John M. Constantine, Jr.
CELCO / Yoke Designer
14 Industrial Avenue, Mahwah, New Jersey 07430
The purpose of this paper is to describe some of the electron optical techniques developed to achieve state-of-the-art performance in a cathode ray tube based, ultra high resolution color film recording system.
Proper selection of cathode ray tube, deflection yoke, magnetic focus lens, astigmatism correction, centering and aperture flooding coils will be discussed.
Absolute control over the size, shape and position of the scanning electron beam spot is necessary to produce a display capable of yielding over 7500 resolvable points across the CRT face with resulting image content in excess of 31 million picture elements per field.
No single factor is responsible for the kind of performance outlined here. Rather, it is a refinement of a combination of disciplines integrating optic, electron optics, analog and digital electronics.
High resolution cathode ray tube scanning systems have been used for many years and for many diverse applications. In the early to mid-sixties much work was being done in the field of nuclear research. Here, under precise computer control, the CRT spot was used to follow the tracks of high energy particles in cloud chamber and bubble chamber photographs in search of meaningful nuclear events. This project was referred to as Precision Encoder and Pattern Recognition (PEPR).
CRT based Computer-On-Microfilm (COM) machines made it possible to store huge amounts of data in document form directly from the computer data base. Digitizing master fonts using CRT scanners and the development of the high speed photo-typesetter made other methods of preparing newspaper printing plates obsolete. High-speed finger print scanning for storage and subsequent identification of criminals is now accomplished utilizing a high resolution CRT display.
Satellites and deep space probes send electrical signals to earth that are converted into their informative and often stunning pictorial counterparts via the high resolution CRT display.
CELCO's continuing involvement in the research, development and manufacture of beam formation control and deflection components, related drive electronics and measurement instrumentation has placed the company in a unique position to design and manufacture an advanced CRT color film recording system with performance unmatched in the industry.
This system is presently providing theatre-quality imagery to the rapidly growing computer graphics industry for computer animation and other computer image generation techniques of recent development.
Other applications of the ultra-high-resolution color film recording system include imagery for oil exploration analysis and other geological interpretations.
The size of the spot produced by the CRT and associated electron optics is the limiting factor in determining the amount of information that can be gained from a given CRT display. Spot size as measured at the phosphor plane of the CRT is defined as the width at the 50% point of the Gaussian distribution of light energy in the spot as described by John Constantine, Sr.1, in the paper titled "Two-slit CRT Spot Analyzer Measurement". This technique has been recognized as the industry standard for defining CRT spot size.
In order to fully characterize the amount of information available from the display, a map of the spot size over the entire useful CRT face must be known. This spot must remain stable without jitter or drift in order to maintain the full resolution capability of the spot. the spot size of the CRT display used in the CELCO Color Film Recording System is 0.8 mil (+ or - 10%) at any point within a 6-3/8" useful diameter.
There is an important distinction to be made between the term resolution and the term addressability. Resolution refers to the spot size at the 50% level of the Gaussian curve as previously discussed. This is a measure of the True Image Content capability of the display.
Addressability refers only to the discrete number of positions to which the hardware D/A converter is able to
move the spot along the scan line.
On a typical 4" x 5" inscribed format on the CELCO Machine, this would yield approximately 5000 x 6250 resolvable pixels, or over 31 million resolvable points. By comparison, in terms of addressability, the hardware can move the beam to any one of 8192 locations in x and 8192 locations in y or 63 million addressable locations. (See Figure 1.)
One should not be misled into believing that the number of addressable positions has any relation to the True Image Content or spot size of the display.
Basic to the overall performance of the CRT display is the cathode ray tube itself. Therefore, proper selection is essential.
A magnetically deflected and focused CRT will achieve the highest performance in terms of resolution and brightness. This type of CRT has the ability to operate more practically at higher anode and G2 potentials than it's electrostatic counterpart. Anode voltages of 25 to 30 KV are typical. With magnetic focus, one is able to precisely align the focus lens, and due to larger lens diameter, a better quality lens is produced when compared to the electrostatic type.
The relative ease in applying fast dynamic focus changes in a magnetic lens compared to making rapid
variations of a high voltage, strengthens the case for magnetic focus.
High voltage operation enables the use of electron gun design techniques producing a small crossover region and small beam bundle with good light output. No loss of electrons occurs in the magnetic focus field as in the electrostatic lens. Magnification or demagnification of the CRT beam crossover is determined by the equation for a simple lens, M=I/O, where I is the image distance from the magnetic focus lens to the phosphor screen and O is the distance from the crossover region to the focus lens. (See Figure 2.)
The electron gun must be sturdy, thermally stable, of accurate construction and processed in a clean environment
to minimize arcing and secondary emissions. Low astigmatic abberation for smallest possible spot size is also a
desired quality for achieving optimum CRT performance.
CRTs used in film recording operations must be free from stray emissions. Stray emission is the result of electrons being emitted by a source other than the cathode, such as a sharp metal point in the gun structure. These stray electrons are not controlled by the G1-cathode bias. Therefore, even when the CRT beam is cut off these electrons will continue to reach the phosphor screen and expose the film when no exposure is called for.
A white phosphor with the required spectral characterisitics of the red, green and blue components must be applied to the faceplate using fine grain deposition techniques for resultant low noise operation. The phosphor must be resistant to aging to allow for consistent operation of the system over long periods of time. An optically flat faceplate arranged to maintain precision optical alignment is indicated.
Now that we have chosen a cathode ray tube with all the qualities necessary for the intended film recording application, we must insure its proper use. We must uniformly flood its aperture and focus its divergent beam, correct any astigmatism that exists and, finally, deflect its beam from the center rest position to the 63 million other positions without distorting the beam shape or size so that the picture is as "pretty" around the edges as it is in the middle of the screen.
Easier said than done? It always is. But what follows should make life easier than it might otherwise be. (Refer to Figur 8 for CRT component layout.)
In the beam forming region of the CRT between the cathode surface and the G1 control grid, the electrons are focused by an electrostaic field so that they form a crossover near the plane of the G1 aperture. As the beam emerges from this crossove region it will begin its diverent path, passing through the first accelerator or G2 aperture located less than 50 thousandths of an inch from G1. The beam diameter through both the G1 and G2 apertures is many times smaller than the apertures themselves; thus no beam trimming occurs in this region.
Continuing to diverge at an angle of a few degrees, the beam is accelerated to the final anode potential and toward the final limiting aperture where the outer portion of the beam is extruded through the aperture which is smaller than the beam diameter at this point. This removes the most divergent electrons from the beam, effectively reducing the beam bundle diameter.
If the beam is not aligned to the center of the aperture, less than optimum beam energy transfer to the screen will result and a nonuniform energy distribution in the resulting spot will tend to distort the information in that spot.
To correct for this error, a thin electromagnetic deflection coil is placed ahead of the G1, G2 region. This component, referred to as an aperture flooding coil, produces indepentently controlled x and y transverse fields that, when energized with stable D.C. currents, allow for the positioning of the beam to the center of the limiting aperture.
Proper alignment is accomplished by adjusting the coil currents while observing a low intensity, unfocused, undeflected spot at the center of the CRT. The optimum setting is achieved when the unfocused spot appears most evenly illuminated and the lighted area appears centered within the sharp outline of the aperture. (See Figures 3 and 4.)
Following the beam trimming region a correction to the beam path is made to direct it to the precise mechanical
center of the CRT face. This is usually necessary to correct for small errors in electron gun alignment within the
CRT neck or CRT neck misalignment which will cause the spot to land off center.
To make this correction, a centering coil is placed approximately 1" ahead of the final aperture. Controlled with stable D.C. current, small x and y positional corrections are made to bring the unfocused spot to center. We are now ready to focus the divergent beam.
The electron optical focusing lens, or focus coil, produces an axial magnetic field which changes the direction of the electrons in a rather complex helical trajectory causing them to converge towards the phosphor screen. A single gap thin lens design with both static and dynamic windings is used. The main field produced by the static winding focuses the beam at the center of the CRT with a field strength of several hundred gauss.
The focus coil is positioned at a point along the CRT neck approximately one inch behind the deflection yoke. This location is a practical compromise; if too close, the performance of both deflection and focus will be degraded due to excessive field interaction. If too far apart, the magnification ratio increases producing a larger than necessary spot size.
Precise alignment of the focus lens eliminates the unwanted transverse field components that tend to deflect the beam, ensuring that the electron beam is centered within and parallel to the axis of the field. This will provide the smallest spot size and eliminate spot shift that would occur with changes in focus.
Alignment is accomplished by mounting the focus coil in a CELCO micropositioning device permitting accurate and independent adjustment of horizontal and vertical translation, pitch and yaw. A 50X microscope mounted on a stable X-Y traverse is used to observe the spot location while adjustments are made. The object of the alignment procedure is to obtain a focused spot in both the positive and negative focus polarity that is precisely at the center of the unfocused, undeflected spot.
A high quality focus coil is a prerequisite. It must exhibit low residual magnetism and be of homogeneous metallurgical structure so as to eliminate spot shift and the need to refocus the display during use and from day to day operation. It should have fast dynamic response and good flux settling characteristics. In the event some small spot shift should exist, it will not show up as a relative error in position during a beam dwell as long as the flux settles before the beam is unblanked. The magnetic lens should be free from astigmatism and have low spherical aberration for smallest spot size.
Once the focus coil is properly aligned and the beam focused, we must examine the aberration known as astigmatism.
Astigmatism at CRT center is an aberration caused by asymmetries in the magnetic focus lens and/or in the CRT itself. Focus lens asymmetries may result from errors in mechanical construction, lack of homogeneity in the material that provides the flux path, or non-uniformity of the winding that produces the field.
Astigmatism in the CRT is due in part to lack of circular symmetry in the electrostatic lens elements, errors in alignment of the elements in the gun structure, or uneven static charging of a gun support structure.
This aberration results in a larger spot size and therefore a loss of resolution. In severe cases, up to a 2:1 spot size increase is possible. But even a small amount of astigmatism can be responsible for a display falling short of the design goal. Fortunately, we are able to correct for this aberration by adding an additional magnetic device to the system known as a static astigmatism correction coil. This device produces two independently controlled quadrupole fields arranged at 45 degree angles to one another.
Astigmatism is observed with the aid of a 100X microscope, by slightly reducing the static focus field strength. As the spot begins to defocus it will no longer maintain its round shape if astigmatism is present. Instead, it will become elliptical in shape with the orientation of the axis of the ellipse dependent on the net effect of the causative errors. The magnitude of the distortion is expressed as the largest obtainable ratio of the length to width of the elliptical spot. The maximum ratio is found at a point where the width of the ellipse is at a minimum and just begins to increase in dimension with any further reduction in focus field strength.
Correction is achieved by introducing currents into the two separate phases of the correction coil in the amount and proportion necessary to reshape the elliptical spot into a round one. When this round spot is refocused it will be gratifyingly smaller than it was before correction. (See Figure 5.)
We are now ready to deflect the electron beam. To maintain a focused spot as it is deflected from CRT center towards the edge, it is necessary to reduce the strength of the focus field. This is due to the increase in path length from the focus coil to the flat plane of the CRT face and added focusing action of the deflection yoke fields on the beam. (See Figure 6.)
In order to maintain proper focus on the screen, an approximate 5% reduction in the main field strength is required for a 20 degree deflection. A parabolic wave shape is introduced into the dynamic winding as a function of deflection, bucking the main focus field, reducing its strength so that the spot remains focused at all points on the display. (See Figure 7.) The dynamic winding has a smaller number of turns for low inductance (25-250uH) so that the current through it may be changed rapidly in synchronization with the scanning beam.
In reaching for the ultimate in display performance, the merits of incorporating a high quality deflection
yoke cannot be over emphasized. Display symmetry, off-axis spot size and registration of pixel data are all
affected by deflection yoke characterisitics.
Before these parameters may be properly measured, the deflection yoke must be accurately aligned. A micropositioning assembly similar to that used for the focus coil is required. The micropositioner should have the additional capability for accurate rotational adjustment of the deflection yoke about the axis of the electron beam.
Alignment begins by energizing the horizontal yoke winding with a 60-1000 Hz sine wave or sawtooth sweep covering the full diameter of the CRT. Yoke rotation is adjusted so that, visually, the line appears parallel to the horizontal reference plane. Next, view the line through a 50X microscope which is mounted on a precision x-y traverse and traverse across the diameter; it will most likely appear bowed either up or down. Balance the bow using the rotation adjustment. The observed bow is due to the small vertical deflection components that exist in the horizontal fringe fields and similar components in the main field inside the yoke. These vertical components change sign as they cross the horizontal yoke axis. The object is to align the horizontal scanning beam with the horizontal yoke axis. Translate the yoke vertically to remove this bow. With a few small adjustments the line can be made to coincide with the horizontal reference plane within 0.001 inch.
Deflection is now switched to the vertical yoke axis and the above procedure followed using the horizontal translation adjustment to remove any bow in the vertical line. The rotation control should not be changed, else the horizontal reference established will be lost. Departure of the vertical trace from the true perpendicular reference plane is a measure of the non-orthogonality of the yoke deflection fields. This error should be no greater than 0.1 degree.
The yaw adjustment is made by intentionally introducing 5 to 10 mils of bow into the horizontal scan with the vertical translation. If any imbalance in the bow is observed it can be removed using the yaw adjustment.
The same procedure is applied to the vertical scan using the pitch adjustment to equalize the bow.
Additional fine tuning of the pitch and yaw controls may be accomplished by displaying a full size pincushion square on the CRT face, and adjustments made to achieve the best balance of symmetries around the edges.
When the alignment of one component is changed, interaction among all the components occurs, making it necessary to repeat the alignment procedure starting from the aperture flooding coil, proceeding forward to the yoke. (See Figure 8.) Several iterations of this alignment procedure may be necessary.
The non-linearity between yoke current and the displacement of the electron beam on a flat-face CRT gives rise to
the well known characteristic pincushion distortion. The symmetry of this distortion is one of the criteria examined
in evaluating the performance of the deflection yoke. Other factors will effect this symmetry, and if not properly
accounted for, erroneous conclusions may be derived about yoke performance. Lack of symmetry will result if the
yoke is not properly aligned, if the electron beam is not perpendicular to the CRT face-plate at the center, or
if there is an uneven charge distribution inside the CRT.
A high quality yoke will exhibit a uniform amount of pincushion on each of the sides of the pincushion pattern, with the tilt of any given side not exceeding 0.1% of the line length. This high degree of symmetry will ease the set up of the linearity correction circuitry for an overall display, with an absolute linearity of les than 0.1%. Proper design of deflection field shape and distribution is necessary for low astigmatic aberration of the deflected beam resulting in a maximum of 10% center to edge spot growth.
Under certain conditions of operation, errors in spot position as small as 1/4 mil with respect to previous data recorded, can show up as defects in the image. Elimination these kinds of error requires near perfect deflection settling time characterisitics and near zero residual magnetism. This is a job for the CELCO Deflectron(R).
In cases where the deflection angle and/or the size of the electron beam bundle is too large, the ability for even the finest deflection yoke to limit deflection astigmatism may be exceeded. In these cases it is necessary to add an additional correction coil.
Similar to the static astigmatism corrector, but of low inductance design, this coil is used to reshape the astigmatic electron beam into a round one at all locations on the CRT face. The required correction waveforms are determined by a complete mapping of the CRT face and are applied dynamically as a function of deflection. When all previous steps are properly accomplished, we are nearly there. However an additional consideration follows.
The effects of the Earth's magnetic field on the electron beam inside the CRT can cause various unwanted problems. A CRT with otherwise good beam characterisitics may be disrupted, leading to larger than desired alignment corrections causing degraded linearity and pattern distorion problems.
Once alignment has been accomplished, any movement of the system or movement of magnetic structures near the system will require its complete realignment. The effects of spot jitter and noise produced by electromagnetic interference could render the display useless. Proper magnetic shielding practices utilizing high permeabilitiy nickel alloys will attenuate these unwanted effects.
In conclusion one must recognize that attention to detail is all important when pursuing the ultimate in CRT display performance. The need for precise positioning and alignment of the various components can not be overemphasized. Concern for and recognition of the importance of limiting errors, both long and short term, in the magnetics and the deflection system waveforms to values as small as one part in 50,000 are necesary requirements for the color film recording application.
And finally, when it seems that all has been considered, you can be sure that someone will find some small error in detail that will provide the impetus for further investigation.
Special thanks to Jim Wurtz of Litton Electron Tubes, Tempe, Arizona and Peter Seats of Thomas Electronics, Wayne, New Jersey for their helpful comments. Appreciation is extended to Joe Mays at Electronic Image Systems, Xenia, Ohio our SPIE Technical Session Co-Chairman without whose friendly encouragement this paper would not have been written.
1. Constantine, Sr., J.M., "2-Slit CRT Spot Analyzer Measurement", Published by Society for Information Display (SID), 1966.
1. Septier, A., Focusing of charged Particles, Vol.s I and II, Academic Press, New York and London, 1967.
2. Wang, C.C.T., Computer Calculation of Deflection Aberration in Electron Beams, author witn IBM T.J. Watson Research Center, Yorktown Heights, New York, November 4, 1966.
3. Wendt, G., "On the Image Defects of Magnetic Deflecting fields", dissertation at the Technische Hochschule, Berlin, February 1938.
4. Willder, S.S., "Dynamic Refocusing in Microspot Cathode Ray Tubes", and "The Correction of Deflection Astigmatism in Microspot Cathode Ray Tubes", J. Phys. D: Appl. Phys., Vol. 4, Department of Nuclear Physics, U. of Oxford, MS recd. 29th April 1970, Great Britain, 1971.
5. Zworykin, V.K., and Morton, G.A., Television, 2nd Edition, John Wiley & Sons, Inc., New York, 1954.
C E L C O
Constantine Engineering Laboratories Co.
14 Industrial Avenue
Mahwah, New Jersey 07430
Phone: (201) 327 - 1123
FAX: (201) 327 - 7047
Web Sites: www.celco-nj.com
C E L C O
8660 Red Oak Avenue
Rancho Cucamonga, CA 91730
Phone: (909) 481 - 4648
FAX: (909) 481 - 6899
Web Site: www.celco.com
J A P A N
Shinjuku Nomura Bldg.
1-26-2 Nishishinjuku, Shinjuku-ku
celco-nj.com Web Site
by John H. Dodd / CELCO
HTML pages for this document
prepared by John H. Dodd / CELCO
Graphics ( Figures 1 through 8 )
created by John H. Dodd / CELCO