Voltage doubler improves accuracy
A transistor's low saturation voltage results in less loss.
S Chekcheyev, Tiraspol, Moldova; Edited by Martin Rowe and Fran Granville -- EDN, February 5, 2009
The voltage doubler in Figure 1 provides more accurate voltage doubling than does the conventional voltage doubler in Figure 2 because it uses transistors instead of diodes. You can express the output voltage of the conventional doubler as VOUTDC=2VINAC–2VD, where VOUTDC is the output dc voltage, VINAC is the amplitude of the input ac voltage, and VD is the voltage across the forward-biased diodes. The error of the conventional voltage doubler is 2VD. Transistors Q1 and Q2 in Figure 1 are saturated during the positive and the negative half-cycles, respectively, of the input ac voltage. The operation of the saturated transistors is similar to the operation of the forward-biased diodes in Figure 2. The collector-emitter voltage of the saturated bipolar transistors, however, is substantially smaller than the voltage across the forward-biased diodes. Thus, the error of doubling decreases.
|
Transistors Q1 and Q2 are reverse-biased during the negative and the positive half-cycles, respectively. The reverse beta of the bipolar transistors is small; consequently, the operation of the reversed transistors in Figure 1 is similar to the operation of the reverse-biased diodes in Figure 2. Both circuits underwent tests with a resistive load of 10 kΩ and a 50-Hz, 2V-amplitude sinusoidal signal applied to the input. The measured output voltage of the conventional voltage doubler was 2.8V, and the error of doubling was 2×2V–2.8V=1.2V. The measured output voltage of the proposed voltage doubler was 3.8V, and the error of doubling was 2×2V–3.8V=0.2V.
-
This article is equally usefull when fed from a DC pulse source, it's efficiency can be improved if the base of Q2 is switched by means of the collector (with series resistance added of course) of an pnp transistor with its base (again with a series resistance) driven from the junction of the emitter of Q2.
This way Q2 won't leak the negative voltage back during a positive pulse from the source. Also, the BC337 might be a better choice instead of the BC547, because of its lower saturation voltage and higher current rating.
This design as a charge-pump is better than using schottky diodes, especially at higher loads because the forward voltage of schottky's increases rather quickly when the current increases.
(Improved schematic "layout")
470u BC337
V_-_ o-+-||-+--+-----
Dimitri Princen - 2009-2-9 02:24:00 PDT -
I am sorry, but these circuitry on Figure 1. and 2. are not doubler but voltage inverters.
Peter K. - 2009-5-5 14:19:00 PDT -
This design idea is of dubious value.If I read this correctly, your conventional diodes drop is 1.4V if according to your measurement 2*VD=2.8V So I have a fundamental problem with your measurements, and your equation expressing the output voltage.
There is a simple solution to improving the conventional doubler. Use Schottky small signal diodes instead of conventional diodes.This also reduces concerns about fast recovery when doubling at high clock frequencies.
Cal Gold - 2009-15-3 06:49:00 PDT -
This design idea is nifty and permits use of 2 NiMH cells powering a 555 oscillator to drive white leds. My compliments to Mr. Chekcheyev.
Dave Wuchinich - 2009-5-2 18:57:00 PST -
I can’t believe that I have to spend my time correcting this article, in a prestigious (?) electronics magazine.
Starting from the text, where the input voltage is mentioned only as Vinac, with no specifications (as RMS, pk or pk-pk).
Secondly, the 47K resistor will limit the base current to about 1 mA for a pk positive alternance of 50V of the Vinac. But, not having anything to limit the saturation current in the collector of the BC547, this current can go from 10 to 100mA (and over - I have used 100mA as the datasheet limit). Well, for these currents, the Vce saturation of the BC547 is between 0.25v and 0.6V (see datasheet). So where is the advantage replacing the diodes with transistors?
And I didn’t mention the limitation due to the Vbe, at 6V. Any input voltage over 6Vpk (negative alternance) would destroy the transistor.
Who has given “green light” to this article?
Nicholas Adams - 2009-5-2 14:37:00 PST
