LM5148-Q1: Vout gets stuck at 1.9V (instead of 12V) upon a reboot unless the input and output capacitance is fully discharged.

Hello Ben,

please fill out the Design calculator with customer spec so we can check design against it.  

7737.LM5148-LM25148 Quickstart Calculator.xlsm

Thanks.

David.

Hi David, Thanks for the form, I've filled this out and attached it here. Thanks.

[Removed the unlocked design spec calculator]

Hello Ben,

The spreadsheet I sent you was password protected, the spreadsheet you sent back was not sure how this happened?  Please see attached new calculator that is "locked".  Please can you delete the above spreadsheet you sent me as unlocked spreadsheet as this is not to be shared publicly.  I ask you remove all unprotected copies from your PC.   Please refer to the terms of use in the spreadsheet.

Regarding your design, I derated the output capacitors to suit.  Please use the compensation components therein.  Also ensure your design as input capacitors to suit your needs as schematic above has none.  

LM5148-LM25148 Quickstart Calculator. e2e Ben T xlsm.xlsm

Hope this helps.

David.

Hi David, thanks a lot, I will try reducing the output capacitance. I do have input capacitance (as per the spreadsheet) but it's on a different schematic page. I'll test out reducing the output capacitance tomorrow and let you know how it goes. 

The lock may have got removed when I imported into google sheets as an FYI as I don't have an excel licences at the moment. Apologise for that, I have removed the copies on my end.

Thanks Ben,

Appreciate it very much.  let me know if you still need support.

David.

Ben,

The resistor from FB to VDDA for fixed 12V should be 49.9kΩ. Also, you can try it with a resistive divider from VOUT to FB, e.g. 100kΩ and 7.15kΩ (this is also convenient if you need to take a bode plot measurement).

PS: I don't see the input cap in your schematic.

--

Tim

Hi, I tried the fixes above. Reducing the input / output capacitance does improve the behaviour, but not enough to be classified as working. 

I've noticed that the issue seems to be mostly caused by the residual voltage on the input capacitor (~2V after turning off the supply due to the eFuse input blocking any discharge path below this voltage). I can fix this either by adding a 1k resistor in parallel with the input capacitor to ground, or I have noticed that pulling the high side gate driver high with a 100k resistor also fixes the problem. Does anyone have any idea why this helps fix the behaviour?

I've also noticed substantial ringing on the high side gate, let me know if you think this could be linked to the problem.

Hi Ben,

That is the DCM ring at light load. Tie the PFM pin low to obtain FPWM operation at constant Fsw.

Regards,

Tim

Connecting 100k from VIN to HO is not really advised.

Hi Tim, don't worry I wont leave the 100k, I was just playing with what would help the issue to see if it gave any clues to the route of the problem. 

Okay so the DCM ring is not the issue, so the problem remains, at low load, I cannot restart the chip for 5 seconds as the input voltage remains at ~2V (slowly discharging) for that time.

Have you got a completed quickstart file so I can understand the parameters involved here?