WO2006059067A1

WO2006059067A1 - An electrosurgical generator

Info

Publication number: WO2006059067A1
Application number: PCT/GB2005/004470
Authority: WO
Inventors: Francis Amoah; Colin Charles Owen Goble
Original assignee: Gyrus Medical Limited
Priority date: 2004-12-03
Filing date: 2005-11-18
Publication date: 2006-06-08
Also published as: GB0426648D0

Abstract

An electrosurgical generator for supplying RF power to an electrosurgical instrument for cutting or vaporising tissue has an RF output stage (42) with RF output devices (44), a series-resonant output network (48) and an RF output (74). The generator offers improved cutting and vaporising performance, especially in relation to the reliability with which an arc can be struck when presented with an initial low impedance load. This is achieved by virtue of the output stage being capable of maintaining output pulses of at least 1kW peak by supplying the RF output devices from a large reservoir capacitor (60). An appropriate combination of cutting performance and haemostasis is provided by allowing for pulsed or continuous wave operation once an arc has been established, according to whether or not a surplus energy condition exists, as indicated by the voltage across the reservoir capacitor.

Description

AN ELECTROSURGICAL GENERATOR

This invention relates to an electrosurgical generator for supplying radio frequency (RF) power to an electrosurgical instrument, and primarily to a generator having a series-resonant output network.

Conventionally, electrosurgical generators make use of a configuration comprising a voltage source coupled to an electrosurgical instrument via a coupling capacitor which defines a matched output impedance between 50 and 500 ohms. Such a configuration produces a power- versus-load impedance characteristic having a power maximum at a matched impedance, with power falling off progressively on each side of this peak. In practice, when conducting electrosurgery, the load impedance can change over a very wide range, resulting in unpredictable clinical effects.

To deal with this problem, it is known to provide an RF output stage capable of providing an impedance match over a wide range. This has the disadvantage that* 7^^ψ rapid load impedance changes can produce large output voltage excursions. An alternative approach is to control the DC supply to the RF output stage in response to feedback signals in order that the delivered power is virtually continuous. This may be done by adjusting the power supply DC voltage or by maintaining the supplied DC power constant. These techniques lead to a power versus load impedance characteristic which is virtually flat over a range of impedances, but one limitation is that it is difficult to control the delivery of energy when initiating tissue cutting or vaporisation (as opposed to tissue coagulation). To cut or vaporise tissue using radio frequency power, the initial low impedance load presented by the tissue or surrounding fluid needs to be brought to a higher impedance in order to strike an arc. Delivering too much energy can result in burns adjacent the operative site, excessive smoke, or instrument failure. Delivering too little energy causes a significant delay and can result in unwanted tissue coagulation.

It is also known to use an electrosurgical generator to supply a bipolar electrosurgical instrument with pulsed electrosurgical power at very high voltages, e.g. in the region of 1 kilovolt peak-to-peak when removing tissue at an operation site immersed in a conductive liquid, such as saline. The instrument may have an active electrode located at its extreme end to be brought adjacent to or into contact with tissue to be treated, and a return electrode set back from the active electrode and having a fluid contact surface for making an electrical connection with the conductive liquid. To achieve tissue removal, the conductive liquid surrounding the active electrode is vaporised to cause arcing at the electrode. The high voltages used to achieve tissue cutting or vaporisation under varying load impedance conditions are particularly demanding of the generator when the instrument experiences a low load impedance. Indeed, as stated above, under such conditions it is difficult reliably to initiate arcing without unwanted effects. Steps have been taken to increase power density at the active electrode and, hence, improve the reliability with which arcing is started, by reducing the size of the electrode and by roughening its surface, e.g. by applying an oxide layer. The latter technique has the effect of trapping vapour in the irregularities in the surface as a means of increasing power density.

It has been found that operation of such instruments at high voltages tends to cause erosion of the active electrode. The rate of erosion increases as the supply voltage is increased, and is also exacerbated by reducing the size of the electrode and providing a roughened surface, as just mentioned.

Published European Patent Application No. EP1053720A1 discloses a generator for generating high electrosurgical voltages.

The applicant's published International Application No. WO2004/062516A1 discloses an electrosurgical generator for supplying RF power to an electrosurgical instrument, the generator comprising an RF output stage having a least one RF power device, at least one pair of output lines for delivering RF power to the instrument, and a series-resonant output network coupled between the RF power device and the said pair of output lines. In this generator, the output impedance of the output stage at the output lines is less than 2OθHP ohms, where P is the maximum continuous RF output power of the generator in watts. When the generator is configured for wet field surgery, e.g. for use with the electrode or electrodes of the instrument immersed in a conductive fluid such as saline, the maximum continuous power is preferably in the region of 300W to 400W. Accordingly, if the maximum output power is 400W, the output impedance is less than 10 ohms. Dry field electrosurgery, i.e. with the electrode or electrodes not normally immersed, requires less RF output power. In this case, the generator may be configured such that the maximum continuous RF output power is in the region of 16W, in which case that the output impedance is then less than 50 ohms. In both such cases, the figures are obtained when operating with an output voltage for cutting or vaporising tissue, i.e. at least 300V peak. The output impedance is preferably less than IOONP ohms, which yields maximum output impedance values of 5 ohms and 25 ohms at the above power outputs.

When RF energy is supplied to a load in bursts, generally as an RF sine wave, the maximum continuous power is the average power measured over several such bursts.

In this generator the series-resonant output network is configured such that the maximum rate of rise of the output current at the output lines is less than (VP)/4 amps per microsecond, P being as defined above. Accordingly, for a typical maximum continuous RF output power of 400W for wet field electrosurgery, the maximum rate of rise of the output current amplitude, generally when the output lines are short- circuited at the maximum power setting of the generator, is less than 5A/μs. With P at a typical value of 16W for dry field electrosurgery, the rate of rise of the output current amplitude is less than 1 A/μs.

To maintain the constant peak output voltage at low impedances, the power supply to the output stage includes a charge-storing element, preferably a capacitance in excess of ImF, the or each output device being pulsed by a pulsing circuit so that it supplies RF energy in bursts with the timing of the bursts, particularly the termination of each burst, being controlled in response to the output of a voltage sensing circuit coupled to the capacitance. The DC power supply voltage to the output stage is typically 100V or greater. To avoid substantial decay of the supply voltage, the voltage sensing and pulsing circuits are arranged to terminate the individual pulses of RF energy when the sensed voltage falls below a predetermined level, typically set such that pulse termination occurs when the voltage falls by a predetermined percentage value of between 5 percent and 30 percent which, typically, corresponds to the peak RF voltage delivered at the output lines falling to a value between 25V and 150V below its starting value for the respective pulse. The RF energy delivered during each pulse is typically 60 joules for wet field electrosurgery and 2 joules for dry field electrosurgery. Peak power typically reaches at least IkW₅ and preferably 4kW.

The very high peak power capability of the preferred wet field generator (in excess of IkW) allows the impedance transition occurring at the start of a tissue cutting or vaporisation cycle to be completed very quickly since only voltages in excess of those required for arcing are delivered. This significantly reduces the delay and the unwanted coagulation effects of some earlier generators. The substantially constant voltage delivery leads to cutting or vaporisation occurring at consistent rates, regardless of changes in tissue type or engagement.

The generator is particularly useful for use with large electrodes, e.g. with a bipolar electrosurgical instrument having an active electrode formed as a conductive U- shaped loop. Such a loop may be used for excising tissue samples in a TURP (transurethral resection of the prostate) procedure. This electrode places particular demands on the generator in terms of achieving saline vaporisation and arcing.

It has been found that the combination of the above generator and the loop electrode is effective for resecting or vaporizing tissue with a haemostatic effect. The generator can also be set to coagulate tissue via haemostasis. The ability to achieve substantially instant generation of a plasma enveloping the active electrode gives surgeons the confidence of the state of the instrument before tissue engagement and facilitates vaporisation of fronds of tissue which typically result from resection using a loop electrode. Experience has shown, however, that there are variations in efficiency with which surgeons are able to complete a typical TURP procedure. At the end of this procedure, a low level of bleeding is thought to be desirable. Should bleeding be unacceptably high, the procedure is typically completed by identifying the points of bleeding and applying a specifically coagulating waveform to the instrument using the same loop electrode. The variations in efficiency are thought to be due to anatomical variations and differing surgical techniques. It would be desirable to reduce such variations. Accordingly, the present invention provides an electrosurgical generator for supplying radio frequency (RF) power to an electrosurgical instrument for cutting or vaporising tissue, wherein the generator comprises: an RF output stage at least one RF power device, at least one pair of output lines for delivering RF power to the instrument, and a series-resonant output network coupled between the RF power device and the said pair of output lines; a power supply stage coupled to the RF output stage to supply power thereto and having a sensing element for sensing energy levels in the power supply stage; and a control circuit coupled to the sensing element for interrupting the delivery of RF power by the RF output stage, the control circuit being configured to interrupt the power delivery whilst the energy level in the power supply stage exceeds an interruption energy threshold and to cause continuous wave operation of the RF output stage when the energy level is below the threshold. Interrupting the RF output in this way when the generator is in use, connected to an instrument having an electrode such as that referred to above, increases the dwell time of the electrode during tissue resection. In other words, the resistance to movement of the electrode through the tissue is increased. This results in greater heat transfer and cauterisation of tissue adjacent the loop. In this way, haemostasis can be increased without unacceptably reducing the cutting efficacy of the system.

In practice, interrupting the RF output causes the plasma around the electrode to collapse, which, with the preferred generator, increases power dissipation in tissue and conductive liquids adjacent the electrode. By arranging for interruptions to occur only above a given power supply energy threshold, sufficient energy is available to re- establish the plasma at the end of each interruption. Preferably, control circuit is in the form of a pulsing circuit configured to apply the interruptions repeatedly in the form of pulse modulation of the RF output stage. Typically, the OFF periods of the pulse modulation are less than 10ms in duration so that the interruptions are brief enough to be below the thermal time constant of the electrode and its surroundings and so that the electrode creates a consistent surgical effect during a resection/through tissue (i.e., substantially to avoid a corrugated cut surface).

OFF periods of 0.5ms to 1.5ms are preferred. Advantageously the pulsing circuit is configured to cause pulse modulation of the RF output stage with a mark-to-space ratio of between 1:1 and 1:10, preferably 1:2.5 to 1 :5. The pulse ON periods typically have a duration in the range of from 0.5ms to 5ms.

There is significant latitude in the control process, so long as there is sufficient energy to reestablish the plasma, hence the use of a given power supply energy threshold below which this kind of pulsed operation is disabled. The steady state condition in a typical resection stroke in a TURP procedure is preferably a power supply energy level above the threshold relative to generator default settings, the energy reservoir always recovering towards the defined threshold if interruptions have temporarily depleted the energy reservoir. The energy threshold is defined so as to be towards the maximum energy storage capability of the generator to allow reestablishment of the plasma even in a cold environment.

The pulsing circuit of the preferred generator includes a timer which defines the pulse modulation mark-to-space ratio as well as the pulse repetition rate, which may be, typically, in the region of from 100Hz to IkHz.

Preferably, the power supply energy level threshold below which such predetermined pulsed operation is switched off is set at a value which is between 2 and 20 percent below a predetermined maximum power supply energy level. A value less than 16 percent below the maximum energy level is particularly preferred.

The pulsing circuit may also be configured such that delivery of RF power by the RF output stage is substantially terminated when the power supply energy level falls below a second, lower, energy threshold. Thus, the extinguishing of the RF output in low load impedance conditions can be implemented by means of a function which overrides the above-described pattern of operation combining pulse modulation and continuous wave (CW) operation. In effect, the mode of operation described in the above-referenced International Application No. WO2004/062516 takes precedence over the determination of whether the RF output power should be pulsed or continuous. This overriding function is preferably achieved by terminating the delivery of RF power when the power supply energy level falls below a minimum energy threshold with a value of between 5 and 50 percent below the maximum power supply energy threshold and is recommenced when the power supply energy level regains the maximum energy threshold.

It will be understood that the delivery of RF power will be reestablished either with pulse modulation or as a CW waveform according to the impedance presented across the generator output. In the extreme case, if the electrode is surrounded by conductive liquid, such as saline, the power supply energy level will drop rapidly, resulting in a period of CW operation as a plasma is reestablished. If, then, the electrode is brought into contact with tissue to perform tissue resection, for instance, the load impedance rises and, typically, power supply energy levels rise to the extent that operation continues with pulse modulation.

According to another aspect of the invention, there is provided an electrosurgical generator for supplying radio frequency (RF) power to an electrosurgical instrument for cutting or vaporizing tissue, wherein the generator comprises an RF output stage having at least one RF power device, and at least one pair of output lines for delivering RF power to the instrument, and a power supply stage which is coupled to the RF output stage to supply power thereto and which has a sensing element for dynamically sensing energy levels in the power supply stage, the generator being capable of delivering RF power at the output lines with a peak current of at least IA and a simultaneous peak voltage of at least 300V, and wherein the generator further comprises a control circuit coupled to the sensing element to receive an energy level sensing signal from the sensing element and to cause interruptions in the delivery of RF power when the energy level sensing signal is indicative of the power supply energy level being in an upper energy region and to cause continuous wave supply of RF power when the energy level sensing signal is indicative of the power supply energy level being in a lower energy region.

In a preferred generator in accordance with the invention, there is protection circuitry responsive to a predetermined electrical condition indicative of an output current overload, e.g. due to short-circuiting of the output lines, substantially to interrupt the RF power supplied to the series-resonant output network. The protection circuitry is responsive to short-circuiting with sufficient speed that the supply of RF power to the output network is cut off within a time period corresponding to no more than 20 cycles of the delivered RF power. The protection circuitry is preferably much faster, e.g. being operable to interrupt power delivery within 3 cycles or even 1 cycle. The effect of the series-resonant output network is to delay the build-up of current in a fault condition such as when a very low impedance or a short circuit appears across the output lines. The applicants have found that an impedance transition from open to short circuit results in an effective short circuit across the RF power device only after several RF cycles. By arranging for the protection circuitry to respond quickly, the output stage can be disabled before that happens. In general, the protection circuitry is responsive to application of a short-circuit across the output lines sufficiently quickly to disable the RF power device before the current passing therethrough rises to a rated maximum current as a result of the short-circuit.

The use of an RF output stage with a relatively low output impedance means that the RF voltage output is substantially directly related to the DC supply voltage applied to the output stage (specifically to the RF power device or devices which it contains). In the preferred embodiment of the invention, each RF power device is operated in a switching mode with the result that a square wave output is applied to the series- resonant output network. The RMS voltage available at the output lines is directly proportional to the supply voltage. It follows that the maximum peak-to-peak output voltage is determined by the DC supply voltage and dynamic feedback to control output voltage is, as a result, not required in this embodiment.

The protection circuitry is preferably capable of disabling the output stage within one- and-a-half RF periods after onset of the above-mentioned predetermined electrical condition. Preferably, the predetermined electrical condition is indicative of an instantaneous current in the output stage exceeding a predetermined level, and the speed of response of the protection circuitry is such that the breaching of the predetermined level by the instantaneous current is detected during the same RF cycle that it occurs. Such detection may be performed by current sensing circuitry including a pick-up arrangement, which is typically a current transformer, coupled in series between the RF power device or devices and the series-resonant output network, and a comparator having a first input coupled to the pick-up arrangement (e.g. to the secondary winding of the transformer) and a second input coupled to a reference level source. The reference level source may be a voltage representation of the instantaneous current, i.e. substantially without filtering, in order to cause a change of state of the comparator output within the same RF half-cycle that the threshold is first exceeded, or within the subsequent half-cycle, depending on whether or not full wave rectification is applied ahead of the comparator. The predetermined instantaneous output level is preferably at least 5A for wet field electrosurgery, and typically 2OA. The output of the comparator is coupled to disabling circuitry to disable the power device or devices when the comparator output changes state in response to the instantaneous current sensed by the pick-up arrangement exceeding the predetermined level as set by the reference source. The current shut-down aspect of the protection circuitry is not limited by impedance.

Generally, it is necessary only to interrupt power delivery for a short time. Consequently the protection circuitry includes a monostable stage and is operable in response to detection of the predetermined condition to disable the power device for a limited period determined by a time constant of the monostable stage which is typically less than 20 cycles of the operating frequency of the generator.

Preferably, the generator has an RF source coupled to the power device or devices, the source including an oscillator defining the operating frequency of the generator. The series resonant output network is tuned to this operating frequency. Generally, the source is arranged such that the operating frequency is substantially constant (e.g. during any given treatment cycle).

The preferred generator is arranged such that, for a given user setting, the RMS RF output voltage is maintained substantially constant within a load impedance range of from 600/VP ohms to 1000 ohms, where P is as defined above. Thus, for instance, the RMS RF output voltage during each burst of RF energy is maintained to within 20 percent of a maximum value. This can be achieved partly as a result of the series- resonant configuration of the output network.

A further aspect of the generator is that it comprises protection circuitry responsive to a short circuit across the output lines, wherein the output impedance of the RF output stage is less than 2OθNP ohms, where P is the maximum continuous RF output power of the generator in watts, and wherein the protection circuitry is responsive to the said short circuit sufficiently quickly to disable the power device before the current passing therethrough rises to a rated maximum current as a result of the short circuit. The or each power device may be disabled in response to application of the short- circuit to the output lines in a time period corresponding to less than three RF cycles.

In a generator in accordance with the invention for supplying radio frequency (RF) power to an electrosurgical instrument for cutting or vaporising tissue in dry field electrosurgery, the output impedance of the RF output stage at the output lines is less than 50 ohms.

The generator is preferably configured to be capable of maintaining a peak output voltage of at least 300V over a load impedance range of from 6OθNP ohms to 1000 ohms, where P is the rated output power in watts. The rated output power is as defined in the International Electrotechnical Commission standard, IEC 60601-2-2.

In the preferred generator, the power supply stage has an energy storage capacitor capable of storing between 3 percent and 30 percent of the maximum continuous power P (in watts) of the generator in joules.

In another aspect, the energy delivery per pulse (in joules) is between 1 percent and 15 percent of the maximum continuous RF output power (in watts).

When terminating and re-establishing the RF output according to stored energy levels (at the minimum and maximum energy thresholds respectively in the preferred embodiment), as may happen when operating with low load impedances, the pulsing circuit may pulse the delivered RF power in such a way that the crest factor of the voltage developed across the output lines increases as the load impedance presented to the output lines decreases whilst the peak output voltage during pulses is maintained at a value greater than 300V. For wet field electrosurgery, the output impedance of the output stage is preferably less than 10 ohms and the crest factor varies by a ratio of at least 2:1 over a load impedance range of from 6OθNP to 1000 ohms (typically from 10 ohms to 1000 ohms). For dry field electrosurgery, the output impedance figure is less than 50 ohms, and the crest factor varies by a ratio of at least

2:1 over a load impedance range of 600/VP to 50 kilohms (typically from 50 ohms to 50 kilohms).

By "crest factor" we mean the ratio of the peak voltage to the RMS voltage. In the case of a pulsed output waveform, the measurement is conducted over plurality of pulses.

At the lowest load impedances defined above, the length of such pulses is preferably between 0.5ms and 5ms, the pulse duty cycle typically being between 1% and 20% and, more preferably, between 2% and 10%.

The preferred generator comprises a source of radio frequency (RF) energy, an active output terminal, a return output terminal, a DC isolation capacitance between the source and the active output terminal, and a pulsing circuit for the source, wherein the source and the pulsing circuit are arranged to generate a pulsed RF output signal at the output terminals, which signal has a peak current of at least IA, a simultaneous peak voltage of at least 300V, a modulation rate of between 5Hz and 2kHz, and a pulse length of between lOOμs and 20ms. In preferred embodiments of the invention, the signal has a peak current of at least 3 A.

With such a generator it is possible to start arcing even under conditions of relatively low load impedance. Once an arc is established, the load impedance tends to rise, to the extent that the arcing can be maintained using a continuous RF output waveform. The preferred generator in accordance with the invention has a resonant output network and is operable to generate, e.g. during at least an initial part of a treatment period, a peak power of at least one kilowatt, and typically at least 3 or 4 kilowatts. Improvements in electrode erosion performance can be achieved by providing means in the generator for limiting the output voltage to a value in the region of 700V to 1100 V peak-to-peak.

In one embodiment, the source and the pulsing circuit are arranged to generate, in an initial period, a pulsed RF output signal at the output terminals, which signal has a peak current of at least IA, a simultaneous peak voltage of at least 300V, a modulation rate of between 5Hz and 2kHz, and a pulse length of between lOOμs and 20ms, and, in a subsequent period, to generate a constant power RF output signal at the output terminals.

Different ways of causing the generator to end the above-mentioned initial period of operation and begin the so-called subsequent period are feasible. One generator embodiment is arranged such that the switchover from the initial period to the subsequent period occurs automatically at a predetermined time interval after the beginning of the initial period. In an alternative embodiment, the generator has means for monitoring, in use of the generator, the load impedance between the active and return output terminals, and is arranged to cause switchover to the subsequent period when the magnitude of the output impedance increases by a predetermined factor, typically between 5 and 20, and preferably 10, or when it exceeds a predefined threshold.

A third switching-over technique may be used, involving the charge-storing element mentioned above. In this case, the source of RF energy includes an RF output stage, and the generator has a power supply including the charge-storing element such as a large capacitor for supplying power to the output stage. When the treatment period includes an initial period and a subsequent period, as described above, the capacitor is used to supply power at least during the initial period. Associated with the charge- storing element is a voltage-sensing circuit for sensing the voltage supplied to the output stage by the charge-storing element, the generator being arranged such that treatment ends or the subsequent period begins in response to the supply voltage as sensed by the voltage-sensing circuit reaching a predetermined voltage threshold. Indeed, it is possible to control the length and timing of individual pulses using the same voltage-sensing circuit. In this case, the voltage-sensing circuit forms part of the above-mentioned pulsing circuit and the timing of at least the beginnings of the pulses produced by the output stage during the initial period being determined in response to the supply voltage reaching the above-mentioned voltage threshold. It is possible to arrange for both the leading and trailing edges of the pulses produced by the output stage to be determined by the supply voltage respectively falling below and exceeding the respective voltage thresholds. The charge-storing capacitance is preferably at least lOOOμF and advantageously has a capacity in excess of 5 J.

As already stated, the preferred generator has a tuned output. Indeed, good results have been obtained using a generator with a resonant output network, the load curve of the generator (i.e. the curve plotting delivered power versus load impedance) having a peak at a load impedance below 50 ohms. Delivery of peak power levels into low load impedances is aided by forming the output network as a series-resonant network comprising the series combination of an inductance and a capacitance, the output of the network being taken across the capacitance. The output may be taken to all output terminals of the generator via a coupling capacitor and, optionally, a step-up transformer from a node between the inductance and the capacitance of the series combination. Whilst it is possible, instead, to take the output from across the inductance, taking it across the capacitor has the advantage of reducing switching transients. As a further alternative, the generator may have its output terminals connected to the resonant output network so that, effectively, when a load is connected to the terminals it is connected as an impedance in series with the inductance and capacitance forming the resonant combination, e.g. between the inductance and the capacitance.

The resonant output network typically provides a source impedance at the output terminals in the range of from 50 ohms to 500 ohms.

Not least because the resonant frequency of the output network can vary with load impedance as a result of coupling capacitance, the RF source may include a variable frequency RF oscillator, the output frequency advantageously being limited to a maximum value below the resonant frequency of the output network when connected to a matching load impedance, i.e. a load impedance equal to its source impedance.

The invention will now be described by way of example with reference to the drawings in which:- Figure 1 is a general diagram showing an electrosurgery system including a generator in accordance with the invention and a bipolar electrosurgical instrument;

Figures 2A and 2B are respectively perspective and side views of a loop electrode assembly forming part of the bipolar instrument shown in Figure 1 ;

Figure 3 is a block diagram illustrating the main components of the generator;

Figure 4 is a simplified circuit diagram of an RF output stage forming part of the generator;

Figure 5 is an illustrative load curve for the generator of Figure 1;

Figure 6 is a more detailed circuit diagram of the RF output stage;

Figure 7 is a flow diagram showing a control process for pulsed operation of the generator;

Figures 8A and 8B are waveform diagrams illustrating operation of the generator when activated with a loop electrode in a wet field;

Figures 9 A and 9B are waveform diagrams illustrating operation of the generator with the electrode in tissue;

Figures 1OA and 1OB are waveform diagrams illustrating operation of the generator with the electrode in a low impedance tissue;

Figure 11 is a block diagram of an alternative electrosurgical generator in accordance with the invention;

Figure 12 is a circuit diagram of a resonant output network of the alternative generator; and

Figure 13 is the load curve of the generator of Figure 11. Referring to Figure 1, a generator 10 has an output socket 1OS providing a radio frequency (RF) output for an electrosurgical instrument in the form of an endoscope attachment 12 via a connection cord 14. Activation of the generator may be performed from the instrument 12 via a control connection in cord 14 or by means of a footswitch unit 16, as shown, connected separately to the rear of the generator 10 by a footswitch connection cord 18. In the illustrated embodiment, the footswitch unit 16 has two footswitches 16A and 16B for selecting a coagulation mode and a cutting mode of the generator respectively. The generator front panel has push buttons 20 and 22 for respectively setting coagulation and cutting power levels, which are indicated in a display 24. Push buttons 26 are provided as alternative means for selection between coagulation and cutting modes. The instrument 12 has a detachable loop electrode assembly 28 with a dual electrode structure and is intended for use in a saline field.

The instrument 12 has a detachable loop electrode assembly 28 with a dual electrode structure and intended for use in a saline field. Figures 2A and 2B are enlarged views of the distal end of the electrode assembly 28. At its extreme distal end the assembly has a U-shaped loop electrode 30 depending from a pair of electrode assembly arms 32 which are mounted side-by-side in a clip 34 intended to be snapped onto an endoscope. The loop electrode 30 is an active electrode. Each of the arms 32 is formed as a coaxial cable, the exposed conductive outer shield of which, in each case, forms a return electrode 36. In operation immersed in a saline field, the loop electrode 30 is typically used for excising tissue samples, the electrosurgical voltage developed between the loop electrode 30 and fluid contacting surfaces of the return electrodes 36 promoting vaporisation of the surrounding saline liquid at the loop electrode 30, and arcing through the vapour envelope so formed.

The loop electrode 30 comprises a composite molybdenum rhenium wire with an oxide coating to promote increased impedance in the electrode/fluid interface and, as a result, to increase power density at the surface of the electrode.

The width of the loop is typically in the range of 2.5mm to 4mm and the wire typically has a diameter in the range of 0.20 to 0.35mm. This loop electrode assembly places particular demands on the generator in terms of starting vaporisation and arc formation.

Efforts to improve the starting of the arc (the "firing up") of this electrode assembly by reducing the wire diameter and forming oxide layers have tended to increase the rate of erosion or resulted in the loop being mechanically flimsy.

It should be noted that generators in accordance with the invention are not limited to use with a loop electrode assembly, nor to use in wet field surgery.

The generator will now be described in more detail with reference to Figure 3. It has an RF source in the form of an oscillator 40 which is connectible to an RF output stage 42. The output stage 42 comprises a mosfet power bridge forming part of a power mosfet and driver circuit 44, a current sensing element 46 and a resonant output network 48. The oscillator 40 is configured to operate at a substantially constant RF frequency and the output network 48 is tuned to that frequency. In general terms, the RF source coupled to the RF power device or devices defines the operating frequency of the generator, and the output network (which, as will be described below, is series- resonant) is tuned to the operating frequency. In this embodiment of the invention the operating frequency is substantially constant.

Power to the RF output stage 42, or, more specifically, to the power mosfets, is supplied from a power supply stage 49 which has a DC power supply 50 coupled to the RF output stage 42 via a supply rail 58. The power supply stage 49 has a 6mF reservoir capacitor 60 connected between the supply rail 58 and ground. The voltage on the supply rail 58 is sensed by a voltage sensing circuit 62, the output of which is fed to a pulsing circuit 63 which has a first transmission gate 64 connected in series between the RF oscillator 40 and driver devices in the power mosfet and driver circuit 44. The pulsing circuit 63 has a data processor 65 which controls operation of the gate 64 according to voltages sensed by the sensing circuit 62.

The current sensing element 46 in the output stage 42 is a series-connected current transformer, the secondary winding of which is coupled to a first input of a comparator 66 which also receives on the other of its inputs a reference signal from a reference input 68. The output of the comparator controls a monostable 70 which, in turn, controls a second transmission gate 72 coupled in series with the gate 64 in the path between the oscillator 40 and the drivers in the power mosfet and driver circuit 44. The output network 48 supplies RF power to an output termination 74 which, in practice, is a pair of output lines, as will be described hereinafter. Operation of the generator may be pulsed insofar as RF energy may be supplied to the output lines 74 in bursts controlled by the combination of the processor 65, the voltage sensing circuit 62 and the gate 64. When the generator is activated, the voltage on the supply rail 58 tends to fall, at least when the load impedance coupled across output lines 74 is relatively low, owing to the discharge of reservoir capacitor 60. When the DC supply voltage on the supply rail 58 falls to a preset value, the output of the voltage sensing circuit 62 changes state and transmission gate 64 is driven to its open circuit condition, thereby disabling the power mosfet and driver circuit 44. The reservoir capacitor 60 then recharges and the voltage sensing circuit 62 and the processor 65 cause the gate 64 to reconnect the oscillator 40 when the supply rail voltage reaches an upper threshold value. In this way it is possible to control the amount of energy delivered in each pulse in low load impedance conditions.

The generator also has a second pulsed mode of operation in which the RF output is interrupted regularly when the DC supply voltage exceeds an intermediate threshold value. For this purpose, the processor 65 is configured to include a pulse timer 65 A.

Further details of the operation of the pulsing circuit will be provided below.

Referring again, however, to Figure 3, the current sensing element 46, the comparator 66, the monostable 70 and the second transmission gate 72 act together as a protection circuit to protect the mosfet power devices in circuit 44 against excessive current drain caused, for instance, by a short circuit across the output lines 74. The storage of energy in output network 48 delays the transfer of the short circuit across the output lines 74 to the power devices in the mosfet and driver circuit 44.

The electrical circuit condition sensed by the current sensing element 46 and the comparator 66 is the current flowing between the power mosfets in circuit 44 and the output network 48 rising to a level which could be indicative of a short circuit having been applied across the output lines 74. When the current reaches a preset current level, as detected by the comparator 66, the comparator output changes state and the monostable 70 causes the second transmission gate 72 to become open circuit, disabling the power mosfets and driver stage 44. The monostable time constant is typically set to 0.5 seconds or less, which allows a warning signal to be generated for alerting the user. However, owing to energy storage in the series-resonant circuit, it is possible to protect the RF power devices with a monostable time constant of about 20 RF cycles at an operating frequency of 40OkHz.

The configuration of the output stage 42 is shown in principle in the simplified circuit diagram of Figure 4. Referring to Figure 4, the power mosfet and driver stage 44 shown in Figure 3 has a power mosfet bridge comprising a first push-pull pair of FET power devices Ql₅ Q2 and a second power FET device push-pull pair Q3, Q4, each pair having a respective output node which, when the pairs are driven 180° out of phase, produces a square wave at the frequency of the oscillator 40 (Figure 2) at the input to the series resonant output network 48. Each pair of power mosfets Ql, Q2; Q3, Q4 is coupled between the supply rail 58 and ground. Accordingly, since each of the mosfets is a virtual short circuit when driven "on", the voltage applied to the output network 48 swings virtually between ground and the supply rail voltages. The reservoir capacitor 60 shown in Figure 3 is, of course, connected across the respective power mosfet pairs, as shown in Figure 4.

The output network is series-resonant in that an inductor Ll and a resonating capacitor Cl are coupled in series between the output nodes 76, 78 of the first and second power mosfet pairs respectively. In this embodiment, the load resistance RL constituted in practice by an electrosurgical instrument coupled between the output lines 74, and the tissue and/or fluid present across its electrode assembly, is connected in series between inductor Ll and capacitor Cl. As explained above, the series- resonant tuned circuit formed by inductor Ll and capacitor Cl acts as an energy storing device which delays the current build-up between the nodes of the power mosfet bridge Q1-Q4 should the load resistance RL drop to a very low value. Another feature of this resonant arrangement is that it is a low impedance at one frequency only, which means that the delivered output signal consists almost exclusively of the fundamental component of the waveform produced by the power mosfets, conditional, of course, upon the frequency of resonance of the network 48 being the same as that of the operating frequency of the oscillator stage 40 (Figure 3).

One of the characteristics given to the generator by the output configuration described above with reference to Figure 4 is that, during each burst or pulse of RF energy it has an approximately constant voltage load curve, as shown by the power-versus-load impedance load curve shown in Figure 5. This characteristic is particularly suitable for cutting or vaporisation of tissue since it provides the high power required at low impedance without voltage overshoot. The low output impedance and high current required are provided by the direct coupling of the power mosfets to the supply rail and ground, and by the reservoir capacitor 60, even if a step-up transformer is coupled between the series-resonant elements Ll, Cl and the output lines 74. It is possible, using this configuration, to keep the output impedance of the generator at the output lines 74 to 2 ohms or less. The implications which this has for peak current delivery in a fault condition leads to the need for a protection circuit such as that referred to above.

The RF output stage 42 is shown in more detail in Figure 6. As shown in Figure 6, the current sensing element 46 is a current transformer, coupled in series between one of the output nodes 76, 78 of the power mosfet bridge and one of the components Ll, Cl of the series resonant output network, in this case between node 76 and the inductor Ll. In this preferred generator, the normal DC supply voltage on supply rail 58 is about 120V. To strike an arc for the purpose of performing tissue cutting or vaporisation, a peak voltage in excess of 380V may be required. Accordingly, and for isolation purposes, the RF output network 48 includes a step-up isolating transformer TRl to lift the peak output voltage to the region of 500V peak. The primary winding of the transformer TRl has a tuning capacitor C2 coupled across it to yield a parallel- resonant circuit tuned to the operating frequency to act as a shunt-connected trap. This improves the rejection of harmonics in the power signal supplied to the output lines 74., particularly when the output impedance is high, with the consequent benefit of reduced RFI (RF interference). DC blocking is provided by a coupling capacitor C3 between the transformer TRl secondary winding and one of the output lines 74.

The actual resonant frequency of the output network 48 is the result of several elements, these being (1) the main tuning elements represented by the lumped inductance Ll and the tuning capacitor Cl, (2) the transformer leakage inductance and cross-coupling capacitance, (3) the DC blocking capacitance, C3, and (4) the inductive and capacitive loading of the connecting cable (not shown) between the output lines 74 and the electrosurgical instrument itself. The delay in the current build-up in a fault condition is due to the energy levels in this tuned arrangement. At resonance, this arrangement has a finite loss that may be represented by a series resistance, albeit a very small one. Dynamically, however, the energy levels in the resonant output network cannot be changed instantly. An impedance transition from an open to short circuit only presents a short circuit to the switching stage after several RF cycles at the operating frequency. The comparator 66 shown in Figure 3 is capable of detecting such an impedance transition within 1 to 1.5 cycles of the transition beginning at the output lines 74. This rapid response, as well as allowing the power mosfet and driver circuit 44 to be shut down before damage occurs, has the effect that the amount of energy delivered during a short circuit fault is very small.

Referring again to Figure 3 and, in particular, the voltage sensing and output stage pulsing circuits 62, 64, the very high peak powers which are achievable with the output stage described above with reference to Figures 5 and 6 have the benefit that, during power delivery into a low impedance, the voltage across the reservoir capacitor 60 decreases progressively after the instant of generator activation. The capacitor value is chosen to be sufficiently large to ensure that the low to high load impedance transition occurring at the start of a tissue cutting or vaporisation cycle can be produced in a single burst of RF energy. Typically, the amount of energy delivered during the initial burst is about 1 joule in a dry environment and between 10 to 20 joules in a wet field environment. The actual energy in the RF pulses or bursts is controlled by the threshold or thresholds set in the voltage sensing circuit 62, specifically by the difference in supply voltage between pulse initiation and pulse termination. Since the output stage has a very low output impedance, this difference voltage is apparent as a change in delivered RF voltage at the output. The capacitor 60 is, therefore, made sufficiently large (in this embodiment 6mF) that the change in voltage represents only a minor proportion of the absolute voltage at the output. Thus, if the delivered output voltage is a sine wave with a peak voltage of 500V, the supply voltage on supply rail 58, the size of the capacitor 60 and the transformer TRl step-up ratio are chosen such that the output voltage drops by no more than 150V peak (30 percent) during an RF burst. In this preferred embodiment, the output voltage drop is about 125 V peak or 25 percent.

One of the effects of preventing the supply of lower voltages to the output is that, in a tissue cutting or vaporisation tissue cycle, the voltage is not allowed to drop to a level at which excessive coagulation effects occur.

The preferred generator in accordance with the present invention allows the DC energy fed to the reservoir capacitor 60 when the output impedance is low to be altered so that the time period during which a cutting voltage is present at the output can be altered. In practice, owing to the low output impedance of the generator, this time period is directly proportional to the stored energy.

The control methodology, whereby RF energy bursts or pulses are controlled according to voltage thresholds sensed across a reservoir capacitor, allows very low duty cycles to be used, permitting tissue cutting or vaporisation at low average powers. Indeed, it is possible to operate with less than 5 watts average power (averaged over several capacitor charging and discharging cycles). Accordingly, the generator has uses in low power as well as high power applications.

Operation of the pulsing circuit will now be described in more detail with reference to the flow diagram of Figure 7 and the waveform diagrams of Figures 8 A and 8B, 9 A and 9B, 1OA and 1OB.

Consider, firstly, activation of the generator with the loop electrode assembly adjacent tissue and immersed in a conductive liquid such as saline. Referring to Figure 7, when the "cut" footswitch is pressed (SlOO) the voltage on the capacitor 60 (Figure 3) is initially zero, i.e. below the upper voltage threshold which, in this example, is 120V, this condition being sensed in comparison step S 102. The processor 65 (Figure 3) causes the DC power supply 50 to be switched on (S 104). There follows another comparison step S 106 in which the capacitor voltage is compared with an intermediate voltage threshold, here 115V. Initially, the result of this comparison is negative; the timer 65 A (Figure 3) is therefore reset (S 108) and another comparison (SI lO) is performed. In this step, the capacitor voltage is compared against a lower voltage threshold of 90V. Initially, of course, the voltage is below 90V, which means that the operation passes to the step Sl 12. The function of turning off the RF output is inapplicable at this stage since it was not turned on previously. Thereafter, operation reverts to step SlOO and the above steps are repeated until (a) the capacitor voltage exceeds the lower threshold (Sl 10) and (b) the capacitor voltage subsequently exceeds the higher threshold at step S 102. The repetition rate of this sequence of steps is higher than IkHz and preferably about 10kHz, so that the capacitor voltage is sensed at least once every 100 microseconds. As soon as the higher threshold is exceeded, operation passes to step Sl 14 in which the DC power supply is switched off and an "RF ready" flag is set.

Since, now, the capacitor voltage exceeds the intermediate threshold and the RF ready flag is set, operation passes from step S 106 to a pulse timer branch Sl 16, Sl 18, Sl 20, S 122, S 124 in which the pulse timer 65A (Figure 3) is enabled and gate 64 (Figure 3) is closed to turn the RF output ON.

The above-described sequence of events has the effects shown in the left hand parts of Figures 8A and 8B where it will be seen that, initially, low RF power is delivered to the electrode assembly (Figure 8A) and the capacitor voltage rises from zero to 120V (Figure 8B). Note that, owing to the electrode assembly being immersed in conductive fluid, the load impedance is low, typically between 10 and 20 ohms. As soon as the capacitor voltage reaches the upper threshold of 120V, the RF output is switched on and RF energy is supplied at very higher power, typically in excess of 3kW, as shown at point a in Figure 8A.

As a result of the high power, capacitor 60 is discharged rapidly, as shown by the steep negative slope from point a in Figure 8B. The result of this voltage drop on the operation of the processor according to the flow diagram of Figure 7 is that at the next instance of step S 102, the flow passes directly through step S 106 to step Sl 12 which results in gate 64 being opened and the RF output being terminated, as shown at point b in Figure 8A.

At point a, the DC power supply is toggled off (step Sl 14) briefly and then toggled back on (step 104) and, as soon as the RF output has been turned off in step S 112, the capacitor voltage begins to rise as shown by the upward slope of the voltage curve starting at point b in Figure 8B. Once again, the processor operationsuccessively executes steps SlOO to Sl 12 until the capacitor voltage once again reaches the upper threshold, as sensed by S 102 and as indicated as point c in Figures 8 A and 8B.

Now, on this occasion, when the RF output is turned on (step Sl 12) a plasma is established around the loop electrode, as shown by the increasing load impedance (see the lower trace in Figure 8B at c to d). In this instance, the rise in impedance is such that the energy drain from the capacitor 60 (Figure 3) is insufficient to cause the capacitor voltage to drop below the lower threshold of 90V (see point d on the capacitor voltage trace on Figure 8B), which means that step SI lO in Figure 7 produces a negative result and the RF output remains ON. The above "fϊring-up" process has taken place in a period of about ¹A of a second in this example. The first attempt to start a plasma failed, but owing to the heat generated in the first attempt, the second attempt succeeded. Normally, a plasma will be formed in low impedance conditions at the first or second pulse of RF energy. Firing-up is achieved as a result of the high power delivery available from the generator output stage when supplied from a large reservoir capacitor. Should the plasma be extinguished subsequently, it can be restarted with a somewhat lower voltage. In accordance with the present invention, when the capacitor voltage exceeds an intermediate threshold, at which time the power supply stage can be considered to be in an energy surplus condition, the generator is operated in a particular pulsed mode of operation to promote a degree of haemostasis as tissue is cut.

Accordingly, referring to Figures 9A and 9B, the RF output is repeatedly interrupted whilst the capacitor voltage exceeds the intermediate threshold. Referring to the upper trace of Figure 9A, showing instantaneous power delivery, at the start of each RF ON period a plasma has to be established. Until then, the load impedance is low and the instantaneous power is high. Subsequently, during each pulse, the power drops as the impedance rises (see the lower trace in Figure 9B). In the example shown in Figures 9A and 9B, the impedance oscillates generally between about 400 and 800 ohms, and power delivery is such that the supply voltage across the reservoir capacitor is maintained at or near the upper threshold (see the upper trace in Figure 9B).

The generator load curve is such that, with pulsed operation, a constant user-defined average power level can be delivered to tissue unless the load impedance is high enough to cause a maximum RF output voltage to be reached. At impedances above this impedance threshold of typically 450 to 600 ohms, the power output from the generator falls progressively as the RF load impedance rises. This condition occurs when the electrode is predominantly engaged in tissue where, for instance, in the middle of a resection stroke with a loop electrode, impedances may reach 900 ohms to 1000 ohms, the power output then being about 10OW. The generator is preferably set to have a higher default power setting for such an electrode, typically 160W, to ensure plasma fire-up in saline. When an interruption has occurred, the RF power is reapplied using energy reserves as necessary for sufficiently long to allow the plasma to be reestablished and power delivery to be maintained at a sufficiently high level to continue cutting or vaporisation of tissue.

This interrupted mode of operation takes place as a result of the processor 65 (Figure 3) following the "yes" output from step 106 to perform the pulse timing steps Sl 16, S120, S122 (when the RF output is ON), and steps Sl 16, S120 and S124 (when the RF output is interrupted). The processor timer 65A is configured to produce a predetermined mark-to-space ratio and a predetermined pulse repetition rate. In a typical setting, the mark-to-space ratio is 1:1, with the pulse ON duration being set in the region of 0.5ms. The OFF duration is preferably at least 0.5ms but also, preferably, less than 1.5ms. It will be understood that other mark-to-space ratios and pulse widths are possible. Good results, i.e. rapid cutting with sufficient haemostasis to reduce bleeding to a negligible level have been achieved with mark-to-space ratios in the region of 2:1 to 1:4, but values outside this range are feasible within the scope of the invention. The preferred generator includes a control for adjusting the parameters of this pulsed mode of operation to achieve different degrees of haemostasis and cutting drag (the physical resistance to movement of the electrode). Tissue cutting may be accompanied by increased bleeding, which, if interruptions to the RF output are applied, lowers the load impedance to the extent that, after a cutting has begun, the capacitor voltage drops below the intermediate threshold. Referring to Figures 1OA and 1OB, it will be seen that depressed load impedance (the lower trace in Figure 10B) results in very high instantaneous RF output power levels to reestablish the plasma. Consequently, the capacitor voltage rapidly decreases to reach the intermediate threshold at point e (see Figure 10B).

Referring to Figure 7, the processor operation is diverted from its timed pulsing mode by comparison step S 106 whereupon, on reaching step SI lO, operation reverts to step SlOO and repeatedly executes the loop SlOO, S 102, S 104, S 106, S 108, SI lO so that the DC power supply 50 (Figure 3) remains on continuously and the RF output is continuous until the capacitor voltage regains the intermediate threshold, as sensed by step S 106. Referring back to Figures 1OA and 1OB, this period is seen as a period/of constant RF output and increasing capacitor voltage. With a CW RF output, the load impedance is maintained at a comparatively high value, allowing the capacitor to recharge.

When the capacitor voltage reaches the intermediate threshold, the RF output is interrupted by the timer loop (steps Sl 16, Sl 18, S120, S124) and when the RF output is next turned ON (step S 122) another very high power spike g is produced as the capacitor once again discharges into the low impedance which has occurred as a result of the plasma being extinguished when the RF output was turned off. This process of CW operation with comparatively infrequent interruptions continues until a less conductive tissue region in reached. Note that interruptions to the RF output occur only when the capacitor voltage is high enough, indicating an energy "surplus" and a sufficiently high reservoir of energy to reestablish the plasma. The control process can be adjusted by the user by altering the generator power setting. A default power setting is 160W. Increasing the power setting above this level increases the likelihood of an energy "surplus" and, therefore, increases the power dissipating interruptions to the plasma, thereby increasing haemostasis. The tissue cutting performance increases with power. Reducing the power setting below the default reduces the likelihood of an energy surplus and eventually causes the control process to cease making interruptions. The result in the latter case is that RF power is normally delivered only in a continuous wave (CW) form with interruptions occurring only when the capacitor voltage drops below the lower threshold defined by comparison step SIlO. However, at a lower power setting of the generator, the cutting characteristics are less aggressive and, consequently, are less likely to result in excessive haemostasis.

An alternative generator for use in the system described above with reference to Figure 1 will now be described with reference to Figure 11. This generator has a variable frequency RF source including a voltage controlled oscillator (VCO) 4OA. In this example, the VCO feeds a divide-by-two stage 4OB which, in turn, feeds a power driver stage 44 A driving an RF output stage in the form of a power bridge 44B. The power bridge 44B feeds a resonant output network 80 which delivers a generator output signal across output terminals 74. In practice, the power driver stage 44 A and the power bridge 44B can have the same configuration as the power mosfet and driver circuit 44 of the generator described above with reference to Figure 3. The power bridge 44B takes its DC supply from the supply rail 58 of the DC power supply 50, but the driver stage 44A has a lower voltage supply. Typical supply voltages are 180V maximum for the power bridge 44B and 16.5V for the driver stage 44A.

To bring the frequency of the combination of the VCO 4OA and divide-by-two stage 4OB to the resonant frequency of the output network 80, the above-described components of the RF source are coupled in a phase-locked loop including a phase sensing element 82 coupled between the power bridge 44B and the output network 80 to sense the voltage phase in the input leads to the output network. This voltage phase signal is applied to one input of a phase comparator 84, the other input of which receives a signal representative of the output of the VCO 4OA, derived from the output of the divide-by-two stage 4OB via a delay stage 86 which compensates for the delay to the RF signal as it passes through the power driver and the power bridge.

As in the first-described generator, the RF output stage 44B is supplied from the DC supply rail 58 attached to the reservoir capacitor 60, which allows large currents to be drawn by the output stage 44B for short periods of time, i.e. currents significantly larger than the current rating of a power supply 50 connected to the DC supply rail 58. It follows that the voltage on supply rail 58 will fall during the time that a large current is drawn. Such variations in voltage are sensed by the voltage sensing stage 62 coupled to the rail 58. Voltage sensing circuit 62 has a control output coupled to the first transmission gate 64 in a line 88 coupling the divided-down output of the VCO 4OA to the input of the power driver 44 A.

As before, the arrangement of the voltage sensing stage 62 and the pulsing circuit 63 are such that when the voltage on supply rail 58 (the voltage supplied to the power bridge 44B) drops below a predetermined voltage threshold, the gate 64 is operated to interrupt the signal path between the VCO and the power driver 44A. When the supply rail voltage rises again, the gate 64 reverts to its conducting state. This may happen when the voltage rises above the threshold mentioned above, or the maximum threshold voltage referred to above.

The second transmission gate 72, connected in series in the signal line 88 with the voltage-operated gate 64, has a control input connected to the output of a 0.5 second monostable 70 which is triggered by current sensing circuitry comprising the current sensing element 46 in one of the input leads to the output network 80 and the comparator 66. These elements act to interrupt the signal line 88 to the power driver

44A for 0.5 seconds when the power bridge output current exceeds a predetermined threshold.

Referring to Figure 12, the resonant output network 80 comprises the series combination of an in-line inductance L₁ and a tank capacitor C₁. The output is taken from across the tank capacitor C₁ (which takes out switching noise) via a first coupling capacitor C₂. This first coupling capacitor C₂ couples to the output (represented by terminals 74) via a step-up matching transformer T with a 1 : 2 step-up ratio. The secondary rewinding of the transformer T couples to the output terminals via a second coupling capacitor C3. In this embodiment, Li is about 4.7μH, the tank capacitor is about 1 OnF and the two coupling capacitors C₂ and C₃ co-operate (one of them via the transformer T) to form a coupling capacitance of about 23nF.

It will be appreciated that when the output terminals 74 are open-circuit, the resonant frequency of the output network is determined by the series combination of Li and C₁. When the output terminals 74 are shorted, the resonant frequency is determined by the series combination of L₁ and the network represented by Ci, C₂, C₃ and T. With the values given, the short-circuit resonant frequency is about 0.55 times the open-circuit resonant frequency.

One of the features of a series-tuned output stage is that peak power delivery inherently occurs at extremely low and extremely high impedances. Referring to Figure 13, the load curve of a series-tuned network (i.e. the delivered power versus load impedance) at resonance is shown by the dotted curve A. The network 80 has minimum power delivery, which may be regarded as the "matched condition", at a load impedance across the terminals 74 (Figures 11 and 12) of about 200 ohms. It will be noted that the part of the curve A which has a negative slope follows a path which is approximately hyperbolic over a major part of its length, which means that this part of the curve is of similar shape to a constant voltage line on the graph of Figure 13.

The applicant has recognised that such a characteristic, when applied to the output stage of an electrosurgical generator, allows output power to be maximised for a given constant voltage limit over a range of load impedances. It has been found that erosion of the active electrode of an electrosurgical instrument operated in a conductive liquid increases markedly when the output voltage rises above a threshold in the region of 900 volts, to 1100 volts peak-to-peak. By arranging for the load curve of the output network 48 to follow an approximate constant voltage curve at about 1000 volts peak- to-peak (340 volts rms) the power delivered into a varying load impedance can be close to the maximum theoretically achievable for that voltage.

In effect, over the range of load impedances of importance in so-called "underwater" electrosurgery, the generator can be made to behave as a constant voltage supply. This can be achieved with a matched output impedance much higher than the load impedance presented by the electrode assembly shown in Figure 2A and 2B in the wetted condition, which, for a 4mm loop is in the region of 25 ohms. This translates to a maximum power of about 4.5kW at 340 volts rms.

The actual load curve achieved with the arrangement shown in Figures 11 and 12 is shown by curve B in Figure 13. This deviates from the series-tuned curve A at low impedances owing to imposition of a current limit using the current sensing stage circuitry 46, 66 monostable 70 and transmission gate 72 (Figure 11). In the present embodiment, the current limit is set at a level of about 13 amps. The actual load curve B also deviates from the inherent series-tuned load curve A towards the lower part of the negative-slope portion of the curve A so that the delivered power follows a continuing negative gradient as the load impedance rises, again mimicking a constant voltage supply. This latter deviation is deliberate inasmuch as extreme power into a very high impedance is undesirable. The reason for this deviation is the movement of the resonant frequency of the output network 80, as described above, coupled with the imposition of a high-frequency limit on the RF frequency output as will be described below. The phase comparator 84 compares the current phase at the input to the output network 80, as sensed by the phase sensing circuit 82 with a delayed version of the output of the divide-by-two circuit 4OB which, in turn, is fed by the VCO 4OA. Accordingly, the phase and frequency of the VCO are varied to maintain a constant phase at the input to the output network 80, subject to the upper frequency limit mentioned above. In the absence of other influences, therefore, the output network 80 is maintained in resonance as the load impedance varies.

Given that the free-running frequency of the phase-locked loop is arranged to be its maximum frequency of operation, the locking characteristics of the phase-locked loop are such that it can be brought into a locked condition at the minimum frequency, corresponding to minimum load impedance, sufficiently quickly to achieve resonance in the early part of the output pulse, but not so quickly that the current limit circuit

(sensing circuitry 46, 66 monostable 70 and gate 72) fails to trip when the current exceeds a predetermined current threshold.

If, now, the output carrier frequency is limited to a value below the frequency of the matched load resonant condition, the delivered power will fall off as the load impedance increases and the resonant frequency correspondingly rises. In fact, the free-run output frequency of the phase locked loop containing the VCO 4OA (Figure 7) is designed to be this maximum frequency. This ensures that the output network always represents a higher source impedance than the impedance of the load, which affords over- voltage protection in the event of a short. Summarising, to achieve optimum resonant frequency, the excitation oscillator (VCO) is phase-locked to the resonant output network. Defining the range of the VCO provides load curve definition in that the delivered output power falls below the theoretical maximum when the output network resonant frequency rises above the maximum frequency of the divided down output of the VCO 4OA. In other words, a match at high load impedance is prevented by preventing the VCO from generating the higher frequencies necessary for resonance. It also follows that, at high load impedances, the maximum output voltage is controlled by virtue of frequency.

It will be seen from Figure 13 that the delivered output power is in excess of IkW over a range of load impedances corresponding to a wetted or partly wetted electrode. Once vaporisation and arcing has been initiated, the impedance rises, and the delivered power falls. To maintain the average output power at 200W or less, the output signal is pulsed when the load impedance is low. It will be understood that with a peak power in excess of 4kW, the pulse duty cycle needs to drop to a level in the region of 5% or less. The pulse repetition rate should be between 5Hz and 2kHz, and is preferably at least 10Hz. These figures are chosen in view of the time taken to initiate vaporisation at the electrode surface. This means that the pulses have a maximum length of about 4 or 5ms into a low impedance requiring maximum power. Typically, the pulse length is in the region of 1 to 2ms. While it is not essential, configuring the RF output stage of the generator as an amplifier amplifying the output of a signal derived from a separate oscillator, rather than having a self-oscillating output stage, is preferred in order that full peak power can be achieved within the above-stated pulse lengths. (In this embodiment, the output stage 44B is an amplifier configured as a power switching bridge for high efficiency.) Should the VCO fail to operate at a frequency corresponding to resonance of the output network 80, as may happen at the start of each pulse, excessive output currents associated with such a mismatch are prevented since the series-tuned output network is low impedance only at resonance.

Pulsing of the output signal can be performed in a number of ways, including simply pulse modulating with predetermined pulse lengths and pulse repetition rates. In the low impedance mode of operation of the alternative generator described here, the output is pulsed only during an initial period from the commencement of treatment, the output signal being a continuous wave (CW) signal or an interrupted signal (as described above with reference to Figure 7) thereafter, i.e. generally when vaporisation and arcing have been achieved and the load impedance is in an upper range. The duration of the initial period may be fixed or it may be determined by monitoring the load impedance and terminating the initial period when the impedance exceeds a predetermined value. In this embodiment, the duration of the initial period and the length and frequency of the pulses are dynamically variable in response to delivered energy, as measured by the supply rail voltage on supply rail 58. As has been explained above, high instantaneous power levels are achieved only by allowing the output stage 44B to draw current from a charge reservoir, here a large capacitance such as capacitor the 6mF capacitor 60. As charge is drawn from the capacitor 60, the supply rail voltage drops. Between pulses, the supply rail voltage rises again. Accordingly, by using gate 64 alternately to allow and prevent the passage of an RF signal along signal line 88 to the power driver 44 A according to the relationship between the supply voltage level and thresholds set in the voltage sensing circuit 62, the output of the generator can be pulsed to achieve maximum peak delivered power whilst operating within a predetermined average power limit. This equilibrium of power consumption and DC supply voltage is achieved by setting the voltage thresholds so that the RF output stage is activated when the supply rail voltage is sufficient to achieve a maximum vaporisation voltage (e.g. 340V rms) and switched off when a lower threshold is reached. The lower threshold defines the maximum energy per pulse and the repetition rate for a given average power level. The initial period referred to above is terminated when the electrode has "fired-up", in other words when vaporisation and arcing have commenced, so that the load impedance rises and the supply rail voltage stays above the switching threshold or thresholds. In this way it is possible to achieve vaporisation of the conductive liquid surrounding the electrode at impedances as low as 20 ohms without unacceptable erosion of the electrode surface.

Claims

1. An electrosurgical generator for supplying radio frequency (RF) power to an electrosurgical instrument for cutting or vaporising tissue, wherein the generator comprises: an RF output stage having at least one RF power device, at least one pair of output lines for delivering RF power to the instrument, and a series-resonant output network coupled between the RF power device and the said pair of output lines; a power supply stage coupled to the RF output stage to supply power thereto and having a sensing element for sensing energy levels in the power supply stage; and a control circuit coupled to the sensing element for interrupting the delivery of

RF power by the RF output stage, the control circuit being configured to interrupt the power delivery whilst the energy level in the power supply stage exceeds an interruption energy threshold and to cause continuous wave operation of the RF output stage when the energy level is below the threshold.

2. A generator according to claim 1, wherein the output impedance of the output stage at the output lines is less than 200/VP ohms, where P is the maximum continuous RF output power of the generator in watts.

3. A generator according to claim 1 or claim 2, wherein the control circuit is a pulsing circuit configured to apply the said interruptions repeatedly in the form of pulse modulation of the RF output stage with a mark-to-space ratio of between 2:1 and 1 :10.

4. A generator according to any preceding claim, wherein the control circuit is configured such that the RF power ON periods between successive interruptions have a duration in the range of from 0.2ms to 5ms.

5. A generator according to any preceding claim, wherein the control circuit is configured such that the interruptions in the delivery of RF power have a duration of less than 10ms.

6. A generator according to claim 5, wherein the control circuit is configured such that the interruptions have a duration of less than 1.5ms.

7. A generator according to claim 5 or claim 6, wherein the control circuit is configured such that the interruptions have a duration of at least 0.5ms.

8. A generator according to any preceding claim, wherein the said energy level threshold is set at a value which is between 2 and 20 percent below a predetermined maximum power supply energy level.

9. A generator according to any preceding claim, wherein the control circuit is configured such that delivery of RF power by the RP output stage is substantially terminated when the power supply energy level falls below a minimum energy threshold which is lower than the interruption energy threshold.

10. A generator according to claim 9, wherein the control circuit is configured such that the delivery of RF power is resumed after the said termination only when the power supply energy level reaches a maximum power supply energy threshold which is higher than the interruption energy threshold.

11. A generator according to claim 10, wherein the minimum energy threshold is set at a value which is between 5 and 50 percent lower than the maximum energy threshold.

12. A generator according to any preceding claim, wherein the power supply stage includes a charge-storing element for supplying power to the RF output stage and wherein the sensing element comprises a voltage sensing circuit arranged to sense the voltage supplied to the RF output stage by the charge-storing element, the or each said energy level threshold being defined as a respective threshold value of the sensed voltage.

13. A generator according to any preceding claim, wherein the control circuit is a pulsing circuit configured to apply pulse modulation with a predetermined mark-to- space ratio and pulse repetition rate.

14. An electrosurgical generator for supplying radio frequency (RF) power to an electrosurgical instrument for cutting or vaporizing tissue, wherein the generator comprises an RF output stage having at least one RF power device, and at least one pair of output lines for delivering RF power to the instrument, and a power supply stage which is coupled to the RF output stage to supply power thereto and which has a sensing element for dynamically sensing energy levels in the power supply stage, the generator being capable of delivering RF power at the output lines with a peak current of at least IA and a simultaneous peak voltage of at least 300V, and wherein the generator further comprises a control circuit coupled to the sensing element to receive an energy level sensing signal from the sensing element and to cause interruptions in the delivery of RF power when the energy level sensing signal is indicative of the power supply energy level being in an upper energy region and to cause continuous wave supply of RF power when the energy level sensing signal is indicative of the power supply energy level being in a lower energy region.

15. A generator according to claim 14, wherein the circuit is a pulsing circuit configured to apply the said interruptions repeatedly in the form of pulse modulation of the RF power.

16. A generator according to claim 15, wherein, during pulsed operation, RF power is delivered with a peak voltage of at least 300V during the entire pulse length.

17. A generator according to any of claims 14 to 16, wherein the pulsing circuit includes a timer operable to define a pulsed operation mark-to-space ratio in the region of from 2:1 to 1:10.

18. A generator according to claims 17, wherein the timer is configured such that the mark-to-space ratio lies between 1:1 and 1:5.

19. A generator according to claim 17 or claim 18, including a control for adjusting the said mark-to-space ratio.

20. A generator according to any of claims 14 to 19, wherein the output impedance of the output stage at the output lines is less than 10 ohms.

21. A generator according to any of claims 14 to 20, arranged such that RF power delivery to the output lines is cut off when the power supply energy level falls below a predetermined minimum value.