LM5164: Shaky switching signal in LM5164

hi, 

I have use webench as below, get the different inductor and RC value, could you please try again. 

and I share a calculate tool here, please try .

snvc226b.zip

Elena

Hi Elena,

Thanks for your comments, I know this tool, and I did a calculation on this one, but the inductor should be at least 373uH, which is selected 68uH automatically. I think it doesn't work.

Do you know what is the reason for shaky switching signal? 

hi, Farhad,

This looks like the loop compensation is not good, please change R9 C12 C16 in your SCH , and try to have a loop test if you could .

suggest to refer to below, C16 is too large, try C16=  56pf, and I suggest to calculate Iout leave some buffer ,i use 100ma below.

1778.Copy of LM5163, LM5164 Quickstart Calculator rev3.xlsm

50mA example from webench, please verify if this test result good.

Elena

I actually updated like this circuit:

and again, same result (Shaky signal)

I also checked the circuit with these values, but the result is still the same as before.

L = [(48-5V)*5V] / [200kHz*0.5*1A*48V] = 45 uH (if k=50%), is it the inductor is too large in you calculate ?

Elena

I know this formula, but as I said before, the maximum output current which I want to use is less than 100mA (around 30mA)  therefore, if you consider 100mA the result becomes much higher.
Here you considered output current ( Io = 1A)

But I tried to decrease the resistor in the T-Network (Rr = R9 in my schematic) to 47K the switching signal become more stable (90% improved) but the shaky problem is still there.

hi, Farhad,

could you upload your layout  and sch, i will check the layout.

This is the 3D model of the product (a part of it) I can not put all the layout files because of the confidential things.

Just please note, the R14 is not placed.

and this is the layout of this section:

I hope it can help you to understand the build-up of the PCB.

Thank you for sharing this, I will check and give you a feed back later.

Elena

I did some more investigation on this issue and I found if I decrease the Rr resistance to around 18K the shaky signal becomes very low like this:

But of course, If I zoom into the one pulse I can still see vibration on the pulse:

The other thing is if you look at the signal there is an undershoot in the signal, 

Do you know what is the reason for this issue?

hi, Farhad,

please try to use Rr = 68K, and then try to increase Vout cap to 6.8uf or 10uf , 16V , please check if this will improve. 

Elena

I already checked these values, does not work, The resistor (Rr) should be around 18K and output capacitor around 1-2uF

hi,

how about increase the Vou cap, for inductor is 470uh, is it 4.7uf in your test, how about 10uf result is not much change ?

Elena

Hi Elena,

Increasing the output caps does not really impact on the stability, I checked different values: 1uF / 4.7uF/ 10uF / 22uF, and none of them had any impact on the shaky signal!

I will check within team to see if other people have an idea. will reply to you later.

Elena

Hi Farhad,

The PCB layout looks OK.

When you measure the waveform, the time Trigged point (X-axial) should be within the scope, but on your above waveform with 200ns/Div, the Trig'd  sign was on right hand and out of the scope range. it's trigged at -6.27us instead of 0--2us range. pls adjust the Trig delay time.  

B R

Andy