TPS 54560 Vout dropping down to zero on loading

Dear John,

Thank you very much for your prompt reply. I have made the suggested change to the circuit but the problem in persistent. Can you help any further?

Few doubts/queries I have:

1) How do you check if clock is working fine and switching is actually taking place correctly?

2) What is the exact use of COMP terminal? Is it used for overcurrent limiting? If yes, what is the guiding relation?

3) How should the circuit be loaded? Is there any particular way of doing so? Currently I'm starting buck converter unloaded and then loading it with 10 ohm resistor in one step.

Thanks in advance

You hav set the switching frequency using a resistor from RT to GND using equation 6 or 7. the output of the SW pin should be a pulse train at that frequency of amplitude Vin and duty cycle of Vo/Vin. It only applies under load. If there is a light load condition, then you will be in eco-mode and the resulting apparent frequency may be quite low. Additionally the TPS54560 has frequency foldback. If the voltage at the VFB pin is lower than the reference voltage, the frequency may fold back. See figure 14.

COMP is the output of the internal error amplifier. Connecting series R and C along with a parallel C sets the frequency compensation. It s important to get proper compensation for a stable circuit. COMP pin voltage is proportional to the peak switch current. The maximum COMP clamp voltage corresponds to the current limit.

Dear John,

Thank you very much for your valuable suggestions.
I have made the suggested changes (connecting FB resistors properly and connecting appropriate RC-C across comp terminal) in the circuit. It is working good!
A new problem has come up. The IC is getting overheated and producing efficiency of mere 71%.
Test conditions:
f = 300kHz
Vin = 30V
Iin = 0.94A
Vout = 12V
Rout = 7.2ohm
Therefore,
Pin = 28.2W
Pout = 20W
efficiency = 71%

I would expect efficiency to be well over 90%. I may need to look at your pcb layout.

Please find attached my eagle schematic and board files (and their pdfs). There are total 5 files.

Dear John,

Were you able to figure out something?

Is there any additional information you need?

The layout is certainly less than ideal. The circulating current path is relatively long and made of etch traces rather than copper fills. That may be contributing to your problem. I suggest you study the layout of the EVM design. What you could do is order an EVM, then modify it to your component values. See if the performance improves.