AnsweredAssumed Answered

LTC4364-2 produces noisy, undervoltage output

Question asked by bbone on Jun 5, 2018
Latest reply on Jun 5, 2018 by bbone
We're bringing up a board that uses the LTC4364-2 and are having some trouble. 
We're using this chip to provide some protection on a 44V battery line that feeds some sensitive electronics downstream. Our design closely follows the Typical Application shown on Page 1 of the Datasheet with only minor changes; we've set the UV and OV thresholds to be 36 and 58V respectively, and successfully simulated the design in LTSpice (attached). However, when we apply power, the surge protector does not allow power through. We're observing some pretty large noise on the input line and gate, shown below. We suspect that the chip is oscillating in and out of some fault condition, but the current and voltage thresholds are in spec. The input noise is perfectly correlated to the gate signal: 
(Yellow/1 is the gate signal for Q1, Blue/2 is the drain of the same. We're getting 70Vpp spiking on a 48V input signal)
Here are the gate/drain signals for Q2 - better but still significant. You can see that the DC output voltage from the surge protector is only about 9V. 
Zooming out, the sawtooth gate signal appears to be modulated onto a lower frequency sawtooth: 
This feeds into two LC filters and buck converters. When we bypass the surge protector, the board runs fine. What could be the cause of this issue? 
We've inspected the board and all components appear to be placed and assembled properly. The components have been verified for proper value. 
Any help would be greatly appreciated. 

Outcomes