« PreviousContinue »
June 27, 1967 F Germann Etau 3,328,596
D.C.-A.C. CONVERTER FOR PRODUCING HIGH FREQUENCY OUTPUTS Filed July 3, 1963 2 Sheets-Sheet 1
June 27, 1967 F. Germann Etal 3,328,596
D.C.-A.C. CONVERTER FOR PRODUCING HIGH FREQUENCY OUTPUTS
Filed July 3, 1963 2 Sheets-Sheet 2
United States Patent Office Pat_ J»£
commutation during that period of time during which the 40 by either one or both of two conditions: One is that the
3 328 596
D.C.-A.C. CONVERTER FOR PRODUCING HIGH FREQUENCY OUTPUTS Fritz Germann, Ruthen, and Manfred Grote, Belecke, Germany, assignors to Licentia Patent-VerwaltungsG.m.b.H., Frankfurt am Main, Germany
Filed July 3,1963, Ser. No. 292,708 Claims priority, application Germany, July 4, 1962, L 42,387 5 Claims. (CI. 307—12)
The present invention relates to the art of D.C.-A.C. conversion.
It is often required to have A.C. voltages available having a frequency well above the conventional 50 or 60 c.p.s. It is conventional to employ frequency multiplication circuit networks to produce such higher frequencies. However, a sufficiently high power output is available only with the aid of quite expensive circuitry. Electric D.C.-A.C. converters are better than multipliers, but also better than machine type converters because electrical circuitry producing D.C.-A.C. conversion usually operates without mechanical contacts, though contact type converters have been developed and used. The known converters using controllable rectifier elements have yet not been able to produce satisfactory results in case of output frequencies and currents sufficiently high so as to be usable, for example, for inductive heating.
It is an object of the present invention to provide for a new and improved D.C.-A.C. converter. It is another object of the present invention to provide for improvements in the art of A.C.-D.C.-A.C. conversion.
According to one aspect of the present invention in a preferred embodiment thereof it is suggested to provide a bridge circuit having four bridge branches, each branch including at least one controllable semiconductor rectifier, for example, a silicon controlled rectifier. The bridge circuit defines a pair of D.C. input terminals and a pair of A.C. output terminals. An inductance is used to connect the D.C. input terminals to a D.C. voltage source. The latter source may be a rectifier in case A.C.-D.C.-A.C. conversion is to be provided for.
A resonance circuit which includes the load reactance is connected to the A.C. output terminals of the bridge circuit. In other words, for a given load, the ohmic and the non-ohmic component is determined, and by way of an additional reactance (inductance or capacitance as required) the load circuit is completed so that the nonohmic load component plus the additional reactance form a resonance circuit; the resonance frequency is to be below the desired A.C. output frequency as ultimately applied to the load. The semiconductor rectifier elements are controlled by a source producing control pulses at a rate frequency twice that of the desired A.C. output frequency; thus, this pulse rate frequency is more than twice the resonance frequency of the above mentioned resonance circuit of the load. How much the pulse rate frequency has to exceed twice this resonance frequency will be more fully explained below.
While the specification concludes with claims particularly pointing out and distinctly claiming the subject matter which is regarded as the invention, it is believed that the invention, the objects, and features of the invention and further objects, features and advantages thereof will be better understood from the following description taken in connection with the accompanying drawing in which:
FIGURE 1 illustrates a network diagram of an improved A.C.-D.C.-A.C. converter including an embodiment of the novel D.C.-A.C. converter;
FIGURE 2 illustrates a voltage-time diagram around the oscillatory passage through zero of the voltage at the
A.C. output terminals of the converter shown in FIGURE 1;
FIGURES 3 and 4 illustrate modifications of the network shown in FIGURE 1;
S FIGURE 5 is a circuit diagram of another embodiment of the present invention; and
FIGURE 6 is a circuit diagram of yet another embodiment of the present invention.
Proceeding now to the detailed description of the drawings, in FIGURE 1 thereof there is shown a D.C. voltage source which may be either a battery or, as shown specifically here, a rectifier 1 connected to the mains. A controllable rectifier assembly 10 comprised of four controllable semiconductor rectifier elements 11, 12,
15 21 and 22 is connected in bridge circuit configuration having A.C. output terminals A and B, and D.C. input terminals C and D so as to constitute a D.C.-A.C. converter. Elements 11, 12, 21 and 22 may be silicon controlled rectifier elements. A control or gating pulse source
20 6 delivers current pulses of short duration and of sufficient height to the elements 11,12, 21 and 22.
A choke or inductance 2 is connected in series between terminal C and one pole, here the plus pole, of D.C. voltage source 1. The voltage drop between D.C. input
25 and A.C. output is developed across this choke 2.
The assembly 10 has two A.C. output terminals A and B as stated to which is connected a resonance circuit, such as a tank circuit, comprised of an inductance 3 and a capacitance 4 connected in parallel thereto. An adjust
30 able ohmic resistor 5 may be connected in parallel to this tank circuit and also between terminals A and B.
Since the converter output voltage and frequency appears across terminals A and B the pulse source 6 is preferably linked thereto to operate in synchronism with said
35 output frequency.
Each circuit element of the oscillator circuit may constitute the actual load, but the load may also 'be a combination of such circuits elements or a portion thereof. The controllable semiconductor rectifier elements 11, 12,
40 21 and 22 are known per se, and it is also known that these elements have characteristics similar to that of a thyratron.
A silicon controlled rectifier is of the four layer type having three distinct p-n junctions. The rectifier is ren
45 dered conductive whenever the voltage applied to the end terminals thereof and in forward direction of the two outer junctions exceeds the breakdown voltage as primarily determined by the innermost and reversely directed p-n junction. This innermost p-n junction, however, may
50 also be biased in its forward direction by a sufficiently large current pulse delivered thereto by a gate terminal at one of the inner layers. When this happens, current will flow in forward direction between the end electrodes even if the voltage applied thereto is below breakdown
Once current flow has started across the end terminals of such rectifier elements and has exceeded the current value corresponding to the breakdown voltage, current flow will continue indeed even if the voltage drop across
60 the end terminals remains below the breakdown voltage and even if the current pulse at the gate terminal has disappeared. Current flow will continue as long as the voltage across the end terminals is above the knee or sluice voltage, which is, of course, much below the break
65 down voltage.
Since similar function is present, analogous terminology to thyratron operation is proper and commonly employed; the current pulse applied to the gate terminal of the controlled semiconductor rectifier indeed "fires" it, i.e., it not
70 only renders the rectifier conductive during the application of the pulse, but conduction continues as long as there is a driving voltage across the cathode and anode constituted
by the end terminals. For converter operation, the elements are fired in pairs (11, 12-21, 22) and in synchronism so that the oscillator circuit is placed into the load circuit; blocking of the semiconductor rectifier elements occurs in an analogous manner.
Each controllable rectifier element will be rendered conductive only if (1) there is a minimum voltage equal to the sluice or knee voltage being applied in forward direction at its main or end electrodes and (2) there is a control current signal present at its control or gating electrode, assuming that the operational voltage is below the breakdown voltage. The minimum voltage value to be applied in forward direction at the main electrodes for firing the rectifier is determined by the voltage drop across a semiconductor rectifier element when being rendered conductive, the knee or sluice voltage thereof.
The tank circuit defines and determines the shape and time-dependent magnitude of the voltage across terminals A and B. This voltage further determines when in effect any semiconductor rectifier element having forward voltage applied to its main or end electrodes can be fired or can be blocked.
Before describing the converter operation in detail and before discussing the general conditions of operation, certain aspects with regard to individual semiconductor rectifier operation as well as terminology will be discussed first.
D.C. to A.C. conversion requires commutation of the electric current from a conductive semiconductor element over to a nonconductive element. In the present case, and using a bridge type rectifier assembly, such commutation actually occurs simultaneously for pairs of elements. Current commutation is possible only if the voltage applied to the respectively nonconductive element or elements at commencement of commutation is effective at its main electrodes in forward direction and has a magnitude equal to or above the sluice or knee voltage as stated above. Consequently, any presently conductive semiconductor element can be blocked or cut off only by way of current
conductive semiconductor element must occur at the latest, so that cut off will occur and the current will, in fact, commutate for proper D.C.-A.C. conversion. The sluice voltage of the semiconductor rectifier ele5 ments determines only one portion of phase shift oe0, other influencing factors will be determined as follows. At higher frequencies there is an inherent delay between the occurrence of a firing pulse for a nonconductive semiconductor element and actual extinction of conduction of the previ10 ously conductive semiconductor element, since at higher oscillator frequencies the minimum angle at which the voltage passes through the corresponding critical voltage for firing is negligibly small.
During current commutation the various electric cur15 rents and current components in the circuit network cannot be altered stepwise due to network reactances. Furthermore, the inherent cut off-delay time is a characteristic feature of semiconductor rectifier elements determined by physical properties thereof and is determined by the mini20 mum period of time required to eliminate free charge carriers in the previously conductive element to really terminate conduction therein and to produce cut off thereof. Elimination of charge carriers is had by recombination or discharge (outflow) and must be complete to such an 25 extent that no reversion to conduction occurs unless a succeeding control pulse is delivered. Hence, at high operating frequencies the required minimum firing angle ot0 is not determined by minimum voltage requirements as between terminals A and D (or B and C) as defined above, 30 but by the fact that from the moment of firing one element for cut off of the yet conductive rectifier element, the current cut off in the latter must occur still faster than the natural sloping through zero of the voltage AB as shown in FIGURE 2.
35 In summary, commutation of current must be initiated by firing the nonconductive semiconductor element at an angle a1>a0 measured from the time of oscillatory passage through zero of the voltage across the output terminals (A and B) which minimum angle a0 is determined
oscillator or resonance circuit delivers a voltage in effect opposing the voltage derived from the D.C. voltage source but still having a magnitude sufficiently above hte sluice voltage of the rectifier element to be fired.
Turning now for a moment to FIGURE 2, there is shown a curve illustrative of the following situation: Assuming that elements 11 and 12 are momentarily conductive and assuming first that no current commutation is present, then the voltage across terminals A and B is an oscillating one and has at the end of one half cycle a sloping configuration as shown as curve AB in FIGURE 2.
Without current commutation, this voltage across terminals A and B would oscillatorily pass through zero at t0. However, in order to attain current commutation at all,
downwardly sloping oscillatory voltage must not go below a value at which no firing of the nonconductive semiconductor element is possible anymore (a01). The other condition is that the current extinction in the element from which current is commutated away, must not occur at a rate slower than the rate of voltage decline for reversion at the output terminals by pure oscillator action (a02)The effective a0 then is the respective largest value of <x0i and a02
Now the complete converter operation will be described next, but only stationary converter operation is to be considered.
Assuming that semiconductor rectifier elements 11 and 12 have been rendered conductive, then the positive poten
the existing voltage for, for example, semiconductor 55 tial of source 1 as effective at terminal C is applied to ter
rectifier element 22, must be directed so that terminal A
is still positive relative to terminal B. Otherwise, semicon-
ductor rectifier element 22 could not be fired at all. Par-
ticularly after a voltage reversal at terminals A and B,
no firing of elements 21 and 22 is possible. 00
The voltage across the end electrodes of semiconductor
rectifier 22 is that between terminals A and D, which is
composed of the momentarily effective voltage between
terminals A and B and the voltage drop in conductive
rectifier element 12, which is about the sluice voltage. In 65
FIGURE 2 now, a0 denotes (measured in electric degrees
for the cycle of the reasonance circuit) the angle measured
from the zero passage at curve AB (time t0) back to a
point at which rectifier elements 21 and 22 could still be
fired. Thus, current commutation must occur at a firing 70
angle on such, as for example, representing a time tx prior
to that of angle a0 to ensure due commutation.
«o is the critical angle in electrical degrees measured from the passage through zero (r0) of the oscillator voltage back to the time at which the cut off of the respectively 75
minal A, whereas the corresponding negative potential from terminal D is being applied to terminal B.
At a particular moment ti, semiconductor rectifiers 21 and 22 are fired. This is possible, as long as terminal A is still sufficiently positive relative to B. However, upon firing of element 21 terminal C will temporarily assume the negative potential of terminal B, whereas upon concurrent firing of element 22 terminal D will temporarily assume the yet positive potential of terminal A. In both instances and by concurrent action rectifier cells 11 and 12 will thus be driven to cut off, provided the potential difference between terminals A and B is larger than the voltage drop across elements 21 and 22 when conductive.
Thereafter, the voltage between A and B decreases to zero, not along oscillatory curve AB of FIGURE 2 but by forced current commutation as determined by the cut off delay outlined above. This passage through zero will thus occur at any time between tx and t0. The voltage across A, B reverses and swings again by oscillator reso