@gsuberland I'm completely unexperienced with analog circuit, so excuse me for asking, but am I getting this right? Q1 is limiting current at the start due to the voltage divider of R1/R4/Q4, and in the background R34 let's C44/C45 slowly charge up, which in turn will push Q4 (n-type) to active mode, which then lowers the gate voltage of Q1 (p-type) pushing it to active mode and letting current flow freely? Looks interesting and the notes seem very helpful to me!
@thunfisch yup, exactly. Q1 acts as a linear element, with its Rds being high initially (off) and falling over time as its Vgs grows due to C1/R1/R4. without the Q4 arrangement, Q1's Vgs tops out at about -2.4V, which leaves it barely switched on, meaning it still has pretty high Rds, so when the load turns on it dissipates a bunch of power unnecessarily. so C44/C45/R34/R35 does the same kind of thing, slowly bringing Q4's gate voltage up, eventually switching on and pulling Q1's Vgs to -12V.
@gsuberland Thanks for confirming! My understanding was a bit off because I did not really take C1 and R33 into account but now it makes much more sense.
@gsuberland Oh this is great! Thanks. Much cooler than the old SPICE stuff that I still remember :) I can see a small spike in current once Q4 pulls down Vgs on Q1 fully. I guess that's short enough to not trip the fuse? I'm wondering where that current flows, why this happens. Does not seem obvious to me 🤔
@thunfisch hmm, I don't see it here. did you have the load switch closed during inrush? (it should be open until after everything is done charging; the load is remotely turned on)
@gsuberland And if I got this correct so far: the Zener diodes are for protection mostly, and R35 to discharge C44/C45 after shutdown to reset the circuit?
@thunfisch yup! the resistors should result in Vgs never exceeding ±12V anyway but the zeners ensure that the gates don't get overvolted if the rail overvolts or a component fails. plus, as you said, they do double duty as fast resets on the caps in case of power disconnection, so the inrush circuit doesn't get nullified if there's an intermittent connection upstream.
@petrillic the 2N7002 is actually kinda terrible here due to its wide Vgs(th) range, but it's the best option out of the parts in JLC's basic catalogue due to its Vgs(max) of ±20V. (the AO3400A is ±12V which is too low)
ideally I'd want a FET with a tighter threshold range, to have more predictable timing, but luckily here a 40-150ms range is fine.
Add comment