cupoftea
Advanced Member level 5
Hi,
After all these years , still, If you want to do an offline Synchronous Two Transistor Forward SMPS, or Offline synchronous Full Bridge SMPS, then you have pretty well one main chipset to choose from…
That’s the LTC3723-1 primary side controller which uses a pulse transformer to signal to the secondary side synchronous rectifier controller (ie LTC3901 (Full Bridge) or LTC3900 (Two tran forward).
These chips will set you back around 23 Euros.
They are pretty useless. LTC3900/LTC3901 both feature reverse current sensing……..that alone tells you they are worthless…….if you sense reverse current in a synchronous FET, then turn off that FET, then you have broken an inductive current and will have to quench an enormous amount of energy….otherwise suffer blown up circuitry.
Synchronous rectification should only ever be done in a way in which reversing current in the synchronous FET can never happen.
That is, you should either…
1….Sense the load current, and disable the synchronous FETs when load current gets below that level where the output inductor current goes discontinuous.
2….Sense the output inductor current, with a sensor just downstream of the output inductor, and disable the synchronous FETs whenever the current in the output inductor goes below 1A…..(ie well above zero Amps where it can then reverse.)
Synchronous FETs are probably the most significant factor towards SMPS efficiency that there is today….even better than resonant converters…and yet they are in a rotten pit of ruination and dereliction.
There are even “reactive” synchronous rectifier drivers which have no (direct) communication with the primary side……..they sense eg the secondary side switching node….they are useless….other posts on here have revealed their utter frailties……eg their needing of the fitting of “stray” inductors to cancel out the stray inductance of the synch rect cctry, so that they can try and not go wrong.
Why have synchronous rectifiers simply been left in the cess-pit?
LTC3901
LTC3900
LTC3723
After all these years , still, If you want to do an offline Synchronous Two Transistor Forward SMPS, or Offline synchronous Full Bridge SMPS, then you have pretty well one main chipset to choose from…
That’s the LTC3723-1 primary side controller which uses a pulse transformer to signal to the secondary side synchronous rectifier controller (ie LTC3901 (Full Bridge) or LTC3900 (Two tran forward).
These chips will set you back around 23 Euros.
They are pretty useless. LTC3900/LTC3901 both feature reverse current sensing……..that alone tells you they are worthless…….if you sense reverse current in a synchronous FET, then turn off that FET, then you have broken an inductive current and will have to quench an enormous amount of energy….otherwise suffer blown up circuitry.
Synchronous rectification should only ever be done in a way in which reversing current in the synchronous FET can never happen.
That is, you should either…
1….Sense the load current, and disable the synchronous FETs when load current gets below that level where the output inductor current goes discontinuous.
2….Sense the output inductor current, with a sensor just downstream of the output inductor, and disable the synchronous FETs whenever the current in the output inductor goes below 1A…..(ie well above zero Amps where it can then reverse.)
Synchronous FETs are probably the most significant factor towards SMPS efficiency that there is today….even better than resonant converters…and yet they are in a rotten pit of ruination and dereliction.
There are even “reactive” synchronous rectifier drivers which have no (direct) communication with the primary side……..they sense eg the secondary side switching node….they are useless….other posts on here have revealed their utter frailties……eg their needing of the fitting of “stray” inductors to cancel out the stray inductance of the synch rect cctry, so that they can try and not go wrong.
Why have synchronous rectifiers simply been left in the cess-pit?
LTC3901
LTC3900
LTC3723