Post Go back to editing

LTC4368 - Working with a single MOSFET but Not working with dual mosfets

Category: Hardware
Product Number: ltc4368-2

Hi all, I am using the following setup around LTC4368-2. Mosfets used are SSM6K809R,LXHF (not available in LTSpice). 

In simulation the setup works correctly. In experimental, LTC4368 takes a random number of attempts before achieving connection. 
Most of the times, a small duty appears in Gate, with nFault being asserted high when the gate is applied, and nFault being asserted low after the very short turn on moment indicated. However, i dont see triangles in retry when nFault goes low and readily (<<1200ms) gate is again applied and device turns on succesfully. 

nSHDN is clean and dips a little, but stays way above 20V (not an UV).Therefore i dont think an OV/UV is causing the trip. Why did the device assert a fault and retry a connection before performing the retry cool-down? Any ideas?

Other times, device performs several attempts before sucessfully connecting. In those cases where > 2 attempts are done, retry does have triangles that meet the expected ~1200ms retry period. 

To see if inrush was causing the trip: 

  1. Removed M2 to place a cable loop to measure inrush current. Result >> Setup worked flawlessly every time (!) and inrush current (peak) is ~900mA. 
  2. Re-placed M2 >> random startup attempts appear again.

Therefore, i dont understand the n = 2 cases where retry is not meeting expected cool-down cycles, and I dont see why substituting M2 for a loop for a current clamp to measure inrush is causing the problem to go away... I could be that having M2 causes sharp inrush currents that are not measure when it is substituted by a loop?

In this capture, you can see the 24Vin being cycled, with : 

a) first failed attempt generates a retry cycle

b) second failed attempt is followed readily by a third retry attempt , this thrid one being succesful. Retry timer is not respected in this case

Any debugging ideas?