While you wait for parts, have you thought of simulating the circuit? TI should have tools for doing this, and of course LTSpice if there is a model for the 3845 (someone probably has made one). Great way to evaluate what is going on without the bother of actually letting the smoke out of parts 
The general theory of a current mode converter is that an analog of the inductor current (the voltage across the sense resistor) is compared to the output of the error amp to turn the switch off. The switch is turned on at the start of each cycle. This can be seen in the block diagram posted by Blujay. The point here is that the voltage arcross the current sense resistor must be large enough, say a hundred millivolts or so, for the current sense comparator to have a decent signal when the converter is working with a nominal load. If this condition is not met, the converter may still appear to work, even regulate voltage to a degree, but it is far from optimal. Most would say it is "unstable" as the duty cycle varies widely from cylcle to cycle.
Last edited:
with 50% pedestal current I guess switch current 9 amps peak or so-- to step 12V to 24V, 150 watts
That is only 90mV signal against 1V comparator range -- certainly no transient or overload protection and would be prone to noise and instability.
It may not even be operating in current mode and the fact that comp is type 1 single pole really invites poor phase margin and even oscillation
That is only 90mV signal against 1V comparator range -- certainly no transient or overload protection and would be prone to noise and instability.
It may not even be operating in current mode and the fact that comp is type 1 single pole really invites poor phase margin and even oscillation