Schardt2004_RotatingVesselXrayTube

New x-ray tube performance in computed tomography by introducingthe rotating envelope tube technology

Peter Schardt,a) Josef Deuringer, Jorg Freudenberger, Erich Hell,b) Wolfgang Knupfer,Detlef Mattern, and Markus SchildSiemens Medical Solutions, Vacuum Technology, Erlangen, Germany

~Received 30 January 2004; revised 16 June 2004; accepted for publication 28 June 2004;published 27 August 2004!

The future demands of computed tomography imaging regarding the x-ray source can be summa-rized with higher scan power, shorter rotation times, shorter cool down times and smaller focalspots. We report on a new tube technology satisfying all these demands by making use of a novelcooling principle on one hand and of a novel beam control system on the other hand. Nowadaystubes use a rotating anode disk mainly cooled via radiation. The Straton® x-ray tube is the first tubeavailable for clinical routine utilizing convective cooling exclusively. It is demonstrated that thiscooling principle makes large heat storage capacities of the anode disk obsolete. The unprecedentedcooling rate of 4.8 MHU/min eliminates the need for waiting times due to anode cooling in clinicalworkflow. Moreover, an electronic beam deflection system for focal spot position and size controlopens the door to advanced applications. The physical backgrounds are discussed and the technicalrealization is presented. From this discussion the superior suitability of this tube to withstandg-forces well above 20 g created by fast rotating gantries will become evident. Experience from alarge clinical trial is reported and possible ways for future developments are discussed. ©2004American Association of Physicists in Medicine.@DOI: 10.1118/1.1783552#

Key words: x-ray tubes, anode cooling, electron beam, cardiac CT

raa

papt

sp

sr

rae

fro-th

Jee

t ootu

troe

thngbanvitg

adearthens

thetheionwe

lope,yto-

roliza-

ae

bens is

I. INTRODUCTION

Medical x-ray diagnostics makes use of two types of x-tubes. For low power applications stationary anode tubesthe best compromise between technical design effort andformance. But for most midrange and high performanceplications in terms of focal spot power there is a needutilize rotating anode tubes. These tubes provide a focalpower of 60–100 kW for computed tomography~CT! as wellas angiography. Especially in CT, large amounts of dispated energy are produced by patient scans in the ordesome ten seconds. In contrast to the stationary anodesheat has to be stored in the anode disk and has to be tferred by radiation at high temperatures to the cooling mdium. The main components of such a tube can be seenFig. 1 ~taken from Ref. 1!. As a consequence of this bottleneck for the heat flow, the anode has to store nearlycomplete amount of energy up to several MJ~or expressed inthe more common unit Mega Heat Unit, MHU, where 1 Mequals 1.34 MHU!. The reason is illustrated in Fig. 2. Thtemperature in the focal spot is limited by the material usand for medical tubes somewhere below the melting pointungsten. This temperature for rotating anodes is a comption of the anode disk temperature and the short temperarise of the material passing the target spot of the elecbeam. The temperature rise is a function of material propties, focal spot size, focal spot power and the velocity ofmaterial~for rotating anodes a function of radius and rotatifrequency!. In contrast the disk temperature is determinedthe heat capacity and the applied cooling mechanism. Mimprovements to cool the disk deal with enhanced emissi« or increased areasAR of radiation emitting and absorbin

2699 Med. Phys. 31 „9…, September 2004 0094-2405 Õ2004Õ3

yre

er--

oot

i-of

thens--m

e

df

si-renr-e

yyy

surfaces~optimized for x-ray tubes, e.g., in Ref. 2!. The goalis to maximize the heat flow PR given by theStefan–Boltzmann3 radiation law,

PR5«•AR•~TA42TH

4 !, ~1!

with TA being the temperature of the anode andTH the tem-perature of the housing. Few attempts have also been mto introduce convective cooling by using liquid metal to bethe anode disk and achieve a modest improvement ofcooling.4 Cooling rates of up to 1.4 MHU/min have beereported.5 However, for a further increase of cooling rateone possibility is to get the anode disk itself in contact tocooling medium. One solution for that is a design whereanode disk is part of the tube envelope. This implies rotatof the entire tube with respect to the anode axis, whichrefer to as the class of rotating envelope tubes~RET! incontrast to the rotating anode tubes, where the tube enveis stationary. The Straton tube~Siemens Medical SolutionsErlangen, Germany! is the first tube based on this technologintroduced into the high performance class of computedmography with outstanding power reserves.

Physical details of convective cooling and beam contare discussed in Sec. II: the impact on the technical realtion is presented in Sec. III. We report the results fromlarge clinical field trial in Sec. IV before an outlook to futurdevelopments is given in the concluding section~Sec. V!.

II. PRINCIPLE AND PHYSICAL ASPECTS

A very early appearance of the idea to rotate the tuenvelope and deflect an electron beam by stationary mea

26991„9…Õ2699Õ8Õ$22.00 © 2004 Am. Assoc. Phys. Med.

,

ox---

2700 Schardt et al. : Straton x-ray tube 2700

FIG. 1. Main parts of a medical diag-nostic x-ray tube~1: cathode, 2: anode3: rotor, 4: ball bearings, as taken fromRef. 1!. The large anode disk has tstore the heat generated during the eposure and radiates at very high temperatures to cool down in the intermediate times.

lebyIn

mt,ula

odthth

sitinrmi

dihe

-

rg-herec-

-0ngghEq.

s-n-

ateattelyfore it

ibi

bly.

first documented in a patent of 1917.6 This design is in factvery close to the realized Straton tube. Within this principthe electron beam in the tube is shaped and controlledmagnetic field similar to a miniature electron beam CT.Fig. 3 one can recognize the vacuum housing rotatingbearings outside the vacuum and driven by an externaltor. The tube housing itself consists of a cathode shafceramic high voltage insulator, a tube envelope, an annx-ray window and the anode disk mounted on a second shFor running the tube, a drive motor is attached to the anshaft and a magnetic deflection system is positioned inplane of the tube waist to deflect the electron beam ontofocal spot position. In the following we discuss the phycal aspects of the two major differences between rotaenvelope tubes and rotating anode tubes, namely the thedynamics in Sec. II A and the electron beam dynamicsSec. II B.

A. Thermodynamics

In this concept the cooling surface of the anode is inrect contact with the cooling fluid. The heat transfer for tso called convective coolingPC is given by

FIG. 2. Temperatures of anode disk~dotted! and focal spot~solid! for aspecific applied electrical power at the focal spot. The comparison exhthe difference in temporal behavior for a rotating anode tube~radiationcooling! and a corresponding rotating envelope tube~convective cooling!.Numerical values are for illustration only.

Medical Physics, Vol. 31, No. 9, September 2004

,a

ino-aarft.eee

-go-

n

-

PC5a•AC•DT, ~2!

wherePC is the transferred power,a the heat transfer coefficient, AC the cooling surface andDT the temperature dif-ference between cooling surface and cooling fluid. The laest values ofa are achieved for turbulent flow, because tturbulence causes additional material transport in the dition perpendicular to the cooling surface.7 With the help ofFEM-calculations~numerical computer simulations using finite element methods! one can derive values of 36.00W/m2K for an anode disk of 60 mm radius and a rotatispeed of 150 Hz, as illustrated in Fig. 4. To obtain a rouestimation for the heat flow, one can insert numbers in~2! for the cooling surface area~85 cm2!, the temperaturedifference~250 K! and a meana of 30.000 W/m2K to end upwith a heat transfer rate of more than 60 kW directly tranferred to the cooling fluid. This value corresponds to an uprecedented cooling rate of 4.8 MHU/min compared to stof the art performances of 0.8–1.4 MHU/min. LookingFig. 5 one can recognize that in 20 s the anode is complecooled down to the oil temperature. The cool down timea radiation cooled anode is also visible. With this exampl

ts

FIG. 3. Principle of operation and main parts of a RET housing assem

rant

tc

uab

tioco

tionthbythth

.us-

in-

gncy.lyisge-bleicu-pelly

large

foran

at-thede-uchpotgeV,d ofingge.amces

ted

rnd

Then

alto

h Bnds

reen.


becomes clear that there is no need for a large heat stocapacity at the anode any longer. However, a small amoustill needed for the anode to act as a heat spreader fromarea of the focal spot to the large area of the cooling surfa

Using this setup, the rotation of the anode serves a dpurpose: inside of the tube the rotation maintains a tolerafocal spot temperature rise, outside of the tube the rotacauses turbulent oil flow to maintain large heat transferefficients.

Nevertheless, the requirements of turbulence and rotaspeed result in a substantial amount of dissipated frictiopower which has to be provided by the external motor. Infollowing we describe a way to reduce the drive powergeometrical implications. The functional dependencies offrictional power on geometric and physical properties ofrotating part can be explained by8,9

Pdrive}rn0.2•hRtube

3.9 d20.3• f 2.8, ~3!

where Pdrive is the driving power,r the density andn thecinematic viscosity of the fluid andf the rotational frequencyEquation~3! has been derived for a solid cylinder of radiRtube and heighth rotating in a stationary cylinder. The rotating cylinder and the fixed surrounding housing~stationarycylinder! define an intermediate volume between the cylders of thicknessd, where the friction in the fluid is gener

FIG. 4. Local heat transfer coefficienta as a function of distance from theaxis calculated for the Straton geometry with a tube radius of 60 mm.radius of the focal track is 48 mm. Data are given for two rotating frequcies and are obtained by a fluid dynamical FEM simulation.

FIG. 5. Cooling curve comparison with data from time resolved FEM cculations. Upper curve for rotating anode tube; lower curve for the Straanode.


geishee.l-len-

nale

ee

-

ated. From Eq.~3! it becomes clear that there is a strondependency on the tube radius and the rotational frequeThe viscosity of the fluid and the width of the slit have ona minor impact. A major task for keeping drive power lowtherefore an optimization of the radial dimensions of theometry. As a result of this we found a very compact doucone shaped design described in detail in Sec. III. In partlar we found other solutions to realize a rotating envelotube ~for example, a tube with a beared and magneticafixed cathode10! to exhibit much larger drive power for agiven anode diameter and rotating frequency, because aradiusRtube has to be maintained over a long heighth.

B. Electron beam dynamics

The second physical aspect covers the requirementsthe electron beam deflection and focusing. To achievex-ray focal spot fixed in space while the whole tube is roting, a stationary beam deflection system is necessary. Ineasiest way this is achieved by a dipole magnet systemflecting the electron beam generated on the tube axis sthat the beam hits the anode disk exactly at the focal sposition of the tube housing assembly. In the tube voltarange used for medical imaging of about 40 kV to 150 kthe electrons gain a speed up to about 50% of the speelight. Therefore, the Lorentz force has to be calculated usthe relativistic electron mass and is nonlinear in this ranThe basic formula for the deflection of the electron bewill be found by the basic equations for the Lorentz forand the centrifugal force. The result is the curvature radiuRof the electrons,

R5mn

e0B, ~4!

with m being the relativistic mass of an electron acceleraby the voltageU. Using a simple model~see Fig. 6!,the distanceS of the focal spot from the rotation cente~5focal track radius! depends only on the geometric data aradiusR,

e-

-n

FIG. 6. Simplified model to calculate the required magnetic field strengtfor a given geometry and accelerating voltage U. The magnetic field bethe beam with curvature radius R over the length L1. After the field fsection of length L2 the electrons hit the anode at the focal spot positio

ld

anrot

eeatioeoelticd

nu

g

s

ore

qua-

denin aingfors iscel-pa-

tM

ica-to

u-aineds

ag-entc-eticul-

derformibleticthe

ldthisg

s in

ns of


S5S 12sinS a cosS L1

R D D D •R1L2

tanS a cosL1

R D . ~5!

Using Eqs.~4! and~5! we obtain the required magnetic fiestrengthB for a given deflectionS, which is plotted as afunction of the tube voltageU in Fig. 7.

Despite the nonlinear function for the deflection fieldadditional aspect complicates the situation. The electbeam length~way of electrons from emitter to the focal spo!is about 115 mm and therefore considerably longer thanconventional x-ray tubes~see for example Fig. 1, where thbeam length is about 30 mm!. This is due to the fact that thbeam is created on the rotational axis of the tube, accelerin the direction of the axis and then deflected to the positgiven by the focal track radius. As a consequence therspace charge effects play an important role, because thetrons are not fully relativistic. The electric and magnefields within a long beam with cylindrical symmetry anradiusr are11

Er51

«0r E0

r

r 8r~r 8!dr8, ~6a!

Bw5m0

r E0

r

r 8 j ~r 8!dr8. ~6b!

Upon an individual charge, these fields inflict repulsive aattractive forces, respectively, according to the basic eqtion of motion,

d

dt~m•nW !5qEW 1qnW 3BW . ~7!

Inserting ~6a!, ~6b! into ~7! one ends up with the resultinradial force,

F tot5q

rn«0@12b2#•E

0

r

r 8 j ~r 8!dr8, ~8!

which is defocusing for all but highly relativistic particle(b5v/c, the fractional speed of light!. From ~7! and ~8!, atrajectory equation can be derived under the assumptionhomogeneous radial current density. With constant cur

FIG. 7. Field strength for magnetic beam deflection system as functiotube voltage U.


n

in

ednf,

ec-

da-

f ant

and energy, one can furthermore generate the envelope etion, which gives the beam radiusRbeamas a function ofz inthe implicit form

d2

dz2 Rbeam~z!5qI

2p«0m0~gbc!3 •1

Rbeam~z!. ~9!

This differential equation~9! can be solved numerically anRbeam is plotted in Fig. 8. It can be seen that for the givbeam parameters a beam with a large beam length likeRET is much more spread than in a conventional rotatanode tube. However, this calculation example is onlyillustration, because none of the given ideal assumptionaccomplished in the real tube. In the Straton tube the aceration, deflection and focusing of the beam are not serated in space. Properties of the real beam~e.g., beam size afocal spot position! have therefore to be calculated with FEcalculations.

Tube current and voltage are determined by the appltion and therefore variable. Both are selected in orderoptimize CT image quality with respect to the given attenation properties of the human body. To maintain a certrange of operation with stable focal properties, one neadditional focusing in order to compensateF tot . To supplythis additional force, a quadrupole component of the mnetic field may be used in addition to the dipole componof the deflection field. The calculation of the electron trajetories under the influence of these superimposed magnfield components, while the electrons are accelerated simtaneously, is easily done by FEM calculations even unconsideration of the space charge effects. An examplesuch a calculation is given in Fig. 9. Ignoring the vacuuenvelope of the tube, the tracks of the electrons are visstarting at the emitter and traveling through the magnedeflection and focusing field before reaching the anode atfocal spot.

III. TECHNICAL REALIZATION

Although the idea of turning the tube housing is very oand the underlying physics is easily understandable,technology did not find its way to routine medical imaginbefore 2003. This is mainly because of some obstacle

ofFIG. 8. Beam spread for an electron beam calculated with an initial radiu2.5 mm at the emitter as a function of beam length~100% is the beamdiameter at the emitter! in comparison for rotating anode tubes~typicalvalue! and a RET~120 kV tube voltage and 500 mA beam current!.

agailithidfoa

rolaecmesel

datenthhe

oe

oodin

inexthig

ththi-

. Tac

then-the00

, ata-gelarhthee

esans.ol-

at

gth

leftctive

itter.


technical realization, which had to be overcome. From imquality requirements on focal spot stability a need forintelligent beam control can be derived. Position and stabin shape are crucial especially in computed tomograpTherefore a microcontroller is essential, in particular consering the mostly nonlinear dependencies as discussed beHowever, also mechanical aspects have to be taken intocount. High performance tubes have to run at very hightation speeds and one can imagine, that the traditional gtube technology is not adequate in terms of required prsion and stability. Since the early nineties, the metal ceratechnology is routinely used for medical tubes with highpower demands and is a key factor for manufacturing rable rotating envelope tubes.

Out of these reasons, we describe the realization andcuss some properties of the key components of the Strtube, namely the tube envelope in Sec. III A and the filamfor electron emission in Sec. III B. In Sec. III C we deal withe magnetic deflection system and in Sec. III D with tcooling system.

A. Tube envelope

In Fig. 10 the Straton tube is shown with an anode disk120 mm diameter capable to rotate at 150 Hz. The tubevelope made of nonmagnetic stainless steel is attached tanode disk made of a tungsten zirconium molybdenum bequipped with a tungsten rhenium focal spot track includan annular x-ray window of 0.2 mm thickness12 made out ofstainless steel. In contrast to all other x-ray tubes, the wdow is rotating through the x-ray beam and has to betremely uniform in terms of x-ray attenuation. On the caode side, a ceramic disk insulator ensures a reliable hvoltage insulation of about 170 kV in the vacuum~inside! aswell as in the cooling oil~outside!. The geometries of all thecomponents were chosen very carefully in order to keepdrive power acceptable. The power required to maintainrotation and turbulent flow of the insulating oil at the maxmum rotational speed of 150 Hz does not exceed 6–7 kWkeep frictional power losses low, the vacuum housing is

FIG. 9. FEM calculation of electron trajectories from the cathode throuthe complex magnetic deflection and focusing field to the focal spot atanode.


enyy.-re.c--ssi-icti-

is-ont

fn-they

g

--

-h

ee

o-

celerated~in about half a second! to this high speed onlywhen the maximum electron beam power is impinginganode. All other applications run at lower rotating frequecies with an intelligent power management. Therefore,average driving power can be kept below a level of 500–6W only.

B. Electron emission

Much attention has to be spent on the filament. Ideallycircular uniform emission is needed to avoid unwanted rotional focal spot size modulations, which may reduce imaquality in computed tomography. Our tube uses a circuemitter made of 100mm thick tungsten sheet material, whicis cut by the laser in a way to form meander like paths forfilament heating current.13 In Fig. 11 one can recognize thstructure of the flat emitter inside the focusing device.

This flat emitter technology has important advantagconcerning high speed dose variations during fast CT-scThe limiting parameter for fast emission changes is the coing time of the emitter. Assuming radiation cooling only,high temperatures~.2000 °C! the rate of change for theemitter temperatureT is given by

he

FIG. 10. The Straton tube with a cathode and ceramic insulator at theand an anode at the right. Ball bearings are mounted as well as the indufilament transformer. The largest diameter of the tube is 120 mm.

FIG. 11. Details of the Straton cathode. Inside the focusing head, the emmade of tungsten sheet is visible. The emitter has a diameter of 5 mm

nac

,t

ith

ontana

ucind

e

thca

tfg1

pots

ve.ag-

isbe

byfor

thehed inin-heal

achthe

t issize

ithesthe

thealis

Toe,

heghis

inilouscethehebebe

ainge

eatge

the

ton


dT

dt5

1

csmE•Q, ~10!

with cs andmE denoting the specific heat capacity, mass aQ the rate of change of the thermal energy. Taking intocount, that the temporal change ofQ is caused by radiationand equals therefore the radiated PowerPR given by Eq.~1!,one ends up with

dTE

dt5

«

cs•

AE

mE~TE

42TS4!, ~11!

AE being the area of the radiating surfaces,TE and TS thetemperatures of the emitter and the surrounding surfacesspectively. The rate of temperature decline is dependenthe geometry factorAE /mE for a given material~« and cs)only. Inserting values for a typical helical tungsten wire emter (AE593 mm2,mE5250 mg) and comparing them witthe new developed flat emitter (AE540 mm2,mE537 mg), afactor of three is gained in variation speed. This is in goagreement with measurements. For dose managemeclinical applications, this is a crucial benefit in achievingaccurate dose profile as a function of gantry rotation, in pticular for highest rotating speeds in cardiology.

The tube rotating in lubricated ball bearings has an indtive heat transformer, which transmits the filament heatpower from a stationary primary coil to the rotating seconary coil ~visible in Fig. 10! to avoid any degradation of thinsulation oil.

C. Magnetic deflection system

The deflection system has to perform three tasks: firstdeflection of the beam in the radial direction onto the fospot and to perform a flying focal spot in thez-direction;second the focusing of the beam to determine the size ofelectronic focal spot; and third a deflection to make use oflying focal spot in the phi-direction. For this purpose a manet design similar to a quadrupole is chosen of which Fig.gives a schematic sketch. In contrast to a pure quadrumagnet, the Straton magnet system has three different se

FIG. 12. Scheme of the magnet system with different coils.


d-

re-on

-

din

r-

-g-

el

hea-2leof

coils14 corresponding to the three tasks mentioned aboThe r-coils generate the main dipole component of the mnetic field to deflect the beam onto the focal spot, whichquasistatic in time disregarding the switching when the tuvoltage is turned on or off. The focusing is performedfour q-coils generating a quadrupole component usedforming the electronic focal spot. The third setup~phi-coils!are used to produce a small but fast dipole component forflying focal spot in the tangential direction. Because all tcoil currents depend on the value of the high voltage usethe way as described in Sec. II, a microcontroller drivesdividual current power amplifiers for each coil setup. Tmicrocontroller calculates the required coil current in retime as a function of the measured high voltage. For efocal spot size, a set of coil current tables is managed bymicrocontroller. So, an electronically adjustable focal sporealized having the advantages of low tolerances in spotdimensions, fast switching and high flexibility.15 For ex-ample, our RET uses three different focal spot sizes wonly one emitter, which can be used to adapt to the bimage quality possible. Table I illustrates the variety of tspot sizes for Straton.

D. Cooling system

The cooling system used for transferring the heat fromoil to the gantry air is designed for the large centrifugforces at gantry rotation times of 0.37 s per revolution andcapable of cooling oil temperatures of about 100 °C.avoid any cavitation in the oil due to the fast rotating tubwhich may result in a reduced high voltage stability, tcooling system works at a pressure of about 1 bar. Althouthere is an oil pump in the system, the main oil flowcaused by the rotating tube.16 We measured 25 L/min whenthe tube is rotating, resulting in a very effective coolingcontrast to 8 L/min with only the oil pump running. The oto air heat exchanger is designed to cool down a continupower of 7 kW, which is appropriate for high performanCT. However, this heat exchanger is the bottleneck formaximum continuous power now. While the heat from tanode is directly transferred to the cooling oil, it has tostored by the heat capacity of the oil contained by the tuhousing assembly and the cooling system, until it is agcooled down by the gantry air. In this situation the averapower is only determined by the cooling capacity of the hexchanger. Therefore, in principle much larger averapower would be possible with an appropriate design ofcooling system.

TABLE I. Three electronically adjusted sizes for the focal spots of Straand their maximum load~for 20 s!.

Focal spotSpot size (w3 l,

IEC 60336 values!Maximum load~20 s scantime!

F I 0.630.7 42 kWF II 0.830.8 51 kWF III 0.831.2 60 kW

eld

ne

nbea

s prkg.

t 5An

sentia

tionner

ng.ionndt ofheighhteringnalthe

ing

seer,gethe

anal

be-be

-forli-l

inggal650outperWethe

up

inals.

on


IV. PERFORMANCE RESULTS AND CLINICAL FIELDTRIAL

The Straton tube was intensively tested in a clinical fitrial on the SOMATOM Sensation 16 scanner~SiemensMedical Solutions, Forchheim, Germany!. About 25 installa-tions running from Feb/03 to Oct/03 were monitored onlivia remote control to observe the clinical performance.

The installations with these tubes exhibited essentiallytube cooling times, confirming the difference in the tucooling performance. One clinic is running the tube45.000 scanseconds per month with about 50–60 patientday. An impressive number for the demonstration of a woflow no longer limited by waiting times due to tube coolinThe focal spot F II~medium size, see Table I! was used formore than 90% of all scans, resulting in a statement thakW scanpower is appropriate for routine examination.UHR-mode~ultra high resolution mode! utilizing the FI aswell as the high power modes were only used in a few ca

The Straton tube is the first tube routinely used at a garevolution time of 0.37 s. This mode was used for all card

FIG. 13. Cardiac scanning at 0.37 s per revolution demonstrating the srior resolution applied to the beating heart~courtesy of Dr. A. Ku¨ttner, Univ.of Tubingen, Germany!.

TABLE II. Centrifugal forces at tube axis as function of gantry rotatispeed.

Gantry revolution time

Centrifugal accelerationon tube axis~650 mm! inunits of g (59.81 m/s2)

1 s 30.5 s 100.4 s 160.3 s 300.2 s 65


o

ter-

0

s.ryc

investigations, because of the unprecedented time resolufor a rotating gantry scanner only surpassed by EBT-scantechnology~electron beam tomography!. It demonstrates thebig step forward being made in noninvasive cardiac imagiAn impressive image of the heart taken at 0.37 s revoluttime given in Fig. 13 underlines the possibilities today ademonstrates the strong demand for further improvementemporal resolution. The field trial clearly confirmed, that tball bearings no longer cause tube failures even at the hcentrifugal forces. Here, the small tube with the low weigof only 3.6 kg for the entire tube and the lubrication of thbearings resulted in no observed breakdown due to beafailure. This has to be seen in direct contrast to conventiorotating anode tube technology, where bearings insidevacuum are difficult to lubricate, because of high operattemperatures.

The overall increased lifetime capability could of cournot be demonstrated during the short trial period. Howevwe observed a significant lower number of high voltabreakdowns~mainly because there are no moving parts invacuum! and a lower focal track degradation resulting inmore constant x-ray output comparable to conventiotubes.

V. CONCLUSIONS AND FUTURE DEVELOPMENTS

From all the physical and technical aspects discussedfore, some conclusions for future developments candrawn.

~1! Although the demand for a clinical workflow not limited to cool down times at any situation is a strong driverthe convectively cooled tube technology, the clinical appcations still will ask for further improvement in temporaresolution for cardiac scanning and will need faster rotatgantries in consequence. Table II shows the centrifuforces for future CT machines, assuming a tube axis at amm radius from the center of rotation. Running now at ab20 g the next generation of scanners will perform 0.3 srevolution being not the end of the scanner development.demonstrated that the rotating envelope technology is

e-

FIG. 14. Flying focal spot with the magnetic deflection system not onlyphi, but also in thez-direction. 1: anode disk; 2: focal spot track; 3: focspot position; 4: electron beam; F1, F2: x-rays from different focal spot

mast

e.

iesmrothpa

eof t

irr

ho

ecneitsgc

542,

://

.

ter-

453,

538,

771,

942,

AG,


most appropriate way to cardiac CT scanning with high tiresolution. This has to be seen as an alternative to largerheavier rotating anodes carried by bearings, which arehard to lubricate in vacuum. Maybe the EBT~electron beamtomography! technology is comparable with respect to timresolution, but suffering from other known disadvantages

~2! The microcontroller on board opens new possibilitto improve image quality in computed tomography. In cobination with the magnetic deflection system, the micropcessor controls the beam in two directions of space atfocal spot position. Besides the already mentioned focalsition stabilization, one is now able to create a flying focspot not only within the plane of the gantry rotation~the leftside of Fig. 14!, but also in the radial direction13 ~the rightside of Fig. 14!. This opens the opportunity to acquire twicthe number of slices per revolution with a given detectbecause due to the anode angle a radial movement ofocal spot corresponds to a shift of the focal spot positionz ~patient axis!. Moreover, the microcontroller can monitotube performance and wear out of components in ordepredict the remaining tube lifetime.17 Knowing this number,service calls can be scheduled for a tube exchange witdisturbing the clinical workflow.

~3! In times of compact electronics and solid state dettors, the anode diameter of the x-ray tube mainly determithe size of a CT gantry. Here the RET technology withsmart tube housing assemblies opens a way to compacttry design. It enables larger bores and a better patient acsibility as a future benefit for users.


endill

--e

o-l

r,hen

to

ut

-s

an-es-

a!Electronic mail: [email protected]!Now at BSH Bosch und Siemens Hausgera¨te GmbH, Product Area Re-

frigeration, Giengen, Germany.1H. Bittorf, in Imaging Systems for Medical Diagnostics, edited by E.Krestel ~Siemens AG, Berlin, 1990!, p. 234.

2G. Lavering and R. Treseder, Varian Associates, US Patent 5,689,1996.

3C. Gerthsen, H. Kneser, and H. Vogel, inPhysik, edited by H. Vogel~Springer-Verlag, Berlin, 1989!, p. 219.

4T. Schmidt and R. Behling, Medica Mundi44Õ2, 50 ~2000!.5Website of TOSHIBA Medical Systems, Megacool™ X-ray tube, httpwww.medical.toshiba.com/clinical/radiology/aquilionmulti16-360-360-431.htm, 2004.

6W. Coolidge, General Electric Company, US Patent 1,215,116, 19177K. Stephan, inTaschenbuch fu¨r den Maschinebau, edited by W. Beitz andK.-H. Grote ~Springer-Verlag, Berlin, 1997!, p. D29.

8G. I. Taylor, Proc. R. Soc. London, Ser. A151, 494–512~1935!; ibid.157, 499–523~1935!; ibid. 157, 565 ~1936!.

9H. Schlichting and F. W. Riegels,Grenzschicht-Theorie, Braun-Verlag, 8.Aufl., pp. 491–513~1982!.

10See, for example, J. E. Burke, L. Miller, and S. G. Perno, Picker Innational Inc., US Patent 5,274,690, 1993.

11J. Großer,Einfuhrung in die Teilchenoptik~B. G. Teubner, Stuttgart,1983!, p. 124.

12E. Lenz, P. Rother, and P. Schardt, Siemens AG, patent pending~to bepublished!.

13E. Hell, D. Mattern, and P. Schardt, Siemens AG, US Patent 6,115,2000.



16E. Hell, T. Ohrndorf, and P. Schardt, Siemens AG, US Patent 6,084,2000.

17J. Deuringer, P. Schardt, M. Schild, and J. Freudenberger, Siemenspatent pending~to be published!.

Documents

Schardt2004_RotatingVesselXrayTube