T
treez
Guest
The NCP4306 offers synchronous rectifier control for a flyback in CCM/DCM/BCM.
Do you believe that this IC looks a little dubious?
I would say that this dubiousness is also found in many other manufacturers’ similar offerings
All these type of Flyback sync rect controller ICs make me prefer just putting multiple diodes in parallel , and forgetting sync rects altogether…or, if the current is very high, use a two transistor forward in CCM, which is more conducive to sync rectifiers. (monitoring load current with a micro and disabling the sync rects when it gets toward DCM). A two transistor forward in CCM won’t suffer as bad reverse recovery as a flyback because the leakage inductance is permissibly higher in a two transistor forward.
Page 22 of the NCP4306 datasheet basically confesses that it only really works if you have an SMT package for the synchronous FET (low lead inductance). That may be OK in some cases….however, in most cases, synchronous rectification involves high currents and so you would want a TO220 type synchronous FET….one that can be conveniently fastened directly to a heatsink, which of course, you cant do with an SMT FET.
One may think that very low rds(on) SMT FETs could be used…but the NCP4306 datasheet advises against this, saying it may lead to noisy operation and malfunction. (page 20 of datasheet)
The NCP4306 does offer a “Minimum on time” feature so as to avoid premature turn off of the sync fet..however, on page 26 it confesses that the minimum on time may result in reversal of the secondary current and resultant high voltage spikes on the sync fet……not very appealing at all.
In all, would you agree that the NCP4306 looks a little dubious? Seriously, I think reading that datasheet I would wonder whether I was better off just putting three or four diodes in parallel to minimise flyback diode heating.
However, I do think that the NCP4306 could be used in a microcontroller managed scheme, whereby the flyback was operated in deep CCM and using the NCP4306’s trigger input via a pulse transformer from the primary side gate drive…and the micro would monitor the load current, and when it fell so low that DCM operation would be started, the micro would just disable the NCP4306. This would mean bad reverse recovery losses though….so again, the NCP4306 seems to lead us over thin ice, whatever way we choose to use it.
NCP4306
https://www.onsemi.com/pub/Collateral/NCP4306-D.PDF
Do you believe that this IC looks a little dubious?
I would say that this dubiousness is also found in many other manufacturers’ similar offerings
All these type of Flyback sync rect controller ICs make me prefer just putting multiple diodes in parallel , and forgetting sync rects altogether…or, if the current is very high, use a two transistor forward in CCM, which is more conducive to sync rectifiers. (monitoring load current with a micro and disabling the sync rects when it gets toward DCM). A two transistor forward in CCM won’t suffer as bad reverse recovery as a flyback because the leakage inductance is permissibly higher in a two transistor forward.
Page 22 of the NCP4306 datasheet basically confesses that it only really works if you have an SMT package for the synchronous FET (low lead inductance). That may be OK in some cases….however, in most cases, synchronous rectification involves high currents and so you would want a TO220 type synchronous FET….one that can be conveniently fastened directly to a heatsink, which of course, you cant do with an SMT FET.
One may think that very low rds(on) SMT FETs could be used…but the NCP4306 datasheet advises against this, saying it may lead to noisy operation and malfunction. (page 20 of datasheet)
The NCP4306 does offer a “Minimum on time” feature so as to avoid premature turn off of the sync fet..however, on page 26 it confesses that the minimum on time may result in reversal of the secondary current and resultant high voltage spikes on the sync fet……not very appealing at all.
In all, would you agree that the NCP4306 looks a little dubious? Seriously, I think reading that datasheet I would wonder whether I was better off just putting three or four diodes in parallel to minimise flyback diode heating.
However, I do think that the NCP4306 could be used in a microcontroller managed scheme, whereby the flyback was operated in deep CCM and using the NCP4306’s trigger input via a pulse transformer from the primary side gate drive…and the micro would monitor the load current, and when it fell so low that DCM operation would be started, the micro would just disable the NCP4306. This would mean bad reverse recovery losses though….so again, the NCP4306 seems to lead us over thin ice, whatever way we choose to use it.
NCP4306
https://www.onsemi.com/pub/Collateral/NCP4306-D.PDF