Q3Advanced Power Electronics

(a) Single-Phase to Single-Phase Cycloconverter — Bridge Configuration

A cycloconverter directly converts AC power at one (supply) frequency into AC power at a different (usually lower) output frequency, without any intermediate DC link stage, by appropriately sequencing the firing of thyristors from two anti-parallel-connected converter bridges (a 'positive group' P-converter and a 'negative group' N-converter), each capable of independently producing either polarity of average output voltage depending on firing angle, but each conducting current in only one direction.

For a bridge-type single-phase to single-phase cycloconverter, during the desired positive half-cycle of the (lower-frequency) output waveform, the P-converter bridge is enabled and its thyristors are fired with a firing angle α that is continuously, cyclically varied (following approximately a cosine-modulated pattern) so that the resulting sequence of chopped supply half-cycles approximates the desired low-frequency sinusoidal output envelope; during the desired negative half-cycle of the output, the P-converter is disabled and the N-converter bridge takes over, similarly firing with a continuously varying angle to produce the negative-polarity portion of the output waveform.

Continuous (Circulating-Current) Conduction Mode

In this mode, both the P-converter and N-converter are kept permanently enabled (gated) simultaneously, with their firing angles maintained in a specific complementary relationship (αP + αN = 180°) that keeps their average DC output voltages equal in magnitude and opposite in the sense needed to prevent net short-circuit current, but a small AC ripple-driven circulating current still flows continuously around the loop formed by the two bridges (limited by an inter-group reactor connected between them), even when the load itself draws no current. This mode gives smoother, more continuous output current (avoiding current discontinuity/dead-time near output current zero crossings) and better dynamic response, at the cost of the additional continuous circulating-current loss and the extra size/cost of the inter-group reactor.

Discontinuous (Non-Circulating-Current) Conduction Mode

In this mode, only one converter bridge (whichever corresponds to the current output current direction) is enabled/gated at any given instant, with the other bridge's gate pulses completely blocked, and a brief dead-time/current-zero-detection delay is inserted whenever the load current changes direction (crosses zero) before switching over which bridge is enabled, to prevent any possibility of both bridges conducting simultaneously and short-circuiting the supply through the loop. This mode eliminates continuous circulating current and its associated losses, and does not need an inter-group reactor, but introduces a brief current discontinuity/distortion near each current zero-crossing (since detecting the actual zero-crossing and safely switching over converters takes a finite time), somewhat degrading output waveform quality compared to the circulating-current mode, particularly at low output current levels or unity power factor loads where current crosses zero frequently relative to the output cycle.

(b) Advantages/Disadvantages vs DC-Link Inverters, and Applications

Advantages of cycloconverters over DC-link inverters: direct single-stage AC-AC power conversion, avoiding the losses and additional component count (rectifier stage, DC-link capacitor/inductor, separate inverter stage) of a two-stage DC-link converter; naturally bidirectional power flow capability (regeneration back to the supply) without any additional control complexity, since the same thyristor bridges simply reverse their roles; and the ability to handle very large power levels using relatively simple, robust, naturally line-commutated thyristor technology (no forced commutation circuitry needed, unlike older thyristor-based inverters), making cycloconverters historically well suited to very large motor drive applications.

Disadvantages of cycloconverters compared to DC-link inverters: the maximum practical output frequency is limited to a fraction (typically at most about one-third to one-half) of the input supply frequency, since a reasonable number of supply half-cycles must be 'assembled' to synthesize each output cycle with acceptable waveform quality — this makes cycloconverters unsuitable for applications requiring an output frequency equal to or higher than the supply frequency, unlike DC-link inverters, which can readily produce any desired output frequency (including higher than the input) independent of the supply frequency; cycloconverters draw a poorer displacement power factor from the supply (since thyristor firing angle control inherently draws lagging reactive power, and this reactive burden becomes worse as the desired output voltage is reduced) and inject more significant, complex, and lower-order harmonic content into the supply compared to a well-filtered DC-link inverter; and cycloconverters require a considerably larger number of thyristors (typically 2 to 4 times as many as an equivalent DC-link inverter of similar power rating) along with more complex, continuously time-varying firing-angle control circuitry to synthesize the desired output waveform.

Applications: cycloconverters are traditionally used for very large power, low-speed AC motor drive applications where their output-frequency limitation is not a practical constraint — large gearless mill drives (cement/ore grinding mills), ship propulsion motor drives, and some large variable-speed constant-frequency (VSCF) aircraft/marine power supply systems; with the advent of modern high-power semiconductor devices and matrix converter/multilevel converter alternatives, cycloconverters have become somewhat less dominant than in earlier decades but remain in use for some of the largest-power, lowest-speed drive applications where their direct, single-stage conversion and inherently robust thyristor technology remain economically and technically advantageous.

Output Voltage Synthesis and Harmonic Structure

The instantaneous output voltage of a single-phase to single-phase cycloconverter is obtained by cyclically varying the firing angle of the active converter bridge according to a cosine-modulation law:

where M is the output voltage ratio (analogous to a modulation index) and ωo is the desired output angular frequency; this cosine firing law ensures that the local average of the chopped supply half-cycles traces out a sinusoid at frequency ωo with peak amplitude M times the maximum available output. Because the output is assembled from segments of the supply frequency waveform rather than generated by a continuously-variable-frequency source, the actual output voltage contains, in addition to the desired fundamental at ωo, a family of unwanted harmonic and sub-harmonic components at frequencies of the form |nωs ± mωo| (n, m integers), where ωs is the supply angular frequency — these 'beat frequency' harmonics are generally non-integer multiples of the output frequency and are more difficult to filter than the well-separated, high-frequency harmonics of a PWM inverter, since some fall relatively close to the desired fundamental, particularly as the output frequency ratio ωo/ωs is increased toward its practical upper limit.

Maximum output frequency limit — derivation basis: for acceptable output waveform quality (a recognizable approximation to the desired sinusoid, with the lowest significant harmonic sufficiently separated from the fundamental), it is generally recommended that the output frequency not exceed about one-third of the supply frequency for a single-phase cycloconverter using a modest pulse number, i.e., fo(max) ≈ fs/3; this is one of the more restrictive practical limitations of cycloconverter technology and is a direct consequence of needing a sufficient number of supply half-cycles 'per output half-cycle' to adequately trace out the sinusoidal envelope with tolerable harmonic content — attempting to push the output frequency closer to fs (or beyond) causes the assembled waveform to depart drastically from a sinusoid, since too few supply segments remain available within each output half-cycle to approximate the desired shape.

Comparison with Matrix Converters

A closely related, more modern AC-AC direct-conversion alternative to the classical thyristor cycloconverter is the matrix converter, which uses an array of bidirectional (four-quadrant) semiconductor switches (built from series/anti-parallel IGBT-diode combinations) to directly connect any input phase to any output phase at any instant, under high-frequency PWM-style control rather than low-frequency phase-angle control; matrix converters can achieve output frequencies both below and above the supply frequency (unlike the cycloconverter's output-frequency ceiling), give better input current waveform quality and displacement power factor control, and eliminate the large, low-frequency circulating-current reactor needed in continuous-conduction-mode cycloconverters, but require fully controllable bidirectional switches (rather than simple natural-commutation-friendly thyristors) and more sophisticated real-time space-vector-based switching control, and have historically been adopted more slowly in the very largest power ranges where classical thyristor cycloconverters remain established.