Circuit forms simple, low-cost, 1-kV driver
Edited by Bill Travis
Tai-Shan Liao and Prasit Champa, National Science Council, Hsinchu, Taiwan -- EDN, May 13, 2004
High-voltage drivers have recently received much attention, because they play an important role in driving piezoelectric and electro-optical components, for example. Figure 1 shows a simple, low-cost, 1-kV driver. The circuit uses offline, current-mode-control techniques and a flyback switching-power-supply design. IC1, a UC3844, is the major control component, using a switching frequency of 100 kHz. The IC provides frequency modulation to reduce the switching frequency under light- and no-load conditions. The feedback voltage, which you derive from the output of the error amplifier, serves as the indicator for load conditions. Once the feedback voltage becomes lower than the green-mode threshold voltage, the switching frequency starts to decrease.
All the power losses are in direct proportion to switching frequency. These losses include the switching losses of the transistor, core losses in the transformer and inductors, and the power loss of the snubber. The frequency modulation in the PWM-controller IC reduces the power consumption in the supply under light- and no-load conditions. But the frequency modulation has no effect on the PWM operation under normal- and high-load conditions.
Pin 2 (the feedback pin) of the UC3844 sums the current-sense signal, the output-voltage feedback signal, and any added slope compensation. The feedback-control circuit uses a TL431 adjustable shunt regulator to detect the output signal. A PC817 passes the signal to the feedback pin of the UC3844. The TL431 acts as an open-loop error amplifier with a 2.5V temperature-compensated reference. When the output voltage is lower than the desired level, the feedback to the UC3844 automatically compensates the pulse-width modulation of the output triggering signal. Ceramic bypass capacitors (0.1 µF) from VCC and VREF to ground provide low-impedance paths for high-frequency transients. This design uses a Tomita (www.tomita-electric.com) EI25-2E6 core set to fabricate the transformer. To prevent core saturation, the gap is approximately 1 mm. The primary winding has 70 turns of 28-gauge wire. Both the secondary windings have 105 turns of 34-gauge wire. The primary and secondary auxiliary windings have five and six turns, respectively, of 34-gauge wire. The dc output voltage of the circuit in Figure 1 is 1 kV (fixed). You can adjust the output voltage in a 50V range by adjusting VR1. Both load and line regulation are less than 1%, and power efficiency is 80% at full load.
Check out our Best of Design Ideas section!
-
Nice circuit, but it's not goiing to work....
Q1 should be connected to the anode of D1.
The connection between anode of D1 and anode of the other FR137 diode should not be there.
Now Q1 switches between ground and ground, at least it won't get hot this way.
Hein Godwaldt - 2004-23-6 23:53:00 PDT -
Take a look at the Circuit connections on the primary side of the transformer, I do not think it will work the way it is shown with Q1 connected to the bottom side of T5 and that same side grounded.
Dan Leiwe - 2004-20-5 13:27:00 PDT -
I'm sure I'm not the first to notice, but there is an error in Figure 1.
The line going up from Q1 should not connect to winding T5 and there should not be a connection between windings T1 and T5. Instead, the drain of Q1 should connect to the dotted end of winding T1.
In other words get rid of the two dots at each end of T5 and put the segment back connecting Q1 and T1.
That's the kind of mistake I catch (and make) all the time.
Daniel Carroll - 2004-14-5 22:00:00 PDT -
I think there is a problem with the drawing of Figure 1. I suspect that the link between T1 and T5, and the link between T5 and Q1 Drain, should not be there - rather the anode of D1 should be connected to the Drain of Q1. The bootstrap winding T5 is connected to ground, so Q1 would be rather ineffective!
Also, the gate and source of Q1 are shown connected together!
The efficiency could be improved by using a start-up regulator in place of the 30K 1W resistor connected to pin 7 and the bootstrap circuit.
Finally, I presume that the ground connection on the secondary side is isolated from the ground on the primary side, even though the same symbols have been used?
Steve Winder - 2004-13-5 10:52:00 PDT
