-
Low power loss full-bridge rectifier after transformer 230VAC 50Hz with mosfets
Posted by eneuro on 28 Jul, 2014 20:13 -
Hello,
Just looking for way to minimize power loses on transformer secondary full-bridge rectifier.
I do not know if someone tried something like this:
Idea is to add in pararell to small full-bridge rectifier diodes (up to 5A) ... mosfet switches with optoisolated gate drivers and simply detect higher currents than 1A using Hall sensors and switch ON mosfets in pararell with conducting diode to have voltage drop not 0.5V-0.7V, but.. in the case of IRFZ44N mosfet with RDSON 0.018 Ohm @ 25*C and current 25A we have about 11 W loses in mosfet switch, while diode with 0.7V voltage drop has about 17 W lose.
Difference is much bigger when we use not 1 mosfet but much more depending on desired acceptable power loses, while when 2 mosfets are used than those loses in one rectifier switch will be about 5W, so in whole bridge about 10W, while when only diodes used and 2 of them conducting about 35 W and we can do nothing with this.
I do not saw such solution so far, maybe I've made simple mistake, but simulation shows it might work, while mosfets with internal diodes looks like can of couse rectify AC voltage and when synhronized switched on it looks like I should have in practice power loses limited only by number of pararell mosfets used
What do You think about adding or replacing diodes in full-bridge rectifier with mosfets with internal diodes?
Maybe no fast Shottky diodes needed, while it is only 50Hz AC current rectified, but I'm interested in higher rectified DC power above 250W and maybe at lower voltages around 14V in 3 phase full-bridge rectifier configuration, so current might be up to 50A I guess.
BTW: I've succesfully calibrated today another current sensor designed for 25Amax currents based on analog SS495A Hall sensor and it is quite cheap and easy detect currents >100mA regardless of its polarity (direction in which current flow) with galvanic isolation, so it looks like I have everything needed to test this thing which is very interesting to see how it performs in real higher power application if it can work at all
-
Google "synchronous rectification".
-
Thank you for quick hints.
It looks like I do not "reinvent wheel", however I saw many car battery chargers schematics and never saw synchronous rectification designs in transformer secondary full-bridge, so started to investigate this thing while higher currents are needed
Google "synchronous rectification".
This is nice TI intro Synchronous Rectification in High-Performance Power Converter DesignGood idea but there is already a solution for it:
Just took a look into its datasheet and it looks like there is very little information about its external mosfets drivers gate currents in this integrated ideal diode bridge controller, so I have no idea for the moment how to estimate if 2-4 external mosfets can be driven by this IC?
http://www.linear.com/product/LT4320
There is also somethig else:Quote"The LT4320/LT4320-1 ideal diode bridge controller is designed for input frequencies from DC to 60Hz/600Hz, and operates with rectified outputs from 9V to 72V, supporting 24VAC, 32VAC and a wide variety of DC applications, but not 3-phase designs."
So, LT4320I looks nice for classic 24VAC secondary rectifier ideal diode bridge controller, but in one of my projects I have 3 phase output from... alternator.
Did not figured out for the moment which is its output 3 phase ACs maximum frequency @ lets say up to 3000 RPM -50Hz rotation frequency, but it looks like no way even LT4320-1 rated up to 600Hz is not designed for 3 phase full-bridge rectifier.
I will try scan internet, but maybe someone knows common nice 3 phase ideal diode bridge controller capable to drive a few mosfets in pararell like IRFZ44N 1500pF input capacitance each at frequency I guess not higher than 1000 Hz, so maybe those internal mosfet drivers in those ICs can do the job without external mosfet drivers...
-
but.. in the case of IRFZ44N mosfet with RDSON 0.018 Ohm @ 25*C and current 25A we have about 11 W loses in mosfet switch, while diode with 0.7V voltage drop has about 17 W lose
You may also find that a diode passing 25 A is dropping rather more than 0.7 V (unless it is a really BIG diode)... -
You may also find that a diode passing 25 A is dropping rather more than 0.7 V (unless it is a really BIG diode)...
One can find many things in datasheets. I don't know if those Shottky rectifier diodes MBRXXYY are big, but i've considered use it in another of the projects and in their primary characteristics datasheet we can find that its Vf is 0.73 V at 30 A, 0.65 V at 15 A so it is close to 0.70V in curents range of my interest, so for estimated power loses and quick comparision to synchronised rectifier based on low RDSON mosfets it should be fine.
Just trying to figure out if three of those ideal diode bridge controller LT4320I could be used in similar cnfiguration like classic 3xKBPC rectifiers below for 3 phase AC?
Internal LT4320I diagram could give the answer, but its datasheet has only block diagram and it is not clear if I can connect IN1 <-> IN2 together and make 3 phase ideal diode bridge controller using three LT4320Is...
-
Instead (multiples) of IRFZ44 think about something like this http://www.digikey.fi/product-detail/en/IRFS7530-7PPBF/IRFS7530-7PPBF-ND/4772487
Difference is much bigger when we use not 1 mosfet but much more depending on desired acceptable power loses, while when 2 mosfets are used than those loses in one rectifier switch will be about 5W, so in whole bridge about 10W, while when only diodes used and 2 of them conducting about 35 W and we can do nothing with this.
-
Instead (multiples) of IRFZ44 think about something like this
QuoteIRFS7530-7PPBF: Rds On (Max) @ Id, Vgs 1.4 mOhm @ 100A, 10V
Yep, realy nice RDSON, but mosfet switch selection is not such a big problem while I'm not limited by PCB space.
Design of three phase AC synchronized rectifier and proper driving of those mosfets is a challenge for the moment, however I think I'll make such prototype myself using a few opamps and own current sensors with optoisolated gates power supply.
If I find nice 3 phase ideal diode bridge controller It will be nice-for the moment just have to ensure in simulation and have working prototype to see it can work as expected
-
Quote from: eneuro
Yep, realy nice RDSON, but mosfet switch selection is not such a big problem while I'm not limited by PCB space .... however I think I'll make
I used to "roll" my own for ages, and it is VERY VERY involved. There are all sorts of issues to deal with, that you won't find until you try it.
such prototype myself using a few opamps and own current sensors with optoisolated gates power supply ....
It took me months to get it near perfect, and the parts count was huge.
I now use the LT4320 exclusively, mainly with AUIRFS8409-7Ps. At around 0.5mR, I can run them up to 30A without a heatsink !! The PCB is
double layer 4+4oz, but only ~2.0" square. The IC can drive gates up to 20,000pF and possibly more. It's an awesome chip !! -
I now use the LT4320 exclusively, mainly with AUIRFS8409-7Ps. At around 0.5mR, I can run them up to 30A without a heatsink !!
Unfortunately, LT4320 is useless in three-phase rectififer, I guess
IRF AUIRFS8409 mosfet datasheet shows typical 0.97 mR in its nice SMD version with thick source pin
Just played in simple circuit simulator with modeled three-phase AC ideal rectifier and do not burned any mosfets so far
I think, I know where put my bipolar Hall current sensors to make analog automatic switching of those mosfets in this ideal three-phase rectifier.
It looks like there is a lot of dead time between switching high and low mosfets in the same AC phase and mosfets are switched on/off at quite low currents in the case of nice sine AC I expect at alternator outputs or three-phase BLDC motors in regenerative brake mode, where there is a lof of power to rectify, so we'll see what happends in real hardware when prototype PCB will be ready for testing, but it looks like "piece of cake" after played this three-phase 50Hz ideal rectifier "game" in simulator for a few seconds, however for first runs will be better put cheaper mosfets into this ideal rectifier and see if the facts fits into theory
Note: It will not be ideal rectifier design, while idea is wait for currents close to 1A, while at 1A loses are only about 1.5W or less, so it does not matter, when I control hundreds wats of power...
-
Hi.
I think that in the case of 3 phases, well it depend how many diodes you use too but...
In one phase diode bridges heat a lot because the they do not conduct all the time of the period.
As a result, thez are passed by peaks of current and this make them heat a lot and can even break them
espacialy if you use a big filtering capacitor. With 3 phases the current is spread in time and there are nearly no dead time.
Some switching step down converter do not use capacitors at all in 3 phases and in monophase, thez use a smaller
capacitor than it should. It is the voltage difference with entry and ouput that make it possible.
But it is evident that if you have a strong current and a 4 diodes bridge in monophased current distribution , you will have many losses. If you yse mosfets rectifiers. The mosfets choice has to consider the current peaks such in the cas of a big filter capacitor.
Sorry for my english
-
Hi.
I've already solved those power loses in multi phase AC systems by using ynchronized rectifies based on current sensing in both directions with a litte help of Hall effect current sensors, which detects currents rushe above 1A or so easy, so no problem-rectifier diodes are bypassed by mosfets above some current threshold
I think that in the case of 3 phases, well it depend how many diodes you use too but...
-
TEA2206T happy days!
Also see "bridgless PFC" techniques