# Patent application title: Magnetic Particle Resuspension Probe Module

##
Inventors:
Martin Trump (Pforzheim, DE)
Martin Fletcher (Winchester, MA, US)
Ann Jacob (Stormville, NY, US)

IPC8 Class: AB01L302FI

USPC Class:
422 32

Class name: Process disinfecting, preserving, deodorizing, or sterilizing using disinfecting or sterilizing substance treating bulk material

Publication date: 2012-02-09

Patent application number: 20120034132

## Abstract:

An acid injection module (100) comprising a dual probe nozzles (102).## Claims:

**1.**A resuspension nozzle module for enabling resuspension of accumulated particles in a reaction vessel comprising: a probe housing mountable in an analytical instrument; a mounting recess on a rear surface of the probe housing for interfacing the module to a source of resuspension liquid; plural channels, defined within the probe housing, in fluid communication with the mounting recess; plural probes, defined within the probe housing, each in fluid communication with a respective one of the channels; and plural probe nozzles, formed on a front surface of the probe housing, each in fluid communication with a respective one of the probes, wherein the plural probes and probe nozzles are mutually parallel.

**2.**The resuspension nozzle module of claim 1 wherein the plural probe nozzles are configured to dispense parallel streams of resuspension liquid.

**3.**In an automated analysis system having at least one resuspension liquid nozzle and serially conveyed reaction vessels in which solids particles are accumulated against an interior wall of each reaction vessel and require resuspension, a method of verifying each nozzle can project a stream of resuspension liquid within a target field on the interior wall of each reaction vessel relative to the accumulated solids particles, the method comprising the steps of: defining linear and angular dimensional offset values between the nozzle and the target field necessary for the resuspension liquid stream to hit the target field; for all system components having non-zero positional tolerances in a respective dimension or dimensions, identifying the respective positional tolerance; determining the nominal value and tolerance for the difference in the respective dimension or dimensions between the nozzle and the respective target as a closure value; calculating the total deviation of the closure value in the respective dimension or dimensions for all components having a non-zero positional tolerance in the respective dimension; for all system components having an asymmetric tolerance distribution in the respective dimension or dimensions, determining the mean values and deviation for each; for all system components having a folded normal tolerance distribution in the respective dimension or dimensions, determining the mean values and deviations for each; determining the statistical closure value with tolerance based upon the determined mean values and deviations and the calculated total deviation in the respective dimension or dimensions; from the total deviation of the closure value in each dimension or dimensions, calculating the arithmetic deviation of the target; from the statistical closure values in each dimension, estimating the total statistical deviation from the target; and from the total statistical deviation, deriving the statistical error from the target with tolerance.

**4.**In an automated analysis system having at least one resuspension liquid nozzle and serially conveyed reaction vessels in which solid particles are accumulated against an interior wall of each reaction vessel, the solid particles requiring selective resuspension, a method of verifying each nozzle can project a stream of resuspension liquid within a target field on the interior wall of each reaction vessel relative to the accumulated solid particles, the method comprising the steps of: defining ideal linear and angular offset values for the nozzle, relative to the cuvette, along vertical and horizontal linear dimensions and in the angular plane defined thereby, the ideal linear and angular offset values enabling resuspension liquid to impact the cuvette within the target field; identifying all structural components contributing to tolerance stack-up in each linear dimension and angular plane; identifying the nominal linear dimension or angular orientation and the tolerance range, if any, for each structural component in each linear dimension and angular plane, respectively; arithmetically calculating a nominal closure value for each linear dimension and angular plane as the difference between the ideal linear and angular offset values and the respective summed structural component nominal linear and angular measurements; determining the tolerance range for the nominal closure value for each linear dimension and angular plane as the arithmetic sum of the respective tolerances of the structural components in the respective linear dimension and angular plane; determining tolerance mean and deviation values for each structural component in each linear dimension and angular plane; statistically calculating the nominal closure value for each linear dimension and angular plane as a distributed average of the tolerance mean values of the respective structural components; statistically calculating the respective tolerance zone for each linear dimension and angular plane from the tolerance deviation values of the respective structural components; and orienting the nozzle in the linear dimensions and the angular plane according to the respective statistically calculated nominal closure values and tolerance zones.

**5.**The method of claim 4, wherein the step of orienting comprises compensating for a cuvette sidewall draft.

**6.**The method of claim 4, wherein the solid particles are accumulated against an interior wall of each reaction vessel by action of a magnet array, and wherein the step of orienting includes orienting according to statistically calculated nominal horizontal closure values and tolerance values associated with the magnet array.

**7.**The method of claim 4, wherein the step of orienting comprises compensating for the effect of gravity on a stream of resuspension liquid.

**8.**The method of claim 7, wherein the step of compensating for the effect of gravity comprises further identifying the rate at which the resuspension liquid flows and the diameter of each nozzle.

**9.**The resuspension nozzle module of claim 1, further comprising a buffer zone defined within the housing intermediate the mounting recess and the plural channels, the buffer zone enabling pressure equalization between the plural channels.

## Description:

**CROSS REFERENCE TO RELATED APPLICATIONS**

**[0001]**This application claims priority to U.S. Prov. Appl. No. 60/574,000, filed May 24, 2004, the entirety of which is hereby incorporated by reference.

**STATEMENT REGARDING FEDERALLY SPONSORED RESEARCH OR DEVELOPMENT**

**[0002]**N/A

**BACKGROUND OF THE INVENTION**

**[0003]**Heterogeneous immunoassays typically require the separation of sought-for components bound to component-selective particles from unbound or free components of the assay. To increase the efficiency of this separation, many assays wash the solid phase (the bound component) of the assay after the initial separation (the removal or aspiration of the liquid phase). Some chemiluminescent immunoassays use magnetic separation to remove the unbound assay components from the reaction vessel prior to addition of a reagent used in producing chemiluminescence or the detectable signal indicative of the amount of bound component present. This is accomplished by using magnetizable particles including, but not restricted to, paramagnetic particles, superparamagnetic particles, ferromagnetic particles and ferrimagnetic particles. Tested-for assay components are bound to component-specific sites on magnetizable particles during the course of the assay. The associated magnetizable particles are attracted to magnets for retention in the reaction vessel while the liquid phase, containing unbound components, is aspirated from the reaction vessel.

**[0004]**Washing of the solid phase after the initial separation is accomplished by dispensing and then aspirating a wash solution, such as de-ionized water or a wash buffer, while the magnetizable particles are attracted to the magnet.

**[0005]**Greater efficiency in washing may be accomplished by moving the reaction vessels along a magnet array having a gap in the array structure proximate a wash position, allowing the magnetizable particles to be resuspended during the dispense of the wash solution. This is known as resuspension wash. Subsequent positions in the array include additional magnets, allowing the magnetizable particles to recollect on the side of the respective vessel.

**[0006]**Once the contents of the reaction vessel have again accumulated in a pellet on the side of the reaction vessel and the wash liquid has been aspirated, it is desirable to resuspend the particles in an acid reagent used to condition the bound component reagent. In the prior art, a single stream of acidic reagent is injected at the pellet. Because the size of the pellet and limitations on the volume and flow rate of reagent, insufficient resuspension may result. To address this inadequacy, prior art systems have resorted to the use of an additional resuspension magnet disposed on an opposite side of the process path from the previous separation magnets. The resuspension magnet is configured to assist in drawing paramagnetic particles into suspension, though the magnetic field is insufficient to cause an aggregation of particles on the opposite side of the vessel from where the pellet had been formed. In addition, since the prior art approach utilizes a resuspension magnet, there is less motivation to accurately aim the acid resuspension liquid. Any inhomogeneity in the suspended particles is addressed by the resuspension magnet.

**[0007]**It would be preferable to provide a system in which the use of a resuspension magnet is obviated.

**BRIEF SUMMARY OF THE INVENTION**

**[0008]**An improved acid injection module includes dual, parallel injection probes. A high-precision aiming strategy is employed to ensure that complete, homogenous resuspension of accumulated solid-phase particles is achieved, obviating the need for subsequent resuspension magnet positions.

**[0009]**The dual, parallel injector probe nozzles are spaced by a degree necessary to provide substantially adjacent impact zones on the reaction vessel wall, also referred to as "hit zones" or "hit points." Through careful control over lateral spacing of the two nozzles, and thus the two hit zones, and by performing an exacting analysis of the various physical tolerances which can effect hit zone location relative to the solid-phase pellet, thorough resuspension can be achieved without use of a resuspension magnet.

**[0010]**Other features, aspects and advantages of the above-described method and system will be apparent from the detailed description of the invention that follows.

**BRIEF DESCRIPTION OF THE SEVERAL VIEWS OF THE DRAWINGS**

**[0011]**The invention will be more fully understood by reference to the following detailed description of the invention in conjunction with the drawings of which:

**[0012]**FIG. 1 illustrates an optimal orientation of resuspension probes relative to a reaction vessel according to the presently disclosed invention;

**[0013]**FIG. 2 illustrates certain physical parameters employed in defining the optimal orientation of the probes of FIG. 1;

**[0014]**FIG. 3 illustrates additional physical parameters employed in defining the optimal orientation of the probes of FIG. 1;

**[0015]**FIG. 4 illustrates additional physical parameters employed in defining the optimal orientation of the probes of FIG. 1;

**[0016]**FIG. 5 illustrates additional physical parameters employed in defining the optimal orientation of the probes of FIG. 1;

**[0017]**FIG. 6 pictorially illustrates system components which contribute vertical tolerances and which must be accommodated in defining the optimal probe orientation of FIG. 1;

**[0018]**FIG. 7 is a vector diagram representation of the tolerance contributors of FIG. 6;

**[0019]**FIG. 8 pictorially illustrates system components which contribute horizontal tolerances and which must be accommodated in defining the optimal probe orientation of FIG. 1;

**[0020]**FIG. 9 is a vector diagram representation of the tolerance contributors of FIG. 8

**[0021]**FIG. 10 is a perspective view of a probe module according to the presently disclosed invention;

**[0022]**FIG. 11 is a front view of the probe module of FIG. 10;

**[0023]**FIG. 12 is a section view of the probe module of FIGS. 10 and 11 taken along lines A-A; and

**[0024]**FIG. 13 is a section view of the probe module of FIGS. 10 and 11 taken along lines B-B.

**DETAILED DESCRIPTION OF THE INVENTION**

**[0025]**The presently disclosed concept finds particular applicability to automated laboratory analytical analyzers in which paramagnetic particles are drawn into a pellet on the side of a reaction vessel as part of a separation and wash process. In particular, in an analyzer in which chemiluminescence is utilized for determining analyte concentration, the accumulated particles must be thoroughly resuspended to obtain an accurate reading. One approach in such systems is to resuspend the accumulated, washed particles in acid prior to introducing a base, and thus triggering the chemiluminescent response, at an optical measuring device such as a luminometer. However, it is noted that the presently disclosed concept is also applicable to any environment in which thorough resuspension of accumulated particles is required.

**[0026]**FIG. 1 illustrates a reaction vessel (also referred to as a cuvette), a probe nozzle, and the ideal orientation of the probe with respect to the cuvette. Note that two probes are employed, though only one is visible in the profile illustrated of FIG. 1. Linear distance values are given in millimeters. As shown, the ideal distance below the cuvette top plane where the liquid stream hits the cuvette wall, referred to as the hit point, is 25.98 mm. In the illustrated embodiment, this hit point is 5.74 mm above the centerline of a magnet array which forms the solids pellet and represents an empirically determined ideal locus of the hit point for achieving thorough particle resuspension. The probe is angled 6.9 degrees from vertical, with the probe tip being located 0.92 millimeters behind the cuvette centerline and 2.304 mm above the cuvette top plane. These values are obtained, as described below, by calculating the worst-case tolerance errors which could effect the hit point and by finding the locus where, even assuming all tolerances have a maximum deviation, the hit point will still be above the magnet centerline.

**[0027]**One practical aspect not accounted for in the configuration described in FIG. 1 is the effect of gravity on the liquid stream itself. The ideal hit point illustrated in FIG. 1 is calculated by extending the axis of the probe towards the cuvette wall. Because of the effect of gravity, the actual hit point is slightly below the one illustrated in FIG. 1. That distance is calculated in the following.

**[0028]**With respect to FIG. 2, given values include:

**TABLE**-US-00001 Pump flow rate V = 1300 μl/s Probe inner diameter id = 0.65 mm Probe inclination from vertical φ = 6.9° Vertical probe tip to hit point h = 27.868 mm Axial probe tip to hit point l = 28.071 mm

**[0029]**The speed v

_{0}of the liquid at the probe tip can be derived from the pump flow rate and the needle inner diameter:

**A**= π r 2 = π ( 0.325 mm ) 2 = 0.332 mm 2 ##EQU00001## v 0 = V . A = 1300 l / s 0.332 mm 2 = 1300 mm 3 s 0.332 mm 2 = 3915 mm s = 3.91 m s ##EQU00001.2##

**With reference to FIG**. 2, the horizontal distance between the probe tip and the cuvette wall can be calculated from:

**s**= {square root over (l

^{2}-h

^{2})}

**s**=3.37 mm=3.37 10

^{-3}m

**With reference to FIG**. 3, the arc of the liquid stream can now be calculated:

**with h**' = v 0 t sin α ' - g 2 t 2 and t = s v 0 cos α ' ##EQU00002## h ' = s sin α ' cos α ' - g 2 ( s v 0 cos α ' ) 2 = s tan α ' - g 2 ( s v 0 cos α ' ) 2 = 2.78 10 2 m - 2.52 10 - 4 m ##EQU00002.2##

**The first part of the term is equal to h and the second part gives the**difference between the ideal shown in FIG. 1 and actual hit point.

**[0030]**Overall, there are four tolerance chains which can affect the hit point:

**[0031]**1) Height tolerances--Tolerance summation of parts which affect the vertical gap between the probe tip and the cuvette top plane;

**[0032]**2) Axial tolerances--Tolerance summation of parts which affect the horizontal or axial gap between the probe tip and the central axis of the cuvette;

**[0033]**3) Angle tolerances--Tolerance summation of parts which affect the probe injection angle (any angle tolerances of the cuvette transport system are considered to result in a height error and therefore are considered part of the height tolerance chain); and

**[0034]**4) Magnet array--Tolerance summation of parts which affect the vertical distance between the cuvette and the magnet array centerline.

**[0035]**In the following, every tolerance chain is treated individually. Eventually, the total tolerance is estimated by adding the results of the individual tolerance chains.

**[0036]**The calculations for the individual tolerance chains are performed by executing the following steps:

**[0037]**Identification of related parts and their respective tolerances, providing a graphical description of the tolerance chain;

**[0038]**Graphical vector analysis of the tolerance chain;

**[0039]**Generation of a table of dimensions, tolerances, maximum dimensions, minimum dimensions;

**[0040]**Calculation of the ideal closure dimension;

**[0041]**Calculation of the arithmetic maximum and minimum closure dimensions and the arithmetic tolerance;

**[0042]**Identification of mean values from asymmetric tolerance zones and means values of shape and positional tolerances;

**[0043]**Generation of closure dimension as distribution average;

**[0044]**Identification of deviation σ/variance σ

^{2}for every dimension and calculation of the total error according to the theorem of error propagation; and

**[0045]**Evaluation of statistical closure dimension and tolerance.

**[0046]**The dimensions of all parts are considered to have normal, Gaussian distributions with a deviation of ±3σ. This means that 99.73% of all parts are inside the tolerance zone. This assumption is realistic for lot sizes of 60 to 100 parts and greater. The shape and position tolerances have a folded normal distribution.

**[0047]**For statistical calculation of the hit point tolerance, a mathematical description of the hit point depending upon linear position and angle of the probe is necessary. The arc of the liquid stream is omitted at this point for simplicity, but is factored in subsequently.

**[0048]**A simplified arrangement of a probe module and cuvette is shown in FIG. 4. The draft or outward curvature of the cuvette wall is omitted. h is the distance between the cuvette top plane and the hit point on the inner wall of the cuvette. The width of the cuvette is assumed to be constant. cw thus gives half the width of the cuvette such that cw=2.73 mm.

**h**= h g - y ##EQU00003## and h g = x + cw tan Φ ##EQU00003.2## h = x + cw tan Φ - y ##EQU00003.3##

**The draft angle**β, not taken into account in the foregoing, is 0.5°.

**[0049]**FIG. 5 illustrates the offset produced by the cuvette wall draft. The value h

_{r}has to be deducted from h to get the actual value of the hit point h

_{real}.

**h real**= h - h r ##EQU00004## with h r = l cos Φ ##EQU00004.2## and l = h sin ( 180 ° - β - Φ ) sin β h real = x + cw tan Φ - y - [ ( x + cw tan Φ - y ) 1 sin ( 180 - β - Φ ) sin β cos Φ ] ##EQU00004.3##

**(Eq. 1). Substituting the projected values from FIG. 1 into x, y, and φ as control gives the correct value for h**

_{real}, 25.98 mm.

**[0050]**Height tolerances are now considered with respect to FIG. 6, which illustrates all parts which add tolerances in height. These parts include a washer plate on which is mounted the acid injection probe module, the probe module, a cuvette transport ring segment in which the cuvettes are disposed, and a transport ring on which the ring segments are disposed. The transport ring is supported by a taper roller bearing and opposing circlips. Both the washer plate and the taper roller bearing/circlips are supported upon an incubation ring.

**[0051]**For the worst case in terms of height, it is assumed all tolerances are at their maximum, so that clearance between the washer plate and the cuvette is minimal. The hit point is thus lowered towards the bottom of the cuvette. To achieve this, parts of the left side of FIG. 6 must be at their minimum thickness whereas the parts on the right side must be at their maximum thickness. These requirements are illustrated in FIG. 6 by the large arrows.

**[0052]**The vector diagram of FIG. 7 shows all dimensions with their maximized or minimized direction. M0 is the so-called closure dimension, or the vertical gap between the probe tip and the cuvette upper plane. In the equations, this value is referred to as y. The const. vector sums the two constant values shown in FIG. 6, the thickness of the cuvette top plane and the vertical distance between the probe tip and the washer plate.

**[0053]**In the following table, all factors with the respective maximum and minimum values and resulting tolerance zones are provided:

**TABLE**-US-00002 Max. Min. vector Dimension dimension G

_{o}dimension G

_{u}Tolerance zone +const. 3.596 3.596 3.596 0 +M1 0 0.1 -0.1 0.2 -M2 90 90 89.95 0.05 +M3 6.15 6.17 6.13 0.04 +M4 1.75 1.75 1.69 0.06 +M5 15 15.2 15 0.2 +M6 1.2 1.2 1.14 0.06 +M7 52 52.04 51.96 0.08 +M11 0 0.2 -0.2 0.4 +M8 0 0.2 -0.2 0.4 +M9 5 5.1 4.9 0.2 +M10 3 3.1 2.9 0.2

**[0054]**The nominal closure dimension M

_{OH}:

**M**0 H = M i + - M i - ##EQU00005## M 0 H = 3.596 + 0 - 90 + 6.15 + 1.75 + 15 + 1.2 + 52 + 0 + 0 + 5 + 3 = - 2.304 ##EQU00005.2##

**[0055]**The arithmetic maximum closure dimension y

_{max}:

**y max**= G o i + - G u i - ##EQU00006## y max = ( 3.596 + 0.1 + 6.17 + 1.75 + 15.2 + 1.2 + 52.04 + 0.2 + 0.2 + 5.1 + 3.1 ) - 89.95 = - 1.294 ##EQU00006.2##

**[0056]**The arithmetic minimum closure dimension y

_{min}:

**y min**= G u i + - G o i - ##EQU00007## y min = ( 3.595 + ( - 0.1 ) + 6.13 + 1.69 + 15 + 1.14 + 51.96 + ( - 0.2 ) + ( - 0.2 ) + 4.9 + 2.9 ) - 90 = - 3.184 ##EQU00007.2##

**[0057]**The arithmetic closure dimension with tolerance zone is thus:

**M**0 H = y = 2.304 + 0.88 - 1.01 ##EQU00008##

**[0058]**Some statistical calculations are necessary to account for component fluctuations. The mean values from asymmetric tolerance zones M2, M4, M5 and M6 are now defined. For M2:

**μ 2 = 90 + 89.95 2 = 89.975 ##EQU00009##**

**Similar calculations for M**4, M5, and M6 yield:

**μ**

_{4}=1.72

**μ**

_{5}=15.1

**μ**

_{6}=1.17

**[0059]**As for M1, M8, and M11, shape and positional tolerances are distributed with a folded normal distribution. Mean values and deviations must therefore be calculated with the following equations. A deviation of 3σ is thereby assumed.

**σ 1 = F 1 3 = 0.2 3 = 0.066 μ F 1 = 2 σ 1 2 π = 0.053 σ F 1 = 1 - 2 π σ 1 = 0.04 M1 σ 8 = 0.133 μ 8 = 0.106 σ F 8 = 0.08 M8 σ 11 = 0.133 μ 11 = 0.106 σ F 11 = 0.08 M11 ##EQU00010##**

**[0060]**The closure dimension μ

_{0}H is calculated as a distribution average:

**3.596 + 0.053 - 89.975 + 6.15 + 1.72 + 15.1 + 1.17 + 52 + 0.106 + 0.106 + 5 + 3 = - 1.974 ##EQU00011##**

**[0061]**The deviation σ

_{0}H of the closure dimension:

**0.04 2 + ( 0.05 6 ) 2 + ( 0.04 6 ) 2 + ( 0.06 6 ) 2 + ( 0.2 6 ) 2 + ( 0.06 6 ) 2 + ( 0.08 6 ) 2 + 0.08 2 + 0.08 2 + ( 0.2 6 ) 2 + ( 0.2 6 ) 2 = 0.135 ##EQU00012## T SH = 6 σ 0 H = 0.81 ##EQU00012.2##**

**[0062]**The statistical closure dimension with tolerance zone is:

**M**

_{0}H=y=μ

_{0}H±(T

_{SH}/2)=1.974±0.405

**[0063]**Axial tolerances are now considered. FIG. 8 illustrates the components which contribute tolerances in the axial direction. The worst case is reached if the probes are displaced towards the inside of the incubation ring and the cuvette is displaced away from the probes. The large arrows in FIG. 8 illustrate these conditions.

**[0064]**The vector diagram of FIG. 9 the various contributing factors with the respective direction. M0 is the closure dimension, here the horizontal gap between the probe tip and the cuvette centerline. In the equations that follow, this value is identified as x.

**[0065]**The const. vector is the constant value shown in FIG. 8 and represents the horizontal distance between the probe tip and the cuvette centerline. The tolerance of this separation can be neglected due to the construction of a preferred instrument.

**[0066]**In the following table, all of the contributors with their maximum and minimum dimensions and tolerance zones are provided.

**TABLE**-US-00003 Max. Min. Vector Dimension dimension G

_{o}dimension G

_{u}Tolerance zone -const. 0.3 0.3 0.3 0 -M11 23.03 23.13 22.93 0.2 +M12 226 226.02 225.98 0.04 -M13 0 -0.05 0.05 0.1 -M14 215.87 215.92 215.82 0.1 +M15 14.12 14.22 14.02 0.2

**[0067]**The nominal closure dimension M

_{0}A is given by:

**M**0 A = M i + - M i - - 0.3 - 23.03 + 226 - 0 - 215.87 + 14.12 = 0.92 ##EQU00013##

**[0068]**The arithmetic maximum closure dimension x

_{max}is given by:

**x max**= G o i + - G u i - ( 226.02 + 14.22 ) - ( 0.3 + 22.93 + 0.05 + 215.82 ) = 1.14 ##EQU00014##

**[0069]**The arithmetic minimum closure dimension x

_{min}is given by:

**x min**= G u i + - G o i - ( 225.98 + 14.02 ) - ( 0.3 + 23.13 + ( - 0.05 ) + 215.92 ) = 0.7 ##EQU00015##

**[0070]**From these values, the arithmetic closure dimension with tolerance zone is given by:

**M**0 A = x = 0.92 + 0.22 - 0.22 ##EQU00016##

**[0071]**Some statistical calculations are necessary to account for component fluctuations. The mean values for shape and position for tolerance M13 are now defined.

**[0072]**M13: σ

_{13}=0.033 μ

_{F13}=0.027 σ

_{F13}=0.02

**[0073]**Closure dimension μ

_{0}A is given as a distribution average:

**-0.3-23.03+226-0.02-215.87+14.12=0.9**

**[0074]**The deviation σ

_{0}A of the closure dimension is determined from:

**( 0.2 6 ) 2 + ( 0.04 6 ) 2 + 0.02 2 + ( 0.1 6 ) 2 + ( 0.2 6 ) 2 = 0.054 ##EQU00017##**

**[0075]**The statistical closure dimension with tolerance zones is given by:

**M**

_{0}A=x=μ

_{0}A±T

_{SA}/2=0.9±0.162

**[0076]**Injector inclination tolerances are now addressed. The tolerance of the bores in the washer plate is M16=±0.05°. The parallelism of the axis of the probe bore and the axis of the injector outer diameter is M17=0.05 mm. With the length of 18 mm this results in an angle tolerance of:

**TABLE**-US-00004 Φ = M 16 + arctan ( M 17 18 ) ##EQU00018## Max. Min. Vector dimension dimension G

_{o}dimension G

_{u}Tolerance zero M16 6.9° 6.95° 6.85° 0.1° M17 0 0.05° -0.05° 0.1°

**[0077]**The nominal angle φ

_{0}is given by:

**Φ 0 = 6.9 ° + arctan ( 0 18 ) = 6.9 ° ##EQU00019##**

**[0078]**The arithmetic maximum angle φ

_{max}is given by:

**Φ max = 6.85 ° + arctan ( - 0.05 18 ) = 6.69 ° ##EQU00020##**

**[0079]**The arithmetic minimum angle φ

_{min}is given by:

**Φ min = 6.95 ° + arctan ( 0.05 18 ) = 7.11 ° ##EQU00021##**

**[0080]**The closure dimension with tolerance zone is thus given by:

**Φ 0 = Φ = 6.9 ° + 0.21 - 0.21 ##EQU00022##**

**[0081]**Some statistical calculations are necessary to account for component fluctuation. The mean values for shape and position for tolerance M17 are now defined.

**[0082]**M17: σ

_{17}=0.033° μ

_{F17}=0.027° σ

_{F17}=0.02°

**The average angle distribution**μ

_{0}φ is given by:

**μ 0 Φ = 6.9 ° + arctan ( 0.027 ° 18 ) = 6.986 ° ##EQU00023##**

**[0083]**The deviation of the angle error is given by:

**μ F 17 = 0.027 ° σ M 16 = ( 0.1 ° 6 ) σ F 17 = 0.02 ° ##EQU00024## σ 0 Φ = [ M 16 ( M 16 + arctan ( μ F 17 18 ) ) ] 2 σ M 16 2 + [ μ F 17 ( M 16 + arctan ( μ F 17 18 ) ) ] 2 σ F 17 2 ##EQU00024.2## σ 0 Φ = 0.017 ° ##EQU00024.3## T S = 6 σ 0 = 0.102 ° ##EQU00024.4##**

**[0084]**The statistical angle error with tolerance zone is thus given by:

**Φ 0 = Φ = μ 0 ± T s 2 = 6.9 ° ± 0.05 ° ##EQU00025##**

**[0085]**The worst case calculation for hreal can now be calculated by setting the arithmetic maximum values for x

_{max}, y

_{max}, and φ

_{max}into Eq. 1, above.

**x max**= 1.14 ##EQU00026## y max = 1.294 ##EQU00026.2## Φ max = 6.69 ° ##EQU00026.3## h real = x + cw tan ( α ) - y - [ ( x + cw tan ( α ) - y ) 1 sin ( π - β - α ) sin ( β ) cos ( α ) ] ##EQU00026.4## h real = 29.504 ##EQU00026.5##

**[0086]**The arithmetic minimum can be calculated using the analog:

**x min**= 0.7 ##EQU00027## y min = 3.184 ##EQU00027.2## Φ min = 7.11 ° ##EQU00027.3## h real = x + cw tan ( α ) - y - [ ( x + cw tan ( α ) - y ) 1 sin ( π - β - α ) sin ( β ) cos ( α ) ] ##EQU00027.4## h real = 22.725 ##EQU00027.5##

**[0087]**Thus, the arithmetic derivation of the hit point with tolerance zone is given by:

**h real**= 25.98 + 3.52 - 3.25 ##EQU00028##

**[0088]**The hit point μ

_{h}as distribution average with μ

_{0}H=1.974, μ

_{0}A=0.9, μ

_{0}φ=6.986° and employing Eq. 1:

**μ**

_{h}=25.812

**The statistical deviation**σ

_{h}of the hit point, depending upon the variables σ

_{0}A, σ

_{0}H, σ

_{0}φ, can now be calculated using Eq. 1. Using partial derivatives at the distribution average:

**σ h 2 = ( ∂ h real ∂ x ) 2 σ 0 A 2 + ( ∂ h real ∂ y ) 2 σ 0 H 2 + ( ∂ h real ∂ Φ ) 2 σ 0 Φ 2 ##EQU00029##**

**With**μ

_{0}H=1.974, μ

_{0}A=0.9, μ

_{0}φ=6.986° and σ

_{0}H=0.135, σ

_{0}A=0.054, and σ

_{0}φ=0.017°, the result is:

**σ**

_{h}=0.435

**T**

_{SH}=6σ

_{h}=2.61

**[0089]**The statistical error of the hit point with tolerance zone is thus given by:

**h real**= 25.98 ± T SH 2 = 25.98 ± 1.305 ##EQU00030##

**[0090]**In the embodiment in which the pellet is formed by a magnet array, the tolerance of the array relative to the cuvettes must also be accounted for. The magnets, in a preferred embodiment, are fixed in a ring which is suspended under the transport ring. Most of the tolerance of the magnets is addressed in the height tolerances previously calculated. Thus, there are only the following tolerances to be accounted for:

**[0091]**M18--slide bearing;

**[0092]**M19--magnet ring (i.e. the position of the magnet assembly in the magnet ring);

**[0093]**M20--magnet assembly (i.e. the tolerance of the fixture into which the magnet assembly is fixed); and

**[0094]**M21--the slide bearing support.

**All of the above contribute to movement in the same direction**.

**TABLE**-US-00005 Min. Tolerance Vector dimension Max. dimension G

_{o}dimension G

_{u}zone M18 4 4.1 4.05 0.05 M19 7.4 7.45 7.35 0.1 M20 0 0.05 -0.05 0.1 M21 1 1.1 0.9 0.2

**[0095]**The nominal closure dimension M

_{0}M is given by:

**M**

_{0}M=ΣM

_{i}

**4+7.4+1=12.4**

**[0096]**The arithmetic maximum closure dimension P

_{0}M is given by:

**P**

_{0}M=ΣG

_{0}i

**4.1+7.45+0.05+1.1=12.7**

**[0097]**The arithmetic minimum closure dimension P

_{0}M is given by:

**P**

_{0}M=ΣG

_{0}i

**4.05+7.35-0.05+0.9=12.25**

**[0098]**The arithmetic closure dimension with tolerance zone is thus given by:

**M**0 M = 12.4 + 0.3 - 0.15 ##EQU00031##

**[0099]**Mean values from asymmetric tolerance zone M18 is given by:

**μ**

_{18}=4.075

**[0100]**The closure dimension μ

_{0}M as a distribution average is found according to:

**5.075+7.4+1=12.475**

**[0101]**The deviation σ

_{0}M of the closure dimension is given by:

**( 0.05 6 ) 2 + ( 0.1 6 ) 2 + ( 0.1 6 ) 2 + ( 0.2 6 ) 2 = 0.042 ##EQU00032## T SM = 6 σ 0 M = 0.252 ##EQU00032.2##**

**[0102]**The statistical closure dimension with tolerance zone is thus:

**M**0 M = μ 0 M ± T SM 2 = 12.475 ± 0.126 ##EQU00033##

**[0103]**The nominal distance between the magnet centerline and the cuvette top plane at the acid injection position is 31.72 mm. This value can be calculated with the nominal dimensions listed above:

**3.9+12.4+6.35+5+3+1.067=31.717**

**(3.9 being the distance between the upper magnet and the magnet ring, 6.35 being the magnet width).**

**[0104]**The deviation σ

_{h}and the tolerance zone T

_{SH}of the hit point relative to the cuvette top plane was estimated above as 25.98±1.305 mm. The nominal measure between hit point and magnet centerline is thus:

**h**

_{total}=31.717-25.98=5.737

**[0105]**The total deviation σ of the difference between hit point and magnet centerline is thus calculated by:

**{square root over (0.435**

^{2}+0.042

^{2})}=0.437

**T**

_{s}=6σ=2.622

**[0106]**The statistical error of the hit point versus magnet centerline with tolerance zone can then be written as:

**h total**= 5.737 ± T s 2 = 5.737 ± 1.311 ##EQU00034##

**Once**0.25 mm is added to compensate for the arc of the liquid stream, the acid injection is calculated to hit the cuvette wall not deeper than 4.167 mm above the magnet centerline.

**[0107]**One embodiment of a probe housing 100 is illustrated in FIG. 10. This housing, which supports dual probe nozzles 102 is mounted in order to direct a parallel stream of liquid, preferably acid, above a pellet of particles such as paramagnetic particles which have accumulated on the interior wall of a reaction vessel such as a cuvette. By following the tolerance analysis procedure detailed above, the hit point for both acid streams can be assured to be above the pellet, regardless of variations in the physical components of the system.

**[0108]**The linear dimensions in FIGS. 11, 12 and 13 are all given in millimeters. A front view of the probe housing 100 is provided in FIG. 11, showing the mutually adjacent nozzles which produce parallel streams of resuspension liquid. In FIG. 12, a cross-section taken along lines A-A in FIG. 11, it can be seen that ideally a source of resuspension liquid is coupled to the back of the probe housing. As can be seen in FIG. 13, a cross-section taken along lines B-B of FIG. 11, the liquid source feeds both nozzles 102 in generating the parallel streams, five millimeters apart.

**[0109]**On the back of the probe housing 100 is a mounting recess 110 for interfacing to a resuspension liquid-supplying conduit (not shown). Secure attachment of the conduit to the housing 100 is preferably through interlocking threads or other means known to one skilled in the art. Preferably a buffer zone 112 exists between the forward end of the conduit once installed in the recess 110. Liquid from the conduit passes into the buffer zone and then into each of two channels 114 which lead to respective probes 116 and the probe nozzles 102 themselves. In the illustrated embodiment, the probes 116 and nozzles 102 are 0.65±0.02 mm in diameter.

**[0110]**Having described preferred embodiments of the presently disclosed invention, it should be apparent to those of ordinary skill in the art that other embodiments and variations incorporating these concepts may be implemented. Accordingly, the invention should not be viewed as limited to the described embodiments but rather should be limited solely by the scope and spirit of the appended claims.

User Contributions:

Comment about this patent or add new information about this topic: