LM25116: LM25116 Voltage Regulation

Hi Robin,

1st converter: Vin=25V-33V, Vout=24V. The maximum allowable switching frequency with the above condition is ~50kHz-60kHz, but you are trying to operate at 250kHZ which the device won't be able to regulate. Please use design tool properly to avoid such schematic issues: LM5116-5116WG-25116DESIGN-CALC Calculation tool | TI.com

Operations with low input voltage (~25V_vin_min) requires a high a very high duty cycle [ at Vout=24V] to maintain Vout regulation. The maximum possible duty cycle is limited by minimum HO off-time required for sampling DC inductor current. The output will fail to regulate if the following condition is not satisfied

Fsw<(Vout-Vin_min)/(Vin_min*580ns)

Regards

Arpita

Hi Arpita,

Thank you for your reply. I was wondering if you could point out where I can find the maximum operating frequency on the datasheet. When I was checking the datasheet, I only noticed that the switching frequency would be limited when VCCX is less than 6V but right now the VCCX is powered by a 12V buck converter so I am slightly confused why I could only operate at 50kHz-60kHz. Also, the 5V4 line is running at 500kHz and the 24V rail is at 1MHz. So im not sure where the 250kHz is coming from .

Also would it possible to continue this conversation in private such as by emails?Thanks

ARPITA NATH said:

Fsw<(Vout-Vin_min)/(Vin_min*580ns)

Als, i think the equations is incorrect. Vout is alwasy smaller than Vin. Having a negative frequency wont make too much sense

Also according to my calc, the absolute minimum Fsw would be 130kHz for my 24MHz buck converter asssuming the lowest input votage is 24V.

• Fsw <= 7.14MHz=Vout/ Vin_max * 100ns = 24V/(33.6V*100ns) 
• Fsw <= 133kHz=(Vin_min-Vout)/ Vin_min * 580ns = (26V-24V)/(24V*580ns)

Running at 500kHz shouldnt be a problem for the 5V4 buck:  

• Fsw <=1.61MHz=Vout/ Vin_max * 100ns =5.4V/(33.6V*100ns) 

• Fsw <= 1.36MHz=(Vin_min-Vout)/ Vin_min * 580ns = (26V-5.4V)/(24V*580ns) 

In addition, although the 5v4 buck converter is properly regulating the output voltage, its gate(HO pin19) voltage is behaving weird. 

With a 7.9V input voltage, the node seems okay with some ringing at the falling edge.

However, as soon as the input goes to 8V, the gate node has a completely different waveform.

At 13V the node goes wild but sill have a similar shape the the 8V waveform.

Keeping increasing the input voltage would not change the general shape from now on. It will only make the spike higher and narrower. The switching node also resembles a similar shape. Note that there is no gate snubber populated right now. What could cause such behavior and should i be concerned? Is there any way to resolve this issue? The output voltage seems totally fine. 

Hi Robin,

In datasheet, the Forced HO off-time and Minimum HO on-times are given. Let's understand that the switching frequency [for a given application case of Vin, Vout] has to be chosen such that these limitations does not hit. You can calculate the maximum possible switching frequency for a given Vin, Vout from these. 

However, you can directly use the excel tool which has been already shared to you. You will get proper explanation and equations in the excel itself for easier calculation.

As for 1st converter: Vin=25V, Vout=24V, D~0.96, at Fsw=1Mhz [current value set] gives a required off-time of ~40ns where as the forced HO off time iof the device is 580ns. Hope this makes it clear.

VCCX sets the maximum possible oscillator frequency. But while setting operating switching frequency, the above points have to be taken into account.

Regards

Arpita

Sorry for the Typo, it is clearly Fsw< (Vin_min-Vout)/ (Vin_min*580ns) 

Yes 2nd converter is fine 5.4V

For first converter:

Fsw <= 133kHz=(Vin_min-Vout)/ Vin_min * 580ns = (26V-24V)/(24V*580ns)

 Vin_min=25V, Vout=24V for your case

So the correct calculation should be fsw<(25-24)/(25*580ns) which comes out to be Fsw~60kHz. I have suggested to use 50kHz-60kHz for 1st converter.

Thanks for the suggestion. The ho_sw is looking a lot like a square wave but why does the SW mode looks like its having three states, which seems like its in dcm mode. i already put a 10k resistor at demb pin so it should be in full sync mode?

Can you please try changing the DEMB resistor value and recheck?

Also try using increasing load current

Regards

Arpita

Thank you very much again for your reply. I was wondering if you have an email address I could reach you at. I have a video I would like to share, but due to confidentiality and IP considerations, we would prefer not to post it publicly here. I truly appreciate your time and understanding, and I look forward to hearing from you.