High frequency voltage supply control method for multipole or monopole analysers

Abstract

A voltage supply system for supplying an RF voltage to an RF resonant load comprising an ion-optical component of a mass spectrometer is disclosed. The system comprises a Direct Digital Synthesizer (DDS) arranged and adapted to output an RF voltage. The voltage supply system is arranged and adapted: (i) to vary the frequency of the RF voltage output by the Direct Digital Synthesizer, (ii) to determine a first resonant frequency of the RF resonant load comprising the ion-optical component, and (iii) to determine whether or not the generation of an RF voltage at the first resonant frequency by the Direct Digital Synthesizer would also result in the generation of a spur frequency close to the first resonant frequency. If it is determined that a spur frequency would be generated close to the first resonant frequency then the voltage supply system is further arranged and adapted: (iv) to consult a look-up table comprising one or more preferred frequencies, and (v) to direct the Direct Digital Synthesizer to generate an RF voltage at a second frequency which corresponds with one of the preferred frequencies from the look-up table, wherein the second frequency is different to said first resonant frequency.

Claims

1. A voltage supply system for supplying an RF voltage to an RF resonant load comprising an ion-optical component of a mass spectrometer, said system comprising: a Direct Digital Synthesiser (DDS) arranged and adapted to output an RF voltage; wherein said voltage supply system includes a programmable computer configured to: (i) consult a look-up table comprising one or more preferred frequencies; (ii) vary the frequency of said RF voltage output by said Direct Digital Synthesiser by stepping though the one or more preferred frequencies from said look-up table; (iii) determine which one of said one or more preferred frequencies is closest to a resonant frequency of said RF resonant load comprising said ion-optical component; and (iv) direct said Direct Digital Synthesiser to generate an RF voltage at a frequency which corresponds with the one of said one or more preferred frequencies from said look-up table which is determined to be closest to said resonant frequency.

2. A voltage supply system as claimed in claim 1, wherein said one of said one or more preferred frequencies is substantially close to said resonant frequency but does not result in the generation of a spur frequency close to said resonant frequency.

3. A voltage supply system as claimed in claim 1, wherein said Direct Digital Synthesiser is arranged and adapted to output a generally sinusoidal RF voltage having a fixed amplitude.

4. A voltage supply system as claimed in claim 1, wherein said Direct Digital Synthesiser further comprises a Numerically Controlled Oscillator (NCO).

5. A voltage supply system as claimed in claim 4, wherein said Direct Digital Synthesiser further comprises a Digital to Analogue Converter (DAC) coupled to an output of said Numerically Controlled Oscillator.

6. A voltage supply system as claimed in claim 1, wherein said voltage supply system comprises a digital controller arranged and adapted to control the frequency of said RF voltage output by said Direct Digital Synthesiser.

7. A voltage supply system as claimed in claim 1, further comprising one or more amplifiers for amplifying said RF voltage output by said Direct Digital Synthesiser so that an amplified RF voltage is supplied to said RF resonant load comprising said ion-optical component.

8. A voltage supply system as claimed in claim 1, further comprising an RF amplitude measurement device arranged and adapted to determine the amplitude of said RF voltage as supplied to said RF resonant load comprising said ion-optical component.

9. A voltage supply system as claimed in claim 1, wherein the programmable computer is further configured to determine for which of said one or more preferred frequencies the measured amplitude of said RF voltage as supplied to said RF resonant load comprising said ion-optical component is at a maximum or for which of said one or more preferred frequencies the RF is maximum when compared with a drive level.

10. A voltage supply system as claimed in claim 1, wherein said ion-optical component comprises a multipole or monopole mass filter or mass analyser.

11. A voltage supply system as claimed in claim 10, wherein said ion-optical component comprises a quadrupole mass filter or mass analyser.

12. A voltage supply system as claimed in claim 1, wherein said ion-optical component comprises an RF ion trap.

13. A voltage supply system as claimed in claim 1, further comprising an RF amplitude detector arranged and adapted to output a DC voltage or current which is substantially proportional to the amplitude and the frequency of said RF voltage as supplied to said RF resonant load comprising said ion-optical component.

14. A voltage supply system as claimed in claim 1, further comprising one or more fixed inductors which couple said voltage supply system to said ion-optical component.

15. A mass spectrometer comprising a voltage supply system as claimed in claim 1.

16. A mass spectrometer as claimed in claim 15, wherein said mass spectrometer comprises a miniature mass spectrometer.

17. A method of supplying an RF voltage to an RF resonant load comprising an ion-optical component of a mass spectrometer comprising: providing a Direct Digital Synthesiser (DDS) which outputs an RF voltage; consulting a look-up table comprising one or more preferred frequencies; varying the frequency of said RF voltage output by said Direct Digital Synthesiser by stepping though the one or more preferred frequencies from said look-up table; determining which one of said one or more preferred frequencies is closest to a resonant frequency of said RF resonant load comprising said ion-optical component; and directing said Direct Digital Synthesiser to generate an RF voltage at a frequency which corresponds with the one of said one or more preferred frequencies from said look-up table which is determined to be closest to said resonant frequency.

18. A method as claimed in claim 17, wherein said one of said one or more preferred frequencies is substantially close to said resonant frequency but does not result in the generation of a spur frequency close to said resonant frequency.

19. A method as claimed in claim 17, further comprising determining for which of said one or more preferred frequencies the measured amplitude of said RF voltage as supplied to said RF resonant load comprising said ion-optical component is at a maximum or for which of said one or more preferred frequencies the RF is maximum when compared with a drive level.

20. A method of mass spectrometry comprising a method as claimed in claim 17.

Description

BRIEF DESCRIPTIONS OF THE DRAWINGS

(1) Various embodiments of the present invention together with other arrangements given for illustrative purposed only will not be described, by way of example only, and with reference to the accompanying drawings in which:

(2) FIG. 1 shows a known voltage supply circuit for a quadrupole mass filter;

(3) FIG. 2 shows a voltage supply circuit for a quadrupole mass filter according to a preferred embodiment of the present invention;

(4) FIG. 3 shows how according to an embodiment of the present invention a mass ramp signal may be generated;

(5) FIG. 4 shows a DDS output spectrum showing no large spurs close to the fundamental or resonant frequency;

(6) FIG. 5 shows a DDS output spectrum showing a large spur close to the fundamental or resonant frequency;

(7) FIG. 6 shows a mass spectrum which has low sensitivity and is poorly resolved from its isotope and a corresponding ion current plot at 1080.0 Da as a function of time;

(8) FIG. 7 shows a mass spectrum which is poorly resolved, noisy and shows poor sensitivity and a corresponding ion current plot at 1080.0 Da as a function of time;

(9) FIG. 8 shows a mass spectrum which is well resolved from its isotopes and wherein there is little peak top noise and a corresponding ion current plot at 1080.0 Da as a function of time;

(10) FIG. 9 shows a mass spectrum which is well resolved and wherein there is significant low frequency peak top noise and a corresponding ion current plot at 1080.0 Da as a function of time;

(11) FIG. 10 shows a mass spectrum which is well resolved from its isotopes and wherein there is little peak top noise and a corresponding ion current plot at 1080.0 Da as a function of time;

(12) FIG. 11 shows a mass spectrum which is well resolved but which shows significant high frequency noise and a corresponding ion current plot at 1080.0 Da as a function of time.

DETAILED DESCRIPTION OF PREFERRED EMBODIMENTS

(13) A conventional voltage supply circuit for a quadrupole mass filter will first be described with reference to FIG. 1.

(14) A quadrupole mass filter 6 is shown which consists of four rods 6 which are typically circular or hyperbolic in cross section. The application of a sinusoidal voltage to one pair of the rods 6, and its antiphase to the opposite pair of rods 6 causes ions passing axially along an ion guiding region cavity between the rod electrodes 6 to oscillate in a complex manner. Depending upon the mass to charge ration of the ions these oscillations will either typically because of such amplitude that the ions will collide with one of the rods and hence will not pass through the mass filter or else the ions will pass from one end of the quadrupole to the other (i.e. the ions will pass through the mass filter and be onwardly transmitted.)

(15) The quadrupole mass filter 6 is commonly operated as a bandpass filter. Only ions having mass to charge ratios above a low mass to charge ratio cut-off and below a high mass to charge ratio cut-off will pass through and be onwardly transmitted by the mass filter 6. The centre of the pass band is proportional to the amplitude of the sinusoidal RF voltage applied to the rod electrodes 6 and is inversely proportional to the square of the frequency of the sinusoidal RF voltage as applied to the rod electrodes 6.

(16) If a DC signal is also superimposed on the rod electrodes 6 (with approximately equal value but opposite polarity on the rod pairs) in addition to the RF voltage then the range of mass to charge ratios of ions passed by the quadrupole rod set mass filter 6 will be diminished.

(17) At some level of applied DC voltage singly charged ions with a mass difference of approximately one Dalton can be separated by such an mass filter. For ions of a few hundred Daltons or more the ratio of RF to DC required to separate ions 1 Da apart is approximately constant at 5.96:! (i.e. the RF peak amplitude should be 5.96 times that of the DC value.)

(18) It is common practice to ensure that the RF to DC ratio is maintained at approximately 5.96:1 so as to maintain unit resolution which implies the same peak width (of approximately 0.5 Da at half height) for singly charged ions throughout the mass scale.

(19) It will be understood by those skilled in the art that the RF amplitude, the RF frequency and the DC amplitude must be accurately controlled in order for the performance of the mass filter to remain stable and accurate.

(20) Often the amplitude of the voltages required for such quadrupole analysers are in the region of several thousand volts of RF and the RF voltage is supplied at frequencies of around 1 MHz.

(21) The preferred embodiment of the present invention as will be described in more detail below seeks to facilitate the accurate measurement and control of these parameters whilst minimising component cost, setup cost and physical complexity.

(22) FIG. 1 shows a known control or drive circuit which is used to supply RF and DC voltages to a quadrupole mass filter 6. The signal paths shown with bold arrows are digital signals. The other signal paths are analogue.

(23) A fixed frequency generator 1 is provided which produces a fixed RF frequency with substantially a fixed amplitude. The fixed frequency generator 1 is not controlled by a digital controller 2 and the frequency of the RF voltage output by the fixed frequency generator 1 is not variable.

(24) An amplitude modulator 3 amplifies the RF signal output from the fixed frequency generator 1 by an amount proportional to its control input. An inverter 4 follows the amplitude modulator 3 which allows both the RF signal and an identical RF signal with 180 phase shift to be fed to a pair of power amplifiers 5a,5b. The power amplifiers 5a,5b buffer the voltage and feed their AC output currents directly to the rods 6 of the quadrupole via variable inductors 7. The variable inductors 7 are manually tuned so that, along with the capacitive load of the quadrupoles 6, the inductors 7 form a resonant load whose resonant frequency matches the drive frequency fundamental.

(25) At resonance the voltage at the quadrupole rods may be several hundred times higher than that at the power amplifier 5a,5b outputs (dependent upon the quality factor of the circuit, the inductance and the frequency of the input).

(26) An amplitude measurement circuit 8 is provided which utilizes capacitors 9 to produce a current that is proportional to both the frequency and voltage amplitude at the quadrupole 6. Diodes 10 rectify the current and an ammeter is formed through the use of a low value resistor 11. A buffer amplifier 12 outputs a voltage proportional to the average sensed DC current.

(27) The gain of the amplitude measurement circuit 8 may be calibrated by altering an RF adjustment Multiplying Digital to Analogue Converter (MDAC) 13. The output of the RF adjustment Multiplying Digital to Analogue Converter 13 is compared to a mass program level output from a mass program MDAC 14 and the output of that comparison circuit 15 (typically consisting of a difference integrator) is then fed to the amplitude modulator 3 to form a closed loop control system.

(28) The analogue signals ensure that the RF amplitude at the quadrupole 6 is equal to a mass program level multiple by a known fixed constant.

(29) To achieve constant unit resolution across the mass scale, the DC voltages applied to the quadrupole rod electrodes 6 should be approximately +RFpeak/5.96 and RFpeak/5.96. This means that if the High Voltage amplifiers have a suitable fixed gain then the resolution across the mass range will be substantially constant, and this resolution can be altered by adjusting a DC adjustment MDAC 16.

(30) The known system as shown in FIG. 1 suffers from a number of problems.

(31) Firstly, the adjustable high voltage inductors 7 introduce mechanically complexity as well as power losses (which in turn means more power is required to be supplied by the power amplifiers).

(32) Secondly, adjusting the high voltage inductors 7 to allow resonance at the fixed drive frequency requires sensitive manual setup when the system is manufactured or during servicing.

(33) Thirdly, multiplying DACs 13, 14, 16 (MDACs) are more expensive than non-multiplying DACs and typically take up more circuit board area than DACs which have a fixed reference.

(34) It should be appreciated that the arrangement shown in FIG. 1 and the preferred embodiment as shown in FIG. 2 and described below are a simplification. For example, the RF to DC ration required to attain unit resolution at low masses increases substantially and the rectifiers 10 in the amplitude measurement circuit 8 introduce nonlinearities which become significant at lower masses. However, details of the application of corrections to these potential sources of error are not relevant to the principle of the present invention which will be described in more detail below.

(35) According to an embodiment of the present invention an improved drive and control circuit is accomplished through the use of a digitally controlled oscillator. Furthermore, the high voltage variably inductors 7 as used conventionally are replaced with lower cost fixed inductors.

(36) A preferred embodiment of the present invention will now be described with reference to FIG. 2.

(37) FIG. 2 shows a preferred embodiment of the present invention. The signal paths shown with bold arrows are digital signals. The other signal paths are analogue.

(38) A frequency synthesizer 18 is constructed with a Direct Digital Synthesis (DDS) technique. A digital controller 19 selects the required frequency by instructing the frequency synthesiser 18 which outputs a constant amplitude approximately sinusoidal waveform.

(39) An amplitude modulator 20 amplifies the sinusoidal RF voltage output by the frequency synthesizer 18 by an amount proportional to its control output.

(40) Inverter 21 follows which allows both the sinusoid and an identical sinusoid with 180 phase shift to be fed to a pair of power amplifiers 22a,22b. The power amplifiers 22a,22b preferably buffer the voltage and feed their AC output currents directly to the quadrupoles 6 via fixed inductors 23. The fixed inductors 23 along the capacitive load of the quadrupole 6 form a resonant load.

(41) In normal operation the frequency set by the digital controller 19 is predetermined so as to match closely the resonant frequency of this load. At resonance the voltage at the quadrupole 6 may be several hundred times higher than that at the power amplifier 22a,22b outputs (dependent upon the quality factor of the circuit, the inductance and the frequency of the input.

(42) An amplitude measurement circuit 24 is preferably provided and preferably utilizes capacitors 25 to produce a current that is proportional to both the frequency and voltage amplitude at the quadrupole 6. Diodes 26 preferably rectify this current and an ammeter is preferably formed through the use of a low value resistor 27 and buffer amplifier 28 (which outputs a voltage proportional to the average sensed DC current).

(43) The output of the amplitude measurement circuit 24 is then preferably added to an RF adjustment level as output from an RF adjustment Digital to Analogue Converter (DAC) 29 and the resultant signal is then preferably compared to a mass program level as output by a mass program DAC 30. The output of that comparison circuit 31 (which preferably comprises a difference integrator 31) is fed to the amplitude modulator 20 to form a closed loop control system.

(44) Thus the analogue signals ensure that the amplitude measured is equal to the Mass program level less the RF adjustment level. For a given quadrupole design the mass to charge ration selected (i.e. that at the peak of the stability curve) is proportional to the sinusoidal amplitude on the rods 6 and is inversely proportional to the square of the frequency of that waveform. For a given set of capacitor and resistor values in the amplitude measurement circuit 24, its output is proportional to the sinusoidal amplitude on the rods 6 and is also proportional to the frequency of that waveform. Thus is it possible to compute the RF adjustment level that will almost exactly counter the effect of the frequency upon the measured signal as well as the mass to charge ration transmitted despite alternations in the RF drive frequency. Furthermore, by using a configuration similar to that shown in FIG. 2 and incorporating digital multipliers within the programmable logic, the use of expensive Multiplying Digital to Analogue Converters (MDACs) can preferably be avoided.

(45) It should be appreciated that the arrangement as shown in FIG. 2 is not the only configuration that can achieve this functionality. For example, the RF adjustment DAC 29 may be removed and the mass program DAC 30 value may be re-computed to include the adjustment that the RF adjustment DAC 29 provided. This latter arrangement would necessitate a further computation to determine the DAC adjustment DAC 32 required to maintain the resolving DC level.

(46) The digital controller 19 is preferably programmed to sweep the RF frequency whilst applying a fixed amplitude drive. The frequency at which the RF amplitude measurement detector 24 reports the highest RF amplitude at the quadrupole 6 (or the highest level produced by the high voltage amplifiers or the drive level into those amplifiers) is preferably noted.

(47) Once this frequency is known, the digital controller 19 is then preferably set to use this value (or one suitably close to that frequency where significant spurs are known to be absent) during analysis. This procedure may be performed during the manufacture of the instrument, during service or periodically as required.

(48) A further improvement to the known circuit as shown in FIG. 1 is accomplished by removing the need to multiple the measured RF amplitude by a variable amount in order to calibrate the RF amplitude measurement. This change allows the MDACs to be replaced by relatively low cost non-multiplying DACs 29,30,32. To allow this the amplitude measurement corrected is removed from the feedback loop and is added as a feed-forward control. Digital multipliers whose input is primarily determined by the mass program value within an Field Programmable Gate Array (FGPA) can be used to allow the MDAC removal whilst avoiding the requirement for an expensive high fidelity, high speed analogue to digital conversion of the amplitude measurement.

(49) There are some disadvantages with adjusting the RF frequency away from its nominal design value.

(50) Firstly, if the instrument was previously calibrated and working and one or more parts of the resonant load were replaced, then the system would be required to adjust the frequency synthesizer 18 for resonance whereafter: (i) the amplitude measurement system would no longer be calibrated; (ii) the centre of the mass window transmitted would be shifted for the same amplitude of RF at the quadrupole; and (iii) the ratio between the RF amplitude and the resolving DC would be altered.

(51) These effects combine together and cause the spectral resolution and peak position to be altered. This in turn would require the system to be set-up for mass-scale and resolution across the mass scale. Such a set-up is often non-trivial as known calibrant chemicals need to be introduced to the instrument and a skilled operator (or complex and potentially unreliable algorithm) is required to make sure spectral peaks are correctly resolved and positioned without misassignment despite a potentially complex spectra containing singly and multiply charged species.

(52) Secondly, if the system is designed with accurate components and is manufactured consistently, the settings for unit resolution and accurate mass scale calibration (using for example the DC and RF adjustment DACs 29,32) will only vary over a small range. Any variation away from the typical adjustment range would indicate a faulty component and is a useful diagnostic, saving costly diagnosis time during manufacture or in the field.

(53) However, if the frequency is shifted significantly away from the design nominal, the mass and resolution adjustments will have to be varied by a large amount in order to set-up the instrument and this will obfuscate the existence of such faulty components. Both of these disadvantages can be overcome by automatically computing the adjustment required to adjust for these frequency effects. For example, it is possible to define a variable which is the percentage different between the nominal design RF frequency and the frequency found to resonant the load. This variable can then be incorporated into equations that can automatically correct the set-up parameters (for any frequency related effects) provided by the instrument operator.

(54) FIG. 3 depicts one such method of employing this invention.

(55) The Position, Setup and Resolution values as shown in FIG. 3 are those parameters which are used by the user or performed automatically to set-up the instrument for the preferred resolution and mass position over the mass scale of interest.

(56) The F parameters are used to adjust those parameters for any deviation in the actual resonant frequency from the nominal design value.

(57) LMF, HMP, LMS, HMS, dF, HMR, LMR are the adjusted values that are sent to an FPGA within the instrument.

(58) Since for many operations the instrument must scan rapidly over a mass range, the FPGA is preferably used to generate a rapid finely stepping mass ramp signal. This mass ramp signal is sent to the mass program DAC 30 and also used within the FPGA to generate ramping (or static) control values to the adjustment DACs (allowing them to be used calibrate out errors in the system that relate to circuit gain, offsets and frequency squared, and the electronics of the RF amplitude measurement system (in the present embodiment) has a gain proportional to frequency.) This means that the ion beam will be unaffected when the frequency is altered (disregarding abnormalities caused by spurs).

(59) There are also some disadvantages to the use of a variable frequency oscillator such as a Voltage Controlled Oscillator (VCO) or Phase Locked Loop (PLL) and those constructed by Direct Digital Synthesis (DDS) including a Numerically Controlled Oscillator (NCO).

(60) VCOs have poor frequency stability in comparison to crystal oscillators or if they employ a crystal within their design (VCXOs) they have a very limited frequency range.

(61) PLL based frequency generators generate phase noise which is disadvantageous for quadrupole analyser based instruments.

(62) DDS circuits are capable of producing a wide range of frequencies with low phase jitter and excellent frequency stability. However, DDS circuits suffer a potentially significant problem in that they also produce spur frequencies in addition to the intended frequency.

(63) The amplitude of these spur frequencies is not a problem if they occur far from the resonant frequency as they will be heavily filtered. However, if spur frequencies appear at frequencies which are close to the resonance frequency then they can have a significant effect upon an ion beam travelling through the quadrupole 6 causing poor resolution, poor sensitivity and instability.

(64) It is known that spur frequencies occur at frequencies which are a complex function of the DDS update rate, the DAC resolution within the DDS, the number of bits used to encode the phase increment value and the way in which those bits are truncated.

(65) Thus the frequencies of the spurs will vary with the requested output frequency, but will be the same for any requested frequency for all instruments employing the same DDS design.

(66) According to a particularly preferred aspect of the present invention a DDS based frequency generator is utilized for the RF drive circuit and this is preferably combined with a look-up table so that only frequencies that do not cause significant spur related spectral imperfections are preferably selectable and if a frequency other than those is requested of the system it will respond by selecting the nearest known good frequency.

(67) Advantages of DDS Over VCO/PLL Circuitry for Quadrupole Based Instruments

(68) DDS systems and VCO/PLL systems both require a master clock. This clock will have some phase noise. For a VCO/PLL system this phase noise is effectively increased (multiplied) by the frequency divider contained within it. Conversely, a DDS system reduces the phase noise at its output due to its output being a fractional division of its clock. Phase noise broadens the frequency spectrum around the desired centre frequency. Since the centre of the pass-band of a quadrupole filter is proportional to 1/f.sub.out.sup.2 this results in a broadening of mass peaks and a subsequent loss in mass resolution.

(69) The Effect of Spur Frequencies on Spectral Peak Quality

(70) DDS systems are capable of producing stable low distortion sinusoidal outputs with little phase noise. However, due to their digital nature they produce quantisation related noise (e.g. due to phase truncation and amplitude quantisation) which causes perturbations that repeat regularly. This causes small amplitude unwanted frequencies known as spurs in addition to the large amplitude intended frequency (f.sub.out).

(71) The frequency spectrum of the spurs is deterministic and is dependent upon the requested fundamental frequency and the design of the DDS. For a given design the output spectrum from one DDS will be almost identical to the output from an identical DDS given the same programmed parameters (e.g. requested output frequency).

(72) However, the spectrum may change significantly for very small changes in requested output frequency.

(73) FIG. 4 shows a DDS output spectrum showing no large spurs close to the fundamental frequency and FIG. 5 shows a DDS output spectrum showing a large spur close to the fundamental frequency.

(74) The plots shown in FIGS. 4 and 5 show amplitude (on a log scaling) on the y axis and frequency (on a linear scaling) on the x axis. It can be seen that in these example plots the largest peak (f.sub.out) is at almost the same frequency in both cases by that the spur spectrum is very different.

(75) Resonant circuits act as filters, heavily attenuating input signals that have frequencies that are not close to the resonant frequency (f.sub.res) of the circuit. As a result, only spur frequencies close to f.sub.res are likely to produce significant noise at the output of such circuits.

(76) In the output spectrum shown in FIG. 5 it can be seen that a large spur occurs close to f.sub.out whereas in the output spectrum shown in FIG. 4 in the larger spurs are relatively distant from f.sub.out. To avoid spurs causing significant noise at the output of such tuned circuits, it is possible to shift f.sub.out away from f.sub.res slightly so that the attenuation of f.sub.out by the circuit is insignificant whilst spurs close to f.sub.out and hence f.sub.res are small.

(77) One method of doing this is to generate a set of suitable spaced values of f.sub.out close to a nearby set of .sub.fres values that do not show potentially significant ion beam effects. This can then be used for all instruments having the equivalent DDS design. Thereafter, whenever desired (e.g. during manufacturing set-up) frequencies can be stepped through until resonance occurs, and one of the listed known good frequencies can then be selected for f.sub.out that is suitably close to f.sub.res.

(78) Alternatively, known bad frequencies may be listed and the known bad frequencies may be avoided when setting f.sub.out instead.

(79) FIGS. 6-11 illustrate how very small changes in f.sub.out can affect the signal of a mass spectrometer where the signal containing f.sub.out is used as part of the drive waveform for a quadrupole mass analyser.

(80) FIGS. 6-11 shows the effect of shifting the frequency between 1136750 Hz and 1140150 Hz. These frequencies lie within a band close enough to the resonant load to allow a suitable level of voltage at the quadrupole without demanding too much power in the drive circuitry i.e. it is broadly at the resonant frequency.

(81) FIG. 6 shows that at a frequency of 1136750 Hz a peak at 1080 Da has low sensitivity and is poorly resolved from its isotope. FIG. 7 shows that when the frequency is increased to 1137050 Hz the peak at 1080 Da is poorly resolved, noisy and shows poor sensitivity.

(82) FIG. 8 shows that when the frequency is increased to 1136750 Hz the peak at 1080 Da is well resolved from its isotopes and there is little peak top noise. FIG. 9 shows that when the frequency is increased further to 1138050 Hz the peak at 1080 Da is well resolved but there is significant low frequency peak top noise.

(83) FIG. 10 shows when the frequency is increased to 1138100 Hz the peak at 1080 Da is well resolved from its isotopes and there is little peak top noise. FIG. 11 shows that when the frequency is increased further to 1140150 Hz the peak at 1080 Da is well resolved, but shows significant high frequency noise.

(84) It can be seen from FIGS. 8 and 10 that at some drive frequencies the beam is undistorted whilst in other cases performance is affected. For example, the results shown in FIG. 6 show poor sensitivity and may result in the limits of detection of the analyser being below the users requirements. The results shown in FIG. 9 suffer from low frequency amplitude modulation which could result in poor quantitation of analytes.

(85) Determining Frequencies for Look-Up Table

(86) Many significant spur frequencies can be predetermined or calculated as they relate to the set frequency, clock frequency, DDS resolution, update rate phase truncation and/or DAC analogue performance. However these calculated frequencies typically also have aliases. The result is that accurately predicting all significant spur frequencies is not straightforward.

(87) Not only are the spur frequencies and their amplitudes difficult to predict, but they are very hard to measure. For example, it is known that a mass error of 0.2 Da when analyzing a mass to charge ratio of 2000 Da is enough to cause a significant change in sensitivity. This implies that frequency or amplitude modulations of 1 part in 10,000 are likely to cause degradation in analytical performance. However, measuring a signal with an amplitude that is, e.g., 80 dB below a reference signal that is very close in frequency (typically within a few ppm) as would be required to measure relevant spurs is highly challenging, even for specialized test equipment.

(88) Not only are the spur frequencies and their amplitude difficult to determine, but their effect on the ion beam is very hard to quantify. The spurs will affect the RF control loop, causing it to make errors in accurately controlling the drive amplitude. Furthermore, the spurs will inter-modulate and the overall effect on the ion trajectories of the resulting complex time varying waveforms is not well understood.

(89) Consequently the preferred embodiment if the present invention utilizes a look-up-table that is preferably generated through careful experimentation.

(90) To create the look-up-table a number of steps were carried out. A special version of the RF generator was created that used an adjustable capacitor, allowing the resonant frequency to be altered. A known compound was infused into the mass spectrometer. The mass spectrometer was set to scan over a small window around a high mass peak (and its isotopes) of interest.

(91) An acceptable frequency offset or detuning x from the peak resonance f.sub.c was determined such that the drive efficiency was not significantly affected, i.e. (xf.sub.cQ). The following steps were then carried out:

(92) 1. The drive frequency, f.sub.d, was set to f.sub.min, where f.sub.min is the minimum expected.

(93) 2. The drive amplitude was fixed at a constant value.

(94) 3. The capacitor was adjusted to give maximum output RF (i.e. f.sub.d=f.sub.res).

(95) 4. The RF control loop was set to closed loop (i.e. normal operation, allowing mass analysis).

(96) 5. The drive frequency was altered to a value f.sub.d=f.sub.resx.

(97) 6. The system (using f methods described above) altered the output RF and DC levels automatically so that the expected resolution and peak position should remain unaltered (except for effects caused by frequency spurs).

(98) 7. The resulting peak shape was then checked for: (a) resolution (e.g. the valley between isotopes), (b) sensitivity (i.e. response height), and (c) amplitude modulation (i.e. how much the amplitude changes with time).

(99) 8. After recording the results the frequency was incremented by a small amount (e.g. 50 ppm).

(100) 9. The process was repeated for steps 6 through 8, until the frequency exceeded f.sub.res=x.

(101) 10. One or more frequencies were selected for entering into the known good frequency table that showed good performance at both of the last two capacitor settings (unless this is the initial capacitor setting).

(102) 11. The capacitor was adjusted to give resonance at f.sub.res=f.sub.res+x.

(103) 12. The process was repeated for steps 5 through 11 until f.sub.max was reached, where f.sub.max is the maximum expected resonant frequency of a production unit.

(104) The look-up table of the preferred embodiment generated in this manner preferably comprises a list of preferred frequencies that give a known good performance. The frequencies in the look-up table of the preferred embodiment are valid for any RF resonance load between f.sub.min and f.sub.max, and preferably comprise at least one frequency within x of any given peak resonance.

(105) Although the preferred embodiment of the present invention related to driving a quadrupole mass filter, alternative embodiments are contemplated wherein the voltage supply system is used to drive a monopole filter or an RF based ion trap.

(106) Although the present invention has been described with reference to preferred embodiments, it will be understood by those skilled in the art that various changes in form and detail may be made without departing from the scope of the invention as set forth in the accompanying claims.