LM5116: SW signal pulse width is varying.

Hello Lesheng 

1. 18uH to 22uH is good to get the typical recommended RR (Ripple Ratio 25~-35%). 

2. I'd like to recommend using 4 x 2.2uF capacitors similar with EVM component selection. 

3. Please send me the completed Quick Start https://www.ti.com/lit/zip/snvu051  after changing the inductor value.

4. Yes, it is possible. Please locate the Cramp close to the IC and also please connect the ground connection of the Cramp 

-EL

Thanks for quick response.

Attached is the quick start I updated design inputs:

4064.LM5116_quickstart.xls

Do you think what's wrong with the design that caused the SW waveform is abnormal, would you give some advice how to fix this problem?

Current inductor is 12uH, compared with your recommended, it is smaller, smaller inductor value will lead to bigger ripple current, right? we also found inductor is very hot even under 1A current load, If I increase the inductor value to decrease the ripple current, does this help to decrease the power consumption of inductor and it can help to solve the inductor hot issue?

if I don't change the inductor value keep it 12 uH, what can I do to improve the SW signal?

Thanks

Another question: I found that I input the same design parameters into quickstart.xls and webench power designer, they generate different design, lots of components value are different, for example, in Webench power designer, I input Max peak to peak inductor ripple from 20% to 60%, it always recommended the same value inductor, While in QuickStart.xls, if I changed the value of inductor ripple, the recommended inductor value will be changed. 

Which tool should I trust?

Hello Lesheng

• Thanks for the Quick Start. Here are what I found. 1) L1 is small. Please use 22uH 2) C61 is too big. The recommendation is 270nF 3) Need 100kohm Rramp and 300pF Cramp if 12uH. 4) Crossover frequency is too high. please use the recommended Rcomp, Ccomp, and Chf values 5) There is a possibility that the VCC regulator is current limited. Please select lower Qg MOSFETs.      
• I guess you will be able to see an improvement after correcting 1)- 5) items above.
• In general, the power loss of the inductor increases by increasing the inductance because DCR is proportional with the inductance. 
• If the inductor is hotter than your expectation at light load 1) Check copper loss using inductor manufacturer's tool 2) Monitor inductor current and check if saturated 
• Then, correct only 2), 3), 4), 5) items and check if any improvement. 
• Please use Quick Start. 

-EL

Hi,

Crossover frequency is too high. please use the recommended Rcomp, Ccomp, and Chf values, do you mean schematic I pasted has Crossover frequency issue? do you mean I can use QuickStart recommended components value to replace the current schematic components values? but the inductor value is different between two, in QuickStart, if I changed the inductor value, the Rcomp/Ccomp/Chf will be changed accordingly.

Hi Eric,

I would like to update the experiment results for you reference:

1, First I changed the C61 from 1uF to 270nF, then tested the SW signal seems no improvement at all.

2,Then I changed the Ramp/Cramp to 100Kohm/300pf and keeped C61 270nF, then tested the SW signal seems there is a improvement, captured waveform as below: the SW signal at 0.5A current load now looks regular, it always a wide pulse following a narrow pulse, before changing, the pulse width is varying without regular, but it is still not a fixed duty cycle pulse.

When I increased the current load to 3A, the SW signal looks better than 0.5A, waveform looks like below picture:

I continued to increase the current load to 3.8A, the SW signal looks better than 3A, waveform looks like below picture:

I continued to increase the current load to 4A, the voltage output was shutdown and SW signal is disappeared, our design goal of current load is 5A.

3, then I increased the Cramp to 420pF, this value is very close to Cramp value recommended by Quickstart, the SW  signal has no improvement.

4, Then I changed the Ccomp/Rcomp/Chf from schematic value 220pF/232K/22pf to 820pF/82K/22pF(Recommended by QuickStart), then tested SW signal, seems there is no improvement at all.

Because We don't have 22 uH inductor, I didn't do inductor changing experiment, all the experiments were done based on 12uH inductor.

Do you have any comments regarding the experiment results?

I am confused by your saying" the crossover frequency in the Quick Start which you posted is too high", how can I improve this? because all the component value was recommended by tool, not by my inputs?

Thanks 

In QuickStart, for step 4, output capacitor, how should I do inputs for this item? for example, in my design, there are four output capacitors, they are 2 ceramic 2.2uf capacitor, 2 220uf alminium electrolytic capacitors, how should I do inputs for Cout1, Cout2?

Thanks

How can I know the crossover frequency is high or low? how can I improve this? 

Thanks

Hi Leshend 

• Users select the the target crossover frequency. Please check your input at E63.
• Enter the de-rated capacitance value of 2 x 220uF at E24. Enter the de-rated capacitance of 2 x 2.2uF at E25
• Quick Start recommends RCOMP=104kΩ and CCOMP =611pF while you are using RCOMP=232kΩ and CCOMP = 220pF. The crossover frequency increases by increasing RCOMP. Please use the RCOMP, CCOMP values recommended by Quick Start 

-EL

Hi Eric,

Thanks for you explanation.

For E63 input, there is a reminding says"enter a target closed loop bandwidth for switching regulator, the bandwidth should be at least one order of magnitude lower than the switching frequency", I am not sure my understanding is right or wrong? for my design case, the switching frequency is 400KHZ, what number should I do inputs? should I input a frequency less than 40KHz? what a proper number should I input?

In QuickStart, for frequency compensation graph, from these pictures, How can I know my design is a good design, the loop is stable?

Regards

Hi Lesheng 

• Yes, please enter the target crossover frequency which is less than 40kHz. 40kHz is the maximum crossover frequency you can achieve. I usually enter 1/20 of the Fsw (which is 20kHz in your case) and then gradually increase the crossover frequency if required.  
• In general, engineers target PM>60º, GM>15dB and Fcross ~ Fsw/20.

-EL

Thanks for choosing TI. Please feel free to contact us if you have any question.
-EL