Electric power conversion systems – Current conversion – With condition responsive means to control the output...
Reexamination Certificate
2000-08-11
2002-09-10
Berhane, Adolf Deneke (Department: 2838)
Electric power conversion systems
Current conversion
With condition responsive means to control the output...
C323S251000
Reexamination Certificate
active
06449175
ABSTRACT:
FIELD OF THE INVENTION
The present invention generally relates to power converters having magnetic amplifier (magamp) post regulators and, more particularly, to circuitry used to reduce the power and efficiency loss in the control transistor of a set-mode magamp post regulator.
BACKGROUND OF THE INVENTION
The magamp post regulator is a popular power supply topology for regulating the outputs of a power converter in many applications. Modern electronic devices often require several voltage outputs; and need a low cost, energy efficient and well regulated way of providing these outputs. Magamps are typically used to provide an efficient and reliable way of providing precise voltage regulation of independent outputs of a multiple output power converter. A magamp post regulator provides improved regulation of power converter output voltage using a small control current.
The basis function of a magamp is to block a positive incoming voltage for a certain time (t
block
) before allowing it to pass through an output filter. The duty cycle reduction occurs because the magamp delays the leading edge of the voltage waveform. The magamp acts to reduce the duty cycle to the rest of the circuit from the duty cycle of the incoming voltage so as to maintain the required average output voltage.
Conventional magamp post regulator circuits use a reset control to control the magamp using a control transistor operated in a linear mode. 
FIG. 1
 illustrates a prior art example of a conventional reset-controlled magamp circuit 
10
. 
FIG. 2
 illustrates the hysteresis characteristic of the core element of the magamp of the circuit of FIG. 
1
. The conventional magamp circuit includes a magamp 
16
, a diode 
12
, a reset transistor 
20
 and an error amplifier (error amp) 
18
. In 
FIG. 1
, when a power switch 
34
 is turned on, a secondary voltage V
sec 
is developed across a transformer 
14
 secondary winding. Magamp 
16
 is forced into saturation due to the action of the voltage V
sec 
forced upon it. The B-H hysteresis curve in 
FIG. 2
 shows the saturation point, B
saturation 
at the top of the path. Since the magamp 
16
 is “in saturation”, forward biased and highly conductive, current flows through the magamp to a forward output rectifier diode 
22
 after which it is filtered by an L-C circuit, comprised of inductor 
24
 and capacitor 
26
. The output voltage is coupled to a load, not shown, and is also divided by a voltage divider formed by series resistors 
36
 and 
38
 to generate a Voltage sense signal at node 
35
. At the end of the switch “on” time, the magamp 
16
 remains forward biased and in saturation.
When the main power switch 
34
 turns off and the transformer 
14
 voltage reverses polarity to −V
sec 
the current through the magamp 
16
 is caused to ramp down. As a result, a vertical rectifier diode 
30
 must pick up the output current, causing the voltage at node 
15
 to drop. In this off state the magamp voltage V
m 
is not allowed to reach zero. Instead a reset control circuitry supplies a voltage that reversely biases the magamp 
16
, such that the magnetic flux density is reset to a point below remanence (below the point B
remanence 
of the left side of dark shaded area in FIG. 
2
.). Then the main switch is turned on and the transformer 
14
 secondary voltage becomes +V
sec
. Since Magamp 
16
 is well below the saturation point and not conductive, it acts as an open circuit and blocks the secondary voltage. Vertical diode 
30
 continues to provide a path for the output current so the voltage at node 
15
 remains at zero. The magamp voltage Vm then equals +V
sec
. In time, the voltage across magamp 
16
 causes it to reach saturation and become conductive. The current through the magamp rises to the output current level and remaining at this level till the end of the on time.
The flux excursion on the B-H curve of 
FIG. 2
 depends on how much volt-time is applied across the magamp 
16
 during resetting. The amount of volt-seconds is controlled by the output of error amp 
18
. The blocking time equation is given by 
t
block
=
Δ
⁢
 
⁢
B
·
turns
·
A
core
V
;
where A
core 
is the core area, &Dgr;B is the change in flux density, turns is the number of turns for the core, and V is the voltage. It can be seen from this equation that the loop in 
FIG. 2
 corresponding to &Dgr;B
2
 gives a longer blocking time that the loop of &Dgr;B
1
. The cores required for this prior art method of reset control exhibit a relatively square B-H curve. To lower the output voltage and increase the blocking time, the loop followed is the lightly shaded part of the B-H curve as compared to the dark part. The control circuit forces the B-H loop larger by pushing the vertical, descending part of the locus. Thus, the minimum blocking voltage-time is the locus where it just touches the vertical axis. To maximize the difference between maximum and minimum volt-time blocking, the B-H loop of the core material must have a small difference between B
saturation 
and B
remanence
, where it intercepts the vertical axis.
Compared to square loop amorphous core magamps, ferrite magamps are lower cost, better for high frequencies and can run at higher temperature. However, a drawback associated with this conventional reset control approach is that lower cost non-square ferrite cores perform poorly under reset control because the power dissipation at high flux excursion is too large, especially for operation at high frequency.
A prior art example of a conventional circuit for magamp post regulator control without using reset control but instead using a “set” mode with a control circuit in a linear mode, is shown in FIG. 
3
. This set control enables the use of lower cost ferrite cores for the magamp core, however, operation in linear mode leads to unacceptable losses in the circuit. The corresponding B-H hysteresis characteristic of the core member of the magamp of the circuit is shown in FIG. 
4
. For the magamp post regulator 
40
 in 
FIG. 3
, an error amp 
48
 feeds a control transistor 
50
 which is operated in linear mode. When the transformer 
44
 secondary voltage V
sec 
turns negative in response to power switch 
64
, a diode 
42
 and a control transistor 
50
 “catch” the current through magamp 
46
. Depending on the voltage output from error amp 
48
, the current through the loop of diode 
42
, control transistor 
50
 and magamp 
46
 is decreased, and the corresponding change in &Dgr;H and &Dgr;B is achieved (as shown in 
FIG. 4
, the current is related to H by the equation H*L
core
=turns * I.) During the next positive cycle, the magamp 
46
 will block the secondary voltage V
sec 
The blocking time, T
block
, according to the equation described above, 
t
block
=
Δ
⁢
 
⁢
B
·
turns
·
A
core
V
,
is proportional to &Dgr;B (turns, A
core 
and V are constant for the equation). As the curve in 
FIG. 4
 illustrates, set control mode operates only at one quadrant of the B-H curve while the reset control, as shown in 
FIG. 2
, can operate at all four quadrants. In this “set” mode circuit, the control circuit tries to prevent the core from resetting, i.e. tries to make a smaller loop. Since there is no requirement for the core to be square, non-square less costly ferrites can be used.
FIG. 5
 shows another prior art version of set control for a magamp post regulator. For this magamp post regulator circuit 
70
, in addition to the magamp 
76
 power winding, there is an extra magamp control winding 
77
. A driver diode 
72
 and a control transistor 
80
 control the magamp control winding 
77
, with the control elements isolated from the transformer 
74
 secondary power winding. The current through the diode 
72
 and control transistor 
80
 can be reduced depending on the turns ratio of the control winding and power winding. 
FIG. 6
 shows a corresponding set of timing curves for the magamp set control circuit of FIG. 
5
. The top curve 
1
, is the secondary voltage and V
p 
is the transformer 
74
 primary voltage, curve 
2
 is
Cuadra Jason
Melgarejo Manolo Mariano M.
Astec International Limited
Berhane Adolf Deneke
Coudert Brothers LLP
LandOfFree
Switched magamp post regulator does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Switched magamp post regulator, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Switched magamp post regulator will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2871288