LM5156: LM5156

Hello, 

Thanks for reaching out. Unfortunately it is not allowed to short the CS pin to ground as this device is operating in peak current mode control and therefore sensing the current is fundamentally important. The compensation components cannot be left out and are as well of critical importance to the correct behavior of the device. You might take a look at our EVM at this link: https://www.ti.com/tool/LM5156EVM-BST  and use the Excel Quickstart calculator for your design: https://www.ti.com/lit/zip/snvc224 

Let me know if you need further help. 

Kind regards,
EM

Hi EM,

Thanks for your response. Please see attached here the quick start calculator file. It would be of great help if you could review and let me know your thoughts. Please see that the diode I have selected for my application is V40DL45, inductor as SRP1510CA-4R7M and for MOSFET I have selected two parallel CSD18540Q5B. Also, since under voltage feature isn't needed I have shorted the 'UVLO' line to 'BIAS'.

Also, I am a bit of a novice with switch mode converters and so would really appreciate if you could share with some good reference material (videos &/or application notes) on compensation why it is needed and what can go wrong if not handled correctly.

LM5155_56_Quickstart_Calculator_for_Boost_Converter_Design_V1_1_1.xlsx

-Rishi

Hi EM,

Thanks for your response. Please see attached here the quick start calculator file. It would be of great help if you could review and let me know your thoughts. Please see that the diode I have selected for my application is V40DL45, inductor as SRP1510CA-4R7M and for MOSFET I have selected two parallel CSD18540Q5B. Also, since under voltage feature isn't needed I have shorted the 'UVLO' line to 'BIAS'.

Also, I am a bit of a novice with switch mode converters and so would really appreciate if you could share with some good reference material (videos &/or application notes) on compensation why it is needed and what can go wrong if not handled correctly.

2818.LM5155_56_Quickstart_Calculator_for_Boost_Converter_Design_V1_1_1.xlsx

-Rishi

Hello Rishi, 

Thank you for your feedback. Let me share my inputs here: 

• With 15 A of load current and 24Vout, you end up having a very high output power (~360W. which is quite a bit higher than 240W). I see that you selected a diode with a high current rating but if you look at figure 2 in the DS 'Average Power loss VS average forward current' you will see that the diode will need to dissipate quite some power. I would recommend instead a synchronous device (like the LM5123), in this case you can parallel MOSFETs both on the low side and high side and therefore have better performance. The LM5123 also has an Excel Calculator and an EVM that you can take as example for your design (https://www.ti.com/tool/LM5123EVM-BST and https://www.ti.com/lit/zip/slvrbj1 ). 
• If the compensation is not properly design, the converter is not able to regulate properly the output voltage (it might also oscillates and the converter might not react well to load step and this might also cause damage to the MOSFET for example), this is why a good rule of thumb would be to have ~60 degrees of phase margin and >10dB of gain margin. We have this video that explains in general the different compensation techniques: https://training.ti.com/switch-mode-power-converter-compensation-made-easy (and this is the white paper: https://www.ti.com/seclit/ml/slup340/slup340.pdf ). At this link you can, in general, look for many different videos and trainings on different topics: https://training.ti.com/ . 
  
  

Please let me know if you need further support. 

Kind regards,
EM 

Hi EM,

Thanks for your response. Changing my design to LM5123 might not be possible at this stage. I believe I shared the wrong calculation file with you earlier. Please find attached revised one. Also, please see that my application does not require to deliver continuous power and that the load switches at a duty cycle of 1:1 with 10 seconds ON time. After your suggestion on excessive power loss in the diode I have parallel-ed 4 SBR10E45P5 diodes which brings down the forward voltage drop to about 350mV (4A per diode). I hope this would work. Kindly provide your thoughts and that if the selected compensation network would work for me.

Please see if you could also share with me a similar calculator for a boost converter designed using LM3478.

-Rishi

5140.LM5155_56_Quickstart_Calculator_for_Boost_Converter_Design_V1_1_1.xlsx

Hello, 

Thanks for your reply. Let me share here some comments from my side: 

• About the LM3478: for the compensation you might want to use this file https://www.ti.com/lit/pdf/snva067 together with the guidelines of the datasheet that you can find at page 16. The Excel calculator is not available unfortunately for this device but the guidance of the datasheet should be as good as the Excel file. 
• About the diode: paralleling diodes should be avoided due to the temperature dependence of their forward voltage (which will cause anyway one of the diodes to carry most of the current). As you say the converter is not continuously running, it might be ok but please check this in the lab. 
• About the Excel file you attached for the LM5156: it looks fine. but I would lower the value of the selected sense resistor as suggested in the file, otherwise you might hit current limit during operation. 
• One remark from my side: the current in this system is really high so please take care on the thermal design for this PCB. You might keep in mind for future designs that there is the possibility for synchronizing two devices so that they can share the load current like in this design https://www.ti.com/tool/PMP40679#parameters (the LM5155 is a very similar device so it can be takes as reference) 

Please let me know if you have further questions. 

Kind regards,
EM