Forums

Synchronous post-rectification

Started by Piotr Wyderski May 19, 2014
Hello,

my master power supply provides me with synchronously rectified,
but not filtered (!) AC current. For the used 12V toroidal transformer
the output waveform is thus 17V*abs(sin(50Hz)), up to 40A. I want to
use this supply to power a bank of 10 SEPIC switchers. Issume 20A
load in the worst case. It means I must post-rectify the input
current in order to feed a big filter capacitor. The capacitor
cannot be moved to the point directly after the primary rectifier,
because its control circuit monitors the transformer's voltage,
not the input-output difference as a diode would do (a design
feature + I really don't want it that way).

A simple big-enough Schottky diode would solve the problem, but
it would produce too much heat at 20 amps. Alternatively, I think
about a secondary synchronous rectifier based on a PMOS. Below is
what I came up with. The circuit basically simulates OK, but:

1. Is the real performance of LT1716 anywhere close to the
SPICE model, especially if it comes to its advertised "over the
top" capability?

2. The rising edge of the current on R6 contains a 40 mA bump
with negative flow direction (you need to magnify the "plateau"
to see it). It's not much, but I don't like it. Should I care?
Its origin is a kind of mystery, because the JFET-based gate
discharge circuit seems to work like a charm.

3. How can I solve this problem better? The volume of the final
circuit will be whopping 1 or 2 instances (it's a hobby project),
so I am not repelled by exotic parts. The reliability is, however,
very important.

	Best regards, Piotr


Version 4
SHEET 1 1880 708
WIRE 896 -16 640 -16
WIRE 992 -16 976 -16
WIRE 1280 -16 992 -16
WIRE 1408 -16 1360 -16
WIRE 1504 -16 1488 -16
WIRE 1280 0 1280 -16
WIRE 640 16 640 -16
WIRE 1056 32 1024 32
WIRE 1024 64 1024 32
WIRE 1024 64 992 64
WIRE 1408 64 1408 -16
WIRE 1408 64 1328 64
WIRE 112 96 -128 96
WIRE 320 96 112 96
WIRE 336 96 320 96
WIRE 464 96 416 96
WIRE 560 96 560 64
WIRE 560 96 544 96
WIRE 656 96 656 64
WIRE 704 96 656 96
WIRE 768 96 704 96
WIRE 832 96 768 96
WIRE 1024 96 832 96
WIRE 1056 96 1024 96
WIRE 1280 96 1056 96
WIRE 128 160 112 160
WIRE 224 160 208 160
WIRE 320 160 304 160
WIRE 112 176 112 160
WIRE 320 176 320 160
WIRE 656 176 656 96
WIRE 656 176 592 176
WIRE 464 208 464 96
WIRE 624 208 464 208
WIRE 592 240 592 176
WIRE 624 240 624 208
WIRE 1024 240 1024 96
WIRE 576 272 544 272
WIRE 704 272 704 96
WIRE 704 272 640 272
WIRE 544 288 544 272
WIRE 608 336 608 304
WIRE 1024 336 1024 320
WIRE 1024 336 608 336
WIRE 1504 336 1504 -16
WIRE 1504 336 1024 336
FLAG 112 240 0
FLAG 320 240 0
FLAG -128 176 0
FLAG 768 160 0
FLAG 832 176 0
FLAG 544 288 0
SYMBOL schottky 128 160 R180
WINDOW 0 24 64 Left 2
WINDOW 3 42 38 Left 2
SYMATTR InstName D1
SYMATTR Value MBR745
SYMATTR Description Diode
SYMATTR Type diode
SYMBOL voltage 224 160 R90
WINDOW 0 -32 56 VBottom 2
WINDOW 3 32 56 VTop 2
WINDOW 123 0 0 Left 2
WINDOW 39 0 0 Left 2
SYMATTR InstName V1
SYMATTR Value SINE(0 17 50)
SYMBOL res 320 144 R90
WINDOW 0 0 56 VBottom 2
WINDOW 3 32 56 VTop 2
SYMATTR InstName R1
SYMATTR Value 1m
SYMBOL schottky 336 160 R180
WINDOW 0 -31 10 Left 2
WINDOW 3 241 179 Left 2
SYMATTR InstName D2
SYMATTR Value MBR745
SYMATTR Description Diode
SYMATTR Type diode
SYMBOL schottky 336 240 R180
WINDOW 0 -32 63 Left 2
WINDOW 3 -103 34 Left 2
SYMATTR InstName D3
SYMATTR Value MBR745
SYMATTR Description Diode
SYMATTR Type diode
SYMBOL schottky 128 240 R180
WINDOW 0 24 64 Left 2
WINDOW 3 49 24 Left 2
SYMATTR InstName D4
SYMATTR Value MBR745
SYMATTR Description Diode
SYMATTR Type diode
SYMBOL res 816 80 R0
WINDOW 3 57 76 Left 2
SYMATTR Value 10
SYMATTR InstName R2
SYMBOL cap 752 96 R0
SYMATTR InstName C1
SYMATTR Value 1000µ
SYMBOL res -144 80 R0
SYMATTR InstName R18
SYMATTR Value 10
SYMBOL res 432 80 R90
WINDOW 0 0 56 VBottom 2
WINDOW 3 32 56 VTop 2
SYMATTR InstName R3
SYMATTR Value 1m
SYMBOL Comparators\\LT1716 608 272 R90
SYMATTR InstName U1
SYMBOL pmos 560 16 M90
SYMATTR InstName M1
SYMATTR Value Si4427DY
SYMBOL res 992 -32 R90
WINDOW 0 0 56 VBottom 2
WINDOW 3 32 56 VTop 2
SYMATTR InstName R4
SYMATTR Value 1m
SYMBOL zener 1040 32 R0
WINDOW 0 49 24 Left 2
WINDOW 3 18 -13 Left 2
SYMATTR InstName D5
SYMATTR Value BZX84C12L
SYMATTR Description Diode
SYMATTR Type diode
SYMBOL njf 1328 0 M0
SYMATTR InstName J1
SYMATTR Value U309
SYMBOL res 1504 -32 R90
WINDOW 0 0 56 VBottom 2
WINDOW 3 32 56 VTop 2
SYMATTR InstName R10
SYMATTR Value 1m
SYMBOL res 1376 -32 R90
WINDOW 0 0 56 VBottom 2
WINDOW 3 32 56 VTop 2
SYMATTR InstName R11
SYMATTR Value 1k
SYMBOL res 976 -32 R0
SYMATTR InstName R14
SYMATTR Value 1m
SYMBOL res 560 80 R90
WINDOW 0 0 56 VBottom 2
WINDOW 3 32 56 VTop 2
SYMATTR InstName R6
SYMATTR Value 1m
SYMBOL res 1008 224 R0
SYMATTR InstName R5
SYMATTR Value 1K
TEXT -32 264 Left 2 !.tran 0.15





---
Ta wiadomość e-mail jest wolna od wirusów i złośliwego oprogramowania, ponieważ ochrona avast! Antivirus jest aktywna.
http://www.avast.com

Without looking at the circuit;
1. Probably?..

2. Sounds like recovery, or the equivalent?  Nothing can be instantaneous. 
There will always be something, it's up to you if it's too much or not.

3. The first thought is: choke input rectification.  That gets you average 
output voltage without needing funky circuits.  You don't even need a 
large value (relatively speaking), because current through the sync rect 
will be continuous even at light loads.  Your power factor will suffer 
though.

Tim

-- 
Seven Transistor Labs
Electrical Engineering Consultation
Website: http://seventransistorlabs.com

"Piotr Wyderski" <peter.pan@neverland.mil> wrote in message 
news:llcg8r$8bc$1@node2.news.atman.pl...
> Hello, > > my master power supply provides me with synchronously rectified, > but not filtered (!) AC current. For the used 12V toroidal transformer > the output waveform is thus 17V*abs(sin(50Hz)), up to 40A. I want to > use this supply to power a bank of 10 SEPIC switchers. Issume 20A > load in the worst case. It means I must post-rectify the input > current in order to feed a big filter capacitor. The capacitor > cannot be moved to the point directly after the primary rectifier, > because its control circuit monitors the transformer's voltage, > not the input-output difference as a diode would do (a design > feature + I really don't want it that way). > > A simple big-enough Schottky diode would solve the problem, but > it would produce too much heat at 20 amps. Alternatively, I think > about a secondary synchronous rectifier based on a PMOS. Below is > what I came up with. The circuit basically simulates OK, but: > > 1. Is the real performance of LT1716 anywhere close to the > SPICE model, especially if it comes to its advertised "over the > top" capability? > > 2. The rising edge of the current on R6 contains a 40 mA bump > with negative flow direction (you need to magnify the "plateau" > to see it). It's not much, but I don't like it. Should I care? > Its origin is a kind of mystery, because the JFET-based gate > discharge circuit seems to work like a charm. > > 3. How can I solve this problem better? The volume of the final > circuit will be whopping 1 or 2 instances (it's a hobby project), > so I am not repelled by exotic parts. The reliability is, however, > very important. > > Best regards, Piotr
On Mon, 19 May 2014 10:44:11 +0200, Piotr Wyderski
<peter.pan@neverland.mil> wrote:

>Hello, > >my master power supply provides me with synchronously rectified, >but not filtered (!) AC current. For the used 12V toroidal transformer >the output waveform is thus 17V*abs(sin(50Hz)), up to 40A. I want to >use this supply to power a bank of 10 SEPIC switchers. Issume 20A >load in the worst case. It means I must post-rectify the input >current in order to feed a big filter capacitor. The capacitor >cannot be moved to the point directly after the primary rectifier, >because its control circuit monitors the transformer's voltage, >not the input-output difference as a diode would do (a design >feature + I really don't want it that way). > >A simple big-enough Schottky diode would solve the problem, but >it would produce too much heat at 20 amps. Alternatively, I think >about a secondary synchronous rectifier based on a PMOS. Below is >what I came up with. The circuit basically simulates OK, but: > >1. Is the real performance of LT1716 anywhere close to the >SPICE model, especially if it comes to its advertised "over the >top" capability? > >2. The rising edge of the current on R6 contains a 40 mA bump >with negative flow direction (you need to magnify the "plateau" >to see it). It's not much, but I don't like it. Should I care? >Its origin is a kind of mystery, because the JFET-based gate >discharge circuit seems to work like a charm. > >3. How can I solve this problem better? The volume of the final >circuit will be whopping 1 or 2 instances (it's a hobby project), >so I am not repelled by exotic parts. The reliability is, however, >very important. > > Best regards, Piotr
Not sure what the model is intended to demonstrate, but the aim of the synchronous rectifier is to reduce rectifier forward voltage drop. You might get something from an old web page, with LTspice files at: http://www.magma.ca/~legg/20_Years/1994/1994h.html Low frequency rectification or 'orring' circuits don't necessarily need single-source controller parts. RL
On 19/05/2014 18:44, Piotr Wyderski wrote:
> Hello, > > my master power supply provides me with synchronously rectified, > but not filtered (!) AC current. For the used 12V toroidal transformer > the output waveform is thus 17V*abs(sin(50Hz)), up to 40A. I want to > use this supply to power a bank of 10 SEPIC switchers. Issume 20A > load in the worst case. It means I must post-rectify the input > current in order to feed a big filter capacitor. The capacitor > cannot be moved to the point directly after the primary rectifier, > because its control circuit monitors the transformer's voltage, > not the input-output difference as a diode would do (a design > feature + I really don't want it that way). > > A simple big-enough Schottky diode would solve the problem, but > it would produce too much heat at 20 amps. Alternatively, I think > about a secondary synchronous rectifier based on a PMOS. Below is > what I came up with. The circuit basically simulates OK, but: > > 1. Is the real performance of LT1716 anywhere close to the > SPICE model, especially if it comes to its advertised "over the > top" capability? > > 2. The rising edge of the current on R6 contains a 40 mA bump > with negative flow direction (you need to magnify the "plateau" > to see it). It's not much, but I don't like it. Should I care? > Its origin is a kind of mystery, because the JFET-based gate > discharge circuit seems to work like a charm. > > 3. How can I solve this problem better? The volume of the final > circuit will be whopping 1 or 2 instances (it's a hobby project), > so I am not repelled by exotic parts. The reliability is, however, > very important. > > Best regards, Piotr
[snip LTSpice] Maybe you could put a boost converter after the synchronous rectifier, with the output of the boost converter charging a reservoir capacitor that is always at a voltage somewhat higher than the peaks from the synchronous rectifier. By modulating the duty cycle of the boost converter, this would even allow you to do power factor correction, which might make the transformer run cooler. It might be tricky to manage the initial turn-on current, unless you can start up the synchronous rectifier gently, or you use a buck-boost in place of the boost converter. This all assumes that the SEPIC switchers can take a voltage somewhat higher than the peak of the output of the synchronous rectifier, which I guess is fairly likely since you were going to run them from that. Chris