LTC3649 datasheet circuit fails when powered with load

Hello,

As Dong Mentioned, it would be much easier to debug this circuit by testing the LTC3649 by itself.  To do this, Tie the LT3086 SHDN\ pin to GND.  Then, set R8 to various different values to set various output voltages between 0V and 28V.  Look for a stable constant frequency and constant duty cycle on the SW switch node at the various voltage setpoints.  Try adding your load(s) across the LTC3649 output capacitors (C1, etc).

The schematic doesn't show any input capacitors.  Do you have any?  Is there at least 2uF of quality ceramic capacitors close to the Vin and PGND pins of the LTC3649?  Is there additional bulk cap present to help prevent oscillation on the input?  If not, this could be a contributing factor.

Make sure that the exposed pad of the IC is soldered to the board.  As mentioned on page 2 of the datasheet, this is a requirement.

If all of this looks good, try integrating the LT3086, but start with a TRACK voltage of 0V and slowly increase the voltage as you monitor the SW node, Vout of the LTC3649 and Vout of the LT3086.  Look for abnormal behavior as you increase the TRACK voltage. Introducing the LT3086 in constant current mode with the LT3649 setpoint wrapped around it might be causing stability issues.  The ITH components might not be optimized for this condition.

Let us  know what you see with all of this.

Matt

Ok, thanks. After extensive trials as per instructions above - it turnbed to be the input cap was too small, it was some ceramic (not sure about value now), but increasing it solved the issue.  Thansk a lot.

This is good to hear.  You may need additional bulk cap as well to help prevent resonance on the input.

This is good to hear.  You may need additional bulk cap as well to help prevent resonance on the input.