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.

UCC28704: my AC/DC 10W adapter can't reach 2A after following design rules

user5960859
Prodigy 40 points
Part Number: UCC28704

Hello all,

we have used ucc28704 for our AC/DC 5V@2A adapter and we have "copied" the schematic of "UCC28704EVM-724 10-W USB Adapter", the same BOM and the same transformer (750315841 Wurth Elektronik). We have also followed the layout recommendation. Our issue is that the adapter doesn't work like the "UCC28704EVM-724 10-W USB Adapter User guide" says and after tunning some components values we are able to reach 5V@2A but after some seconds, output capacitors and transformer get hot (more than 65ºC) and Vout falls down.

The first time that I connect our adapter Vout is around 3V, I thought the issue was Vdd or Vs, after discarting Vdd I have to reduce Rs2(R12) from 36.5K to 19K in order to match Vout at 5.1V.
When I try to check the maximum current with an electronic load I realise that it protects itself at 0.9A, I appreciate that the switching frequency is 70-85KHz (measuring with an oscilloscope in the anode of the diode of the secundary side). So I think that it is in overcurrent protection.

The next step is to reduce Rcs(R15&R16) from 1.075 to 0.717 Ohms and the maxium current incrase up to 1.5A and 70-80Khz.

After reducing a little more Rcs up to 0.54 ohms, the maxium current incrase up to 2A but 45-50Khz but after some seconds the adapter protects itself.

I reduce RLC (R13) from 715 to 220 ohms like formulas says and I had to increase R10 from 20 to 50 ohms due to Vdd was bigger than 35V.

1. Is this tunning right? I would say that if Rcs is reduced more than 0.8 Ohm, Ipp is increased above the current satuuration of the transformer (800mA) and that is no good.

2. How could I fit a little more in order to reach 2A properly?

3. Is possible to fit the components values to have 2A with 70-85 Khz?

Thanks in advance!

P.D: I have checked everythig several times in order to avoid mistakes

  • 0
    TI__Mastermind 30330 points
    Hello user5960859,

    Thank you for your interest in the UCC28704 flyback controller. I'm sorry that you are experiencing difficulties with your circuit.
    If your design is an exact copy of the EVM design, it should work exactly like the EVM, too.

    Please check the polarity of your output capacitors. Your statement that the output caps and transformer heat up considerably after a few seconds strongly suggests to me that your caps may have been installed backwards.
    Backwards caps will breakdown and cause overcurrent internally and heat up rapidly. The controller will limit the current through the transformer to the overload set point and you may not see very much current at the output because most of it is going through the backwards caps.

    Please check the installation polarity of the caps. Assuming that they are backwards, I recommend that you replace them with new ones, since they are likely to be damaged by the reverse voltage and current. Also, restore all of the other components back to the EVM values. The "tuning" that you did while the caps were backwards will be highly inappropriate for a normal design.

    Regards,
    Ulrich
  • 0 in reply to Ulrich Goerke
    Prodigy 40 points

    Hello Ulrich,

    Don't worry about difficulties, they are a challenge.

    I have checked output capacitors polarity one more time and they are OK. When I said "some second" I had to have said "minutes".

    But let´s start again, our design is a copy of EVM design (with our shape) and when we connect the adapter Vout is around 3V.

    What do you think the issue could be?

    thx!

     

  • 0 in reply to user5960859
    TI__Mastermind 30330 points
    Hello user5960859,

    Okay, I'm glad the caps are in properly. To be blunt about it, there is something wrong with your implementation, because we have a lot of EVMs that work as designed. We just have to find out what the problem is on your particular board.

    It could be a problem with the pcb layout, where a track may be shorting to another track. It could be some other polarized component, like a diode that is in backwards. Since the output is positive voltage and the controller works okay, I don't think it is the output diode or the AUX diode.

    That gives me an idea: please check the polarity of the flying leads of the transformer.
    If these are reversed, you will get a forward-mode type of conversion, so there will be high peak currents of short duration on the primary side. This may be enough to get a few volts out, but not full power. Also, the AUX winding will not see a reflected output voltage, so the controller will tend to regulate that winding instead of the output winding. Thinking about it more, this can very likely be the source of the problem. It would be easy to mix up the leads when installing the transformer into the board. The high peak forward currents might overheat both the transformer and the output caps.

    If that turns out to NOT be the problem, then we are still left with trying to find some other root cause... something not installed correctly, or wrong value, or bad connection, or a part value does not meet its specification.
    As I mentioned before, please restore all of your "tunings" back to the original values. The board should be debugged with the correct parts in place. Once the true problem is found, it won't need the "tuning" changes, and they may actually harm the board.

    Regards,
    Ulrich
  • 0 in reply to Ulrich Goerke
    Prodigy 40 points
    Hi,

    Flying leads checked, they are ok.

    I have mounted a new prototipe with oringinal values and the same issue.

    Vdd measured with a multimeter in DC = 13-15V (I don't have isolated AC).

    Regards
  • 0 in reply to user5960859
    TI__Mastermind 30330 points
    Hi User,

    At this time, I think we'll have to examine some waveforms for more clues to the problem.
    Since you now have a second board with the same problem, it suggests that this is a systematic problem, not a one-time mistake in assembly.

    Without isolated AC, you will be able to only capture waveforms on the secondary side of the transformer.
    Referring to the EVM schematic (page 5 of the User Guide), let's start with secondary current and transformer voltage.
    Please capture the current in one of the flying leads (which ever is more convenient) with a current probe, oriented for positive current.
    Along with current, capture the secondary winding voltage at FL2 with respect to FL1 (should be SGND), and Vout.
    Capture one or two switching cycles during steady-state.
    Also capture a long-time sweep (say 500ms/div) to ensure that the controller and switching is not shutting down periodically and restarting. Please state the operating conditions (input voltage, output load, etc.) for every test. Set you load for Resistance Mode and vary between 10 ohms and 3ohms.

    I have a few other questions:
    1. Is it possible to obtain a 1:1 isolation transformer for your AC input? For 10W, it can be a small one (500VA rating, for instance).
    It may be difficult to debug without being able to probe on the primary side, and you will need isolation from the line to do this.
    2. Did you procure the 750315841 transformer from Wurth Electronik, or did you construct your own using the User Guide details on pages 13 and 14?
    3. The schematic shows secondary snubber parts C4 and R7 with no value (depopulated). Do you have any parts installed in these locations? If so, what values?
    4. Did you make any substitutions for any of the parts in the Bill of Materials?

    Regards,
    Ulrich
  • 0 in reply to Ulrich Goerke
    Prodigy 40 points

    Hello Ulrich,

    finally we have found the mistake, there was a swap between pin 2 and 6 in the transformer footprint. So, we had a half of the turns of the primari coil and a half of the magnetic flux. I must say that it was checked several times but...

    Thank you very much Ulrich for your support.