LM22679: Voltage Drops With Load

Hi Petersen:

1. can you attached the waveform of PIN1 (SW) , with 100mA, 4A and 6A? need to have the waveform to do further judgement

2.can you share the PCB layout

3. if the schematic you show is all the input capacitance at input side, pls add a electrical cap with 100uF capacitance. in the same time, you can also add one more 100uF output capacitor to see any improvement.

Thanks

Thanks for your response Daniel.

Added additional 100uF electrolytic capacitors at both the input and the output.  No improvement noted.

Layout is not optimal.  Capacitors are not placed well and the component connections should be solid.

Have to patch these boards enough to get them working for now.  Won't be able to fix the layout for a month or two.  

Layout:

Vin comes from the backside and enters the LM22679 through 12 vias at pin 2.  Output capacitor ground lead is 25mm distant from the diode ground lead. Connection is through the ground plane (not shown). 

Switching voltage at 100mA load (12V input):

Switching voltage at 4A load (12V input):

Switching voltage at 6A load (12V input) [output current oscillates and there is a loud hum]:

Hi 

Can I know where you add the input capacitor?  it is better to close to  C69.

one possible reason is R85 GND connect to C69 GND, can you try to swtich 180deg C and connect to chip's GND?

Thanks

With C69 being so far out of the current path into the input pin (pin 2), I placed the new 100uF electrolytic cap on top of the LM22679 with the positive lead soldered to pin 2 and the ground connection being made at pin 4.

Daniel Li14 said:

one possible reason is R85 GND connect to C69 GND, can you try to swtich 180deg C and connect to chip's GND?

I'm not sure I understand this request.  Are you asking me to disconnect C69 from its ground plane and wire the ground connection to the LM22679's ground at pin 4?  Please clarify.  Thank you.

HI

yes, that's right 

Hi Daniel,

I disconnected C69 from its ground connection then used 20 AWG wire to wire both output capacitors directly to the LM22679 pin 4 lead.  Sorry to say, no appreciable improvement.  Vout still drops from 5V down to 4.5V over the 100mA to 4A output range.  

I was thinking to try and desolder the inductor, add a loop of wire, then get some scope shots of the current through the inductor. Would that be a worthwhile exercise here?  Or should I be looking at something else?  I'm still at a loss to understand the mechanism that leads to behavior like this.  Slow response I understand.  But this is no loop response.  Any ideas of what to look for are greatly appreciated. 

Thanks so much.

Hi

can you try to replace D33 by Schottky-type power diode, I checked SBR20M150D1Q , it is Super Barrier . 

from datasheet described below: 

A Schottky-type power diode is required for all LM22679 applications. Ultra-fast diodes are not recommended and may result in damage to the IC due to reverse recovery current transients

Thanks

Had to order some parts to try. Please leave this case open.  I'll report back when they arrive in about ten days.  Thanks!

Sure, you can also apply EVM board ,and move EVM board diode to your board.

Thanks

Replaced the diode with a 200V 15A Schotty (IXYS DSA15IM200UC-TRL).  No improvement.  

I ended up raising the output voltage to 5.44V which at full load (4A) drops to 5V.  This is good enough for now.