LM22672 Output noise

Please provide your schematic and a picture of your layout.

Also, what is your load current?

Measurements of your input and output voltage and scope shots of the switch current would also be helpful.

Your schematic values look reasonable. 

Be sure that the inductor is rated for the load current.

Does the output voltage collapse when you apply your load or is it unstable?

You will need to supply some measurements and scope shots of the switch waveform.

Also, the layout is a very important part of the design; please supply some shots of the layout.

Hi,

You can try a couple different stuff:

1) For the device downstream if there is an enable pin on the LDO, you can set a resistor divider such that the device downstream does not pull more current before the 3.3V rail comes up.

2) You can also try to put a 0.47uF capacitor on the SS pin to make the LM22672 soft start function. This way, the output will ramp up slower. See if there is changes on the waveform. Whether the distortion appears or not

Thanks

-Arief

Do you mean FB1 from my schematic?  With my LPF installed to the enable pin of the LDO, as well as adding a 1uF cap to the soft start pin......This is what I get as a waveform.  As you can see the voltage on the 3.3 rail drops to about 2.5 VDC as the 2.5 rail is coming up.  This is too low for my SFP module, and as such is probably throwing it into a strange mode.

Yes that is correct

It does disconnect the LDO............here is the schematic. I'm not sure what you mean by the second part; when I power up the board the voltage dip as in the scope image earlier happens. So when the dip on the 3.3 rail happens, the 2.5 rail is powering up. The 3.3 rail is fully charged before the dip.

I can't use a current probe..............they are in use as of the moment, however at the limit being 1.7-1.8A, I cannot forsee this being reached.  Even with an inrush current I do not have enough load to draw that current.

With the Vout pin removed, the voltage rise on the output from the switcher looks perfect.  However, this is expected, since there is no load on the system besides the SFP, and the LDO is whats giving me the draw that is affecting the rails.  I have changed the input to the switcher from 5-20V and it still performs the same.

I would like someone to verify that my input and output capacitance as well as bypass on the LDO and voltage bias on the switcher look acceptable.

Here is the schematic of how I have the circuit laid out. 

Yes I do. Here is the layout the rest of the board to the right is not necessary for viewing.

I added it to the same thread as the first layout picture.

Hi,

I was able to see the layers now. It seems like it looks ok, except the DAP connection to GND, i would put more via in there. 

Can you check if the GND connection from the diode D1 to PAD is ok? I think the pad is only connected to bottom of C11.  You can see how we normally routed this DAP on the EVM

Also you mentioned about SD signal. I am not sure what is 525i and 650i. But i notice enable pin is floating. Could you force a 2V supply in here? The abs max for this pin is 6V

Thanks

-Arief

Any luck on the testing? I check your schematic and layout and it seems fine. 

The inductor that you choose is this part# correct? IHLP2525CZER220M1A

The diode seems to be sized correctly too. 

Do you see any dip on the VIN when this happen? Also does the SW node skipping cycle? 

Thanks

No luck yet, I have swapped out the power inductor for a 8.2uH and increased the capacitance to accommodate that change, in hopes that the current becomes more readily available.  The part number you listed is correct, and out of all this testing the only thing I haven't seen that agrees with the data sheet is that when I enable the soft start function on the switcher, it should start up in about 26ms, with the 1uF cap I'm using, however it is starting up in about 1.2ms.  So I need to figure out how to get that to go to 26ms.

Hi,

Yes you are correct, it should be starting up in 26 ms.
There is also a component calculator on the website for the loop response just in case you want to play around with the inductor and output capacitor.

Also, make sure that the LM22672 is operating at 500kHz frequency with the RT pin open