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.

LM5161: My converter 24V-100V Vin > 15 V Vout doesn`t work. Any ideas?

Part Number: LM5161
Other Parts Discussed in Thread: LM5160

Hi. I`m making device, where i need to have 15Vout from 24 - 100 Vin. I decided to use LM5161. I tried many designes on TI`s Webench and i ended up at this one: But it doesn`t work. C1, C2, C4 are 1uF foil capacitors for 250V max. C3 has 16V Vmax, but there should be max only 2V on the SS pin. C5 is ceramic capacitor with Umax 200V. C6 is 1uF ceramic capacitor with Umax 16V. So there also shouldn`t be any problems. 

I`ve done some simple measurements on my osciloscope, and i found out, that on pin SS isn`t 2V, but approximately 0.2V. On VCC is 7.38V, on BST 9.8V RMS. On FB 0.2V. Do you have any ideas, what i`ve done wrong? Where i`m making mistake? I`ve read datasheet many times, but i can`t find any mistake. Thanks in advance. Jachym

  • Hi do you have any scope capture? Is the part switching at all?
  • No. It doesnt't. I think that's cause on SS pin wasn't enough voltage (at least 2V, if i'm not wrong). But i can capture some, if you want.
  • Hello,

    Input capacitor should be ceramic.
    Please remove the TVS from the output as it might be engaging early with the regulation point slightly higher than the breakdown of the TVS?
    Make sure the layout is as per datasheet recommendations.
    Also can you share the datasheet for the inductor used? Want to check inductance value, and Isat rating. Also suggest using a Shielded inductor.
    Can you also share the part number of Couts used? Need to check Cout and Cout ESR to ensure correct ripple voltage is being supplied to FB.
    I would make sure connections on your board are as per schematic, not unintended shorts etc, or mis-installed components?
    Can you share your design specification with us? Iout?

    Hope this helps?
  • Why there should be ceramic instead of foil one? If i use enough voltage, then there will be almost no difference, won't be?

    The TVS i didn't even assembly at all, because i did scoping on your website again, and it seems that there is almost 20V at start. Which should damage TVS and maybe also this chip.

    Here is datasheet for TALEMA:

    It has 150uH inductance and Imax 3A. Unfortunatelly in our country i didn't find any shielded inductors with 130uH and such a big current.

    This datasheet is output 47uF capacitor:

    And this one is 330n Cout:

    All max voltages on capacitors are at least two times bigger, than needed. I do this allways.

    Thanks in advance. Jachym

  • Hello,



    My concern with the Foil input capacitor is that its a leaded device, with higher ESR than a ceramic, foil specify DF (Tan d), when you calculate ESR, typically its very high in comparison to a ceramic (mohms). Also Foil capacitors are usually leaded further increasing its ESL. its important to have low ESR and ESL for the input capacitor due to the high rates of di/dt see on the input capacitor of a Buck.

    Also Ceramic capacitors are quite robust when it comes to voltage ratings. i.e. Typically for an absolute max input voltage of 100V, a 100V ceramic can be used? Of course, you will need to verify that the selected input capacitor is right for your application under all conditions.


    Can we please take a look at your layout?

    Thanks.

    David.
  • Of course. Here it is. C9 is part of 5V voltage source, with LM78M05. But it is electrolytic 470uF capacitor with Ubr 25V. So that also shouldn`t cause my problem.
    I tried to put the capacitors as close to the LM5161 as was possible, but fe. C6 is pretty far. Can this cause the problem?

  • Hello,

    Unfortunately, C6 needs to be close to the IC as well as the Cin Caps.  As mentioned, use ceramics to get very close to pins.  If you violate these rules, you could well have issues due to layout?

    Please see the attached recommendations for laying out a buck using the LM5160.

    Hope this helps?

    LM5160 Floor Plan.pptx

    David.

  • I made some changes, i've ordered new ceramic 0603s (my old 1206 were too big to solder directly on pins), i've resoldered the caps directly to the LM5161's pins, but the LM5161 is still not switching. On SS isn't 2, but 0.36V, on FB 0.48V and on out there is only 3.6V! There must be another problem... I though, it can be caused by corrupted LM, so i ordered on Monday two new, have resoldered it yesterday, but suprisingly the voltages are absolutely same... I really don't know, what else i can do...

    I'm designing voltage sources for a quite ling time, but i didn't ever met with something similar...

    I will greatly appreciate if it can be solved.

    Best regards, Jachym Hurtik

  • Hello JachyHm,

    Can you take a scope shot of Vswitch and inductor current [lifting Lbuck and using a current probe] at the instant you turn on?

    This will help determine if the part is trying to start and shutting down for some reason, or the device is not trying to start at all?

    Thanks.

    David.
  • Hi! And sorry for delay. i didn`t have access to oscilocope and even more time to do measurements. Here you have scopes:

    First seven are Vswitch, with some cursors placed. It seems to LM5161 is trying to switch, but then fall down to 0 (-0.254mV) after some time to 2.6V and then again to switching. Can it be caused by BST cap? That the MOSes don`t have enough power to switch? And the last one (8th) is scope after inductor. There is stable 2.6 - 2.8V. Unfortunatelly i don`t own current probe, so i did only voltage measurements.

    Best regards Jáchym.

  • Here are some more with higher sample rate.
    This behind the inductor:
     And this Vsw:

  • Hello Jachym,

    It looks to me that the voltage on the Switch node is ringing due to DCM mode, this is implemented when you tie FPWM to AGnd. the frequency of switch will reduce significantly in deep DCM mode. If you load the output, the ringing should gradually diminish and the frequency will increase to a frequency set by Ron. Can you confirm this is the case?

    Thanks.

    David.
  • Hi. I didn`t exactly understand, what did you mean. I`m not so good at professional English speech. I can only confirm, that i really have FPWM connected to AGND, but i did it, because it was used in WEBENCH drawing. Also the Ron resistor i used that one, which was calculated by WEBENCH. I calculated now the frequency by equation in datasheet, and it equals to 305kHz, which seems to be okay. Can you please tell me, what exactly i should do, to get this thing working? I don`t have any experience with exactly LM5161, but it seems simillar to lot of another voltage sources, which i made with other circuits and there was never simillar problem.
    Thank you in advance, Jachym.
  • Hello Jachym,

    Please see attached drawing.  Can you confirm this is what you are seeing on your scope?  I am not sure your scope is set up correctly to measure the Switch node accurately but the attached drawing shows you what I expect to see as a result of DCM at light loads.  One way to verify, is to increase the load, the ringing on the switch node should reduce and the Switch node should converge to a stable switching frequency set by Ron.

    Also, you may not get exactly the waveform I am showing here, because the device will go into a Pulse skipping mode (variable frequency), which looks like what you are seeing?  Again, increasing the load and seeing the Switch node converge to a periodic switching (Fixed Frequency) is a way to verify things are normal?

    Another thing you could try is tie FPWM to VCC and implement ripple injection if you prefer fixed frequency operation without PFM mode.  Again, the datasheet goes into detail about how to enable FPWM and also how to implement ripple injection.

    Further I went back over your post and noted that your FB is low and your SS voltage is low.  We need to be able to measure the inductor current to rule out the converter is not hitting current limit?  If you do not have a current probe, then I suggest using a low ohm series between the buck inductor and output capacitor. 

    Also I would like to know what the voltage scale is on all your waveforms you have taken as there are no details given?

    Thanks.

    David.

    3730.DCM Ringing Doc1.docx

  • Hi. Sorry for delay, i had some important work to do.

    I put behind inductor 220Ohm resistor between Vout and GND, so there should be aproximatelly 1A through it. The frequency of pulses changed, but the voltage (Vout) remains absolutelly same.

    Scope's voltage scale are in top right corner. Respectively there are voltages on cursors. They aren't multiplied, or divided any, so they are as there shown.

    Vout gives still 3.2 Volts instead of 16.

  • Hello JachyHm,

    As previously mentioned, I thin the layout is not ideal. You should be using ceramic capacitors and not foil capacitors at the input and they should be very close to the device. Please refer to the layout recommendations I posted previously and follow closely the advice therein for a guide on how to layout this device.

    I also recommend using the components called out in WeBench, in particular the inductor, ideally this should be a shielded type.

    Hope this helps?

    David.

  • As i said somewhere before, i removed foil input capacitors and replaced them with 0403 ceramic, which are soldered directly to chip's legs. But nothing changed. Scopes are absolutely identical...

  • Hello JachyHM,

    I re-reviewed your inputs and I would like to focus on the fact that FB/SS is low. I did get side tracked on the layout, and this still maybe the cause of the issue? if SS is low, I would look in to current limit event.
    you will need to measure the indutor current in order to confirm. Also I re-reviewed your layout and the VCC cap and Cboot should be moved closer to the pin to see if this helps. Also I would double check ripple voltage at the feedback node that it is in phase with the inductor current (peaks happen at the same time) and that the peak to peak ripple voltage is ~50mV.

    Hope this helps?

    David.
  • JachyHM,

    Can you please ensure powerpad connection underneath the IC is connected to GND ? I have seen few failures related to the same . How are you soldering the Pad to GND connection .

    I will recommend to use an Evaluation module to evaluate your design from Webench .


    ---Ambreesh