• State Resolved
  • Locked Locked
  • Replies 13 replies
  • Answers 1 answer
  • Subscribers 64 subscribers
  • Views 3377 views
  • Users 0 members are here
Support feedback
Options
Options
Related

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.

TPS54231 startup issue

Juan Luis Arreaga
Prodigy 50 points
Other Parts Discussed in Thread: LM317, TPS54231, CC3200

Hi John.

I was using a TI LM317 supplying a cellular radio too, Vi = 12V, Vo = 5V, this configuration works quite good but generate too much heat. So I decided to change to TPS54231. The GSM cellular radio sink current spikes of 1.5A - 2A max, I have installed other ICs sinking less than 200mA, giving like a 400mA max when GSM has low activity (no data link, no phone call).

But I have a problem that TPS54231 doesn't start-up correctly with full load, it keeps a 2V output instead a 5V, and less than 4mA, I installed a LED but it only flashes like 2-3 times and then turns off. Could it be that I need more than 2A on start-up?

What can you suggest me? Maybe I missed someting when changing my design from a simple LM317 to a little bit complex design for TPS54231. Why this only works as low current source for me? I used SwitcherPro Design tool.

I'll try to find out what's the maximun load current that I can get from TPS54231, and I'll let you know, maybe I can give you more information.

Thanks,

 

  • 0
    TI__Guru**** 191445 points

    I would need to see your schematics and any waveforms you can provide to diagnose your problem,

  • 0 in reply to JohnTucker
    Prodigy 50 points

    John, I can't get any waveforms right know, I'll try to get soon, but first I want to tell you that I made some basic tests:

    - Using a 1K resistor as load, I can get Io=5mA with Vo=5V. Everything is OK.

    - Using a 470 resistor as load, I can get Io=10mA, but Vo=4.8V. Too much drop out, when was designed for 2A output current.

    - Using a 330 resistor as load, I can get Io=14mA only for 1 second or less, and Vo=4.5V. Then drops to Io=5mA and Vo=2.2V.

    - Also, I tried 1k resistor as load, when stable I placed in parallel a 330 resistor to get almost a 250 ohm load, voltage falls down from Vo=5V to Vo=2V, then remove the 330 resistor and the Vo never got back to 5V.

    Here is the schematics, like I said, it was from SwitcherPro Designer Tool:

    I don't need R6 and R7 to connect to EN pin, left floating, it said that resistors were not necesary.

    I hope you can help me with this information, and thanks for you time.

     

  • 0 in reply to JohnTucker
    Prodigy 50 points

    Hi John, the test board I made is like the development board for TPS54231, and next I show you what's the BOM suggested by SwitcherPro Designer Tool, shaded with blue is what I got:

    Name Part Number Description Manufacturer Package Package Model Manufacturer
    C2 GRM31CR60J476ME19L Capacitor, Ceramic, 47uF, 6.3V, 20% muRata GRM31C 1206 1206
    C3216X5R0J476M
    		 TDK
    C3 Standard Capacitor, Ceramic, 0.1uF, 20V, 10% Standard 805 1206 VJ1206Y104KXXMR Vishay
    C5 Standard Capacitor, Ceramic, 8200pF, 10V, 20% Standard 603 0805 VJ0805Y822KXQCW1BC Vishay
    C6 Standard Capacitor, Ceramic, 680pF, 10V, 20% Standard 603 0603 0603YC681KAT2A AVX
    C7 Standard Capacitor, Ceramic, 47pF, 10V, 20% Standard 603 0603 VJ0603A470KXJCW1BC Vishay
    C9 GRM55DR61H106KA01K Capacitor, Ceramic, 10uF, 50V, 10% muRata GRM55D 2220 2220 293D106X9035D2TE3 Vishay
    D1 MBRA340 Diode, Schottky, 40V, 3A On-Semi SMA SMD-TO277A SS3P5L-M3/87A Vishay
    L1 ELLCTV150M Inductor, 15uH, 3.1A, 25mΩ Panasonic ELLATV 3131 LQH88P MuRata
    R1 Standard Resistor, SurfaceMount, 10KΩ, 100mW, 1% Standard 603 1206 10K, 1% KOA Speer
    R2 Standard Resistor, SurfaceMount, 1.87KΩ, 100mW, 1% Standard 603 1206 1.87K, 1% KOA Speer
    R3 Standard Resistor, SurfaceMount, 32.4KΩ, 100mW, 1% Standard 603 1206 32.4K, 1% KOA Speer
    R6 Standard Resistor, SurfaceMount, 348KΩ, 100mW, 1% Standard 603 ---- ---- ----
    R7 Standard Resistor, SurfaceMount, 71.5KΩ, 100mW, 1% Standard 603 ---- ---- ----
    U1 TPS54231 IC, Converter, 8 pins Texas Instruments, Inc. SOIC     TPS54231

     

    Also, I was searchig about switching designs, special the Buck type which TPS54231 works, and looking the critical components: and they are inductor, schottky diode and the C2=47uF capacitor, but this is for the output stage. But also the C9, the input capacitor is very important, having high current spikes, am I right?? Looking this BOM, inductor, schottky diode, C2 and C9 are the only that has name about manufacturer and model. It might be for some reason.

    I hope that all my component fill the requiremets of the suggested BOM, but I think that I made a mistake choosing C9, instead of choosing a Ceramic capacitor, I chose a Tantalum capacitor, It has all the requiremts about operating voltage, capacitance and almost the size, but looking the datasheet it can only deliver 0.45A current spikes. Do you know what's the current spikes I should expect? Looking the books, and making aproximations about numbers I should expect 1A current spikes from this capacitor.

    Do you think that this must be my problem?? Getting a 10uF ceramic capacitor could make it work? Because I don't know what should be the next step to make it work.

    I hope to find a solution with your help, thanks.

     

  • 0 in reply to Juan Luis Arreaga
    TI__Guru**** 191445 points

    It is entirely possiblre that is your problem.  The TPS54231 requires a X5R or X7R dielectric ceramic input capacitor.  An additional eletrolytic such as tantalum can be used as an addition to the ceramic cap, but you must have the ceramic and it should be located as close as possible to the VIN input of the IC.

    If that does not fix your problem you will still need to provide waveforms so we can determine what else might be wrong.

     

  • 0 in reply to JohnTucker
    Prodigy 50 points

    John, all my capacitors are X7R, but not C9.

    I'll buy some 10uF ceramic X7R o X5R and tantalum/electrolitic capacitor, but they are a little bit expensive... so, maybe I'll get next week and see if it works, I'll come back to thank you, or find the way to show you waveforms.

    Thanks,

  • 0 in reply to Juan Luis Arreaga
    Prodigy 50 points

    John, I have a good news... it works!! I can't believe that without that ceramic capacitor it wouldn't work. I didn't check if I get 2A, but I must get about 1.5A for my board... and by mistake I short-circuited the output, and it keeps working.

    I think my old tantalum capacitor is burnt by my last tests, I think it's not made to handle the requirements for a buck converter.

    Thanks for your time,

  • 0 in reply to Juan Luis Arreaga
    TI__Guru**** 191445 points

    Good.  I was pretty sure that was the issue.  We have seen this several times in the past.

  • 0 in reply to JohnTucker
    Prodigy 40 points

    I have startup problems. Below 5.25V the TPS54231 output motor boats roughly between .8 and ~2.8V (Fluke DVM).

    Datasheet shows a graph of 5V in 3.3 out at 2A says it will work with input above 3.5V

    My output 3.3V,  10K & 3.16K

    It works fine with input above 5.3V.

    All connections checked many times.

    Tried three boards, the same.

    Great ground plane.

    INPUT:  2, 10uF 10V X5R 0805, then added 2, 10uF 35V X5R 1206 = 40uF within 3mm from the ic. Then added OS-Con 1000uF.

    Output: 2, 47uF 10V X5R 1206, then added 2 10uF 35V X5R 1206 = 114uF within 3mm from the inductor.

    Changed the 10uH inductor to 5A low mOhm ferrite shielded 12mm type

    Doubled the 0.1uF 50V 0402 boost cap to 0.2uF

    Diode is MBR230 30V 2A schottky.

    Slow start has been 0.015uF to none.

    Enable is unconnected then connected to input.

    Comp network, (47pF 1000pF-30K), then (94pF 5000pF-10K) I'm using Webbench.

    Oh yah, Added a 0.1Uf to the input.

    It is not drawing any mA. It runs seconds on the 1000uF OS-Con.

    No change for any of these.

  • 0 in reply to Gregg Dewenter
    Prodigy 40 points
    Oh yah, Added a 0.1Uf to the input. It is not drawing any mA. It runs seconds on the 1000uF OS-Con.
  • 0 in reply to Gregg Dewenter
    TI__Guru**** 191445 points

    What is the output current?  If there is no load, you may encounter BOOT UVLO issues if Vin is not > Vout + 2.5 V or so.

  • 0 in reply to JohnTucker
    Prodigy 40 points
    Thanks John,
    The output current sits at a few uA most of the time. It is running a WiFi CC3200 MODULE which goes to very low power to high power for short bursts. Adding a 120 ohm load, input to 4.25V the output goes to 2V. A momentary load of 20 ohm pops the output into regulation. Removing the 120 ohm load or Vin below 4.2v it falls out of regulation. We will be running 4 AA batteries and need a solution with a enable (wakeup).
  • 0 in reply to Gregg Dewenter
    TI__Guru**** 191445 points

    See this application note.  It is applicable to TPS45231 as well.

    Let me know if you have additional questions.

  • 0 in reply to JohnTucker
    Prodigy 40 points
    Thank you John,
    I had been stuck on this for three days. Need to send proto boards to customer Monday.
    The application note is a great one. I have a working solution, 10mA load for now (not perfect but working) and will try a diode from input to Boot. With 4 AA batteries the 8V limit of the diode method is close but fine.

    Gregg