Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “IMPINGEMENT”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 577 records · Page 32
All sourcesNASA NTRSDOE OSTI
NASA NTRS1990

Shock interference heating in scramjet engines

Experimental and analytical research sponsored by the NASA Langley Research center and the NASP Structures Technology Maturation Program to define critical aerothermal loads for the NASP engine is summarized. Presented is a review of (1) shock-shock interaction on the engine cowl leading edge that results in a supersonic jet impinging on the leading edge surface and causes the heat transfer rate to be amplified by a factor of 30 or more over the undisturbed (no shock interaction) flow stagnation point heat transfer rate, (2) the effectiveness of supersonic film cooling with and without the effects of an impinging oblique shock wave, and (3) oblique shock impingement in an axial compression corner.

Wieting, Allan R.↗
NASA NTRS1991

Icing characteristics of a natural-laminar-flow, a medium-speed, and a swept, medium-speed airfoil

Tests were conducted in the Icing Research Tunnel at the NASA Lewis Research Center to determine the icing characteristics of three modern airfoils: a natural-laminar-flow, a medium-speed, and a swept medium-speed airfoil. The tests measured the impingement characteristics and drag degradation for angles-of-attack typifying cruise and climb for cloud conditions typifying the range that might be encountered in flight. The maximum degradation occurred at the cruise angle-of-attack for the long, glaze ice condition for all three airfoils with increases over baseline drag being 486 percent, 510 percent, and 465 percent for the natural-laminar-flow, the medium-speed, and the swept, medium-speed airfoils, respectively. For the climb angle-of-attack, the maximum drag degradation (and total extent of impingement) observed were also for the long, glaze ice condition and were 261 percent, 181 percent, and 331 percent, respectively. The minimum drag degradation (and extent of impingement) occurred for the cruise condition and for the short, rime spray with increases over baseline drag values being 47 percent, 28 percent, 46 percent, respectively.

Bidwell, Colin S.↗
NASA NTRS1992

Comparison of turbulence models for powered-lift flow fields

The development of a practical turbulence model for the complex flows with strong streamline curvature associated with powered-lift aircraft was investigated. The approach was to modify existing models that were compatible with the widely used kappa-epsilon model already available in an existing 3D Navier-Stokes code. The state-of-the-art of current turbulence models was examined. Several models that provide curvature corrections to the kappa-epsilon model were numerically investigated. An algebraic Reynolds stress model, which provides additional physics, was also examined. Test cases including an impinging jet, an impinging jet in a crossflow, and two adjacent impinging jets with the resultant upwash fountain were calculated. The first two test cases could be modeled reasonsbly well with some of the models. The upwash fountain could not be modeled with sufficient accuracy, which is consistent with the findings of other investigators.

Roberts, Donald W.↗
NASA NTRS1993

An experimental study of the sources of fluctuating pressure loads beneath swept shock/boundary-layer interactions

An experimental research program providing basic knowledge and establishing a database on the fluctuating pressure loads produced on aerodynamic surfaces beneath three dimensional shock wave/boundary layer interactions is described. Such loads constitute a fundamental problem of critical concern to future supersonic and hypersonic flight vehicles. A turbulent boundary layer on a flat plate is subjected to interactions with swept planar shock waves generated by sharp fins at angle of attack. Fin angles from 10 to 20 deg at freestream Mach numbers of 3 and 4 produce a variety of interaction strengths from weak to very strong. Miniature Kulite pressure transducers flush-mounted in the flat plate are used to measure interaction-induced wall pressure fluctuations. The distributions of properties of the pressure fluctuations, such as their ring levels, amplitude distributions, and power spectra, are also determined. Measurements were made for the first time in the aft regions of these interactions, revealing fluctuating pressure levels as high as 160 dB. These fluctuations are dominated by low frequency (0-5 kHz) signals. The maximum ring levels in the interactions show an increasing trend with increasing interaction strength. On the other hand, the maximum ring levels in the forward portion of the interactions decrease linearly with increasing interaction sweep back. These ring pressure distributions and spectra are correlated with the features of the interaction flowfield. The unsteadiness of the off-surface flowfield is studied using a new, non-intrusive technique based on the shadow graph method. The results indicate that the entire lambda-shock structure generated by the interaction undergoes relatively low-frequency oscillations. Some regions where particularly strong fluctuations are generated were identified. Fluctuating pressure measurements are also made along the line of symmetry of an axisymmetric jet impinging upon a flat plate at an angle. This flow was chosen as a simple analog to the impinging jet region found in the rear portion of the shock wave/boundary layer interactions under study. It is found that a sharp peak in ring pressure level exists at or near the mean stagnation point. It is suggested that the phenomena responsible for this peak may be active in the swept interactions as well, and may cause the extremely high fluctuating pressures observed in the impinging jet region in the present experimental program.

Settles, G. S.↗
NASA NTRS1994

A six degree of freedom, plume-fuel optimal trajectory planner for spacecraft proximity operations using an A* node search

Spacecraft proximity operations are complicated by the fact that exhaust plume impingement from the reaction control jets of space vehicles can cause structural damage, contamination of sensitive arrays and instruments, or attitude misalignment during docking. The occurrence and effect of jet plume impingement can be reduced by planning approach trajectories with plume effects considered. An A* node search is used to find plume-fuel optimal trajectories through a discretized six dimensional attitude-translation space. A plume cost function which approximates jet plume isopressure envelopes is presented. The function is then applied to find relative costs for predictable 'trajectory altering' firings and unpredictable 'deadbanding' firings. Trajectory altering firings are calculated by running the spacecraft jet selection algorithm and summing the cost contribution from each jet fired. A 'deadbanding effects' function is defined and integrated to determine the potential for deadbanding impingement along candidate trajectories. Plume costs are weighed against fuel costs in finding the optimal solution. A* convergence speed is improved by solving approach trajectory problems in reverse time. Results are obtained on a high fidelity space shuttle/space station simulation. Trajectory following is accomplished by a six degree of freedom autopilot. Trajectories planned with, and without, plume costs are compared in terms of force applied to the target structure.

Jackson, Mark Charles↗
NASA NTRS1994

The Effects of Vortex Ingestion on the Flow in a Diffusing S-duct

An experimental study of the effects of an ingested vortex on the flowfield of a diffusing S-duct is reported. The vortex is generated through the use of a stationary pinwheel device mounted upstream of the diffusing S-duct. Three test conditions vary the location of the vortex in the duct inlet crossplane. For each condition of ingested vortex, a baseline S-duct and an S-duct with an array of vortex generators is examined. The data taken consist of duct inlet and exit crossplane surveys of velocity and total pressure. Duct surface flow visualization and static pressure are also recorded. The data acquired in these tests are compared to identical S-duct data taken in the absence of the ingested vortex. The ingested vortex is observed to have a strong influence on the flowfield inside (and exiting) the S-duct, but only when the vortex impinges at the inlet crossplane location coincident with the crossplane location of downstream flow separation within the duct. When the ingested vortex impinges at this location it reduces the extent of flowfield separation inside the baseline duct and promotes stronger crossflow in the exit plane of both the baseline duct and the duct with installed vortex generators. This enhanced crossflow also strengthens the vortices shed from the vortex generators. The other impingement locations of the ingested vortex are found to produce little effect on the flowfield of the duct with or without vortex generators.

Wendt, B. J.↗
NASA NTRS1994

Experimental studies of shock-wave/wall-jet interaction in hypersonic flow

Experimental studies have been conducted to examine slot film cooling effectiveness and the interaction between the cooling film and an incident planar shock wave in turbulent hypersonic flow. The experimental studies were conducted in the 48-inch shock tunnel at Calspan at a freestream Mach number of close to 6.4 and at a Reynolds number of 35 x 10(exp 6) based on the length of the model at the injection point. The Mach 2.3 planar wall jet was generated from 40 transverse nozzles (with heights of both 0.080 inch and 0.120 inch), producing a film that extended the full width of the model. The nozzles were operated at pressures and velocities close to matching the freestream, as well as at conditions where the nozzle flows were over- and under-expanded. A two-dimensional shock generator was used to generate oblique shocks that deflected the flow through total turnings of 11, 16, and 21 degrees; the flows impinged downstream of the nozzle exits. Detailed measurements of heat transfer and pressure were made both ahead and downstream of the injection station, with the greatest concentration of measurements in the regions of shock-wave/boundary layer interaction. The major objectives of these experimental studies were to explore the effectiveness of film cooling in the presence of regions of shock-wave/boundary layer interaction and, more specifically, to determine how boundary layer separation and the large recompression heating rates were modified by film cooling. Detailed distributions of heat transfer and pressure were obtained in the incident shock/wall-jet interaction region for a series of shock strengths and impingement positions for each of the two nozzle heights. Measurements were also made to examine the effects of nozzle lip thickness on cooling effectiveness. The major conclusion from these studies was that the effect of the cooling film could be readily dispersed by relatively weak incident shocks, so the peak heating in the recompression region was not significantly reduced by even the largest levels of film cooling. For the case studies in the absence of film cooling, the interaction regions were unseparated. However, adding film cooling resulted in regions of boundary layer separation induced in the film cooling layer -- the size of which regions first increased and then decreased with increased film cooling. Surprisingly, the size of the separated regions and the magnitude of the recompression heating were not strongly influenced by the thickness of the cooling film, nor by the point of shock impingement relative to the exit plane of the nozzles. The lip thickness was found to have little effect on cooling effectiveness. Measurements with and in the absence of shock interaction were compared with the results of earlier experimental studies and correlated in terms of the major parameters controlling these flows.

Holden, Michael S.↗
NASA NTRS1994

Experimental studies of shock-wave/wall-jet interaction in hypersonic flow, part A

Experimental studies have been conducted to examine slot film cooling effectiveness and the interaction between the cooling film and an incident planar shock wave in turbulent hypersonic flow. The experimental studies were conducted in the 48-inch shock tunnel at Calspan at a freestream Mach number of close to 6.4 and at a Reynolds number of 35 x 10(exp 6) based on the length of the model at the injection point. The Mach 2.3 planar wall jet was generated from 40 transverse nozzles (with heights of both 0.080 inch and 0.120 inch), producing a film that extended the full width of the model. The nozzles were operated at pressures and velocities close to matching the freestream, as well as at conditions where the nozzle flows were over- and under-expanded. A two-dimensional shock generator was used to generate oblique shocks that deflected the flow through total turnings of 11, 16, and 21 degrees; the flows impinged downstream of the nozzle exits. Detailed measurements of heat transfer and pressure were made both ahead and downstream of the injection station, with the greatest concentration of measurements in the regions of shock-wave/boundary layer interaction. The major objectives of these experimental studies were to explore the effectiveness of film cooling in the presence of regions of shock-wave/boundary layer interaction and, more specifically, to determine how boundary layer separation and the large recompression heating rates were modified by film cooling. Detailed distributions of heat transfer and pressure were obtained in the incident-shock/wall-jet interaction region for a series of shock strengths and impingement positions for each of the two nozzle heights. Measurements were also made to examine the effects of nozzle lip thickness on cooling effectiveness. The major conclusion from these studies was that the effect of the cooling film could be readily dispersed by relatively weak incident shocks, so the peak heating in the recompression region was not significantly reduced by even the largest levels of film cooling. For the case studies in the absence of film cooling, the interaction regions were unseparated. However, adding film cooling resulted in regions of boundary layer separation induced in the film cooling layer, the size of which regions first increased and then decreased with increased film cooling. Surprisingly, the size of the separated regions and the magnitude of the recompression heating were not strongly influenced by the thickness of the cooling film, nor by the point of shock impingement relative to the exit plane of the nozzles. The lip thickness was found to have little effect on cooling effectiveness. Measurements with and in the absence of shock interaction were compared with the results of earlier experimental studies and correlated in terms of the major parameters controlling these flows.

Holden, Michael S.↗
NASA NTRS1994

Pressure-sensing performance of upright cylinders in a Mach 10 boundary-layer

An experimental research program to provide basic knowledge of the pressure-sensing performance of upright, flushported cylinders in a hypersonic boundary layer is described. Three upright cylinders of 0.25-, 0.5- and l.0-in. diameters and a conventional rake were placed in the test section sidewall boundary layer of the 31 Inch Mach 10 Wind Tunnel at NASA Langley Research Center, Hampton, Virginia. Boundary-layer pressures from these cylinders were compared to those measured with a conventional rake. A boundary-layer thickness-to-cylinder-diameter ratio of 8 proved sufficient to accurately measure an overall pressure profile and ascertain the boundary-layer thickness. Effects of Reynolds number, flow angularity, and shock wave impingement on pressure measurement were also investigated. Although Reynolds number effects were negligible at the conditions studied, flow angularity above 10 deg significantly affects the measured pressures. Shock wave impingement was used to investigate orifice-to-orifice pressure crosstalk. No crosstalk was measured. The lower pressure measured above the oblique shock wave impingement showed no influence of the higher pressure generated at the lower port locations.

Johnson, Steven↗
NASA NTRS1995

Load measurement system with load cell lock-out mechanism

In the frame work of the project Shuttle Plume Impingement Flight Experiment (SPIFEX), a Load Measurement System was developed and fabricated to measure the impingement force of Shuttle Reaction Control System (RCS) jets. The Load Measurement System is a force sensing system that measures any combination of normal and shear forces up to 40 N (9 lbf) in the normal direction and 22 N (5 lbf) in the shear direction with an accuracy of +/- 0.04 N (+/- 0.01 lbf) Since high resolution is required for the force measurement, the Load Measurement System is built with highly sensitive load cells. To protect these fragile load cells in the non-operational mode from being damaged due to flight loads such as launch and landing loads of the Shuttle vehicle, a motor driven device known as the Load Cell Lock-Out Mechanism was built. This Lock-Out Mechanism isolates the load cells from flight loads and re-engages the load cells for the force measurement experiment once in space. With this highly effective protection system, the SPIFEX load measurement experiment was successfully conducted on STS-44 in September 1994 with all load cells operating properly and reading impingement forces as expected.

Le, Thang↗
NASA NTRS1997

Fundamentals of Mold Free Casting: Experimental and Computational Studies

Researchers are developing the technology of 'Ballistic Particle Manufacturing' (BPM) in which individual drops are precisely layered onto a substrate, and the drops are deposited so as to prevent splatting. These individual drops will ultimately be combined to form a net-shape, three-dimensional object. Our understanding of controlled drop deposition as applied to BPM is far from complete. Process parameters include the size and temperature of the liquid metal drop, its impact velocity and trajectory, and the condition and temperature of the substrate. Quantitative knowledge of the fluid mechanics and heat transfer of drop deposition and solidification are necessary to fully optimize the manufacturing process and to control the material microstructure of the final part. The object of this study is to examine the dynamics of liquid metal drops as they impinge upon a solid surface and solidify under conditions consistent with BPM (i.e. conditions which produce non-splatting drops). A program of both numerical simulations and experiments will be conducted. Questions this study will address include the following: How do the deformation and solidification of the drop depend on the properties of the fluid drop and the solid substrate? How does the presence of previously deposited drops affect the impingement and solidification process? How does the impingement of the new drop affect already deposited material? How does the cooling rate and solidification of the drops influence the material microstructure?

Tryggvason, Gretar↗
NASA NTRS1995

Implementation of Combined Feather and Surface-Normal Ice Growth Models in LEWICE/X

Experimental observations have shown that discrete rime ice growths called feathers, which grow in approximately the direction of water droplet impingement, play an important role in the growth of ice on accreting surfaces for some thermodynamic conditions. An improved physical model of ice accretion has been implemented in the LEWICE 2D panel-based ice accretion code maintained by the NASA Lewis Research Center. The LEWICE/X model of ice accretion explicitly simulates regions of feather growth within the framework of the LEWICE model. Water droplets impinging on an accreting surface are withheld from the normal LEWICE mass/energy balance and handled in a separate routine; ice growth resulting from these droplets is performed with enhanced convective heat transfer approximately along droplet impingement directions. An independent underlying ice shape is grown along surface normals using the unmodified LEWICE method. The resulting dual-surface ice shape models roughness-induced feather growth observed in icing wind tunnel tests. Experiments indicate that the exact direction of feather growth is dependent on external conditions. Data is presented to support a linear variation of growth direction with temperature and cloud water content. Test runs of LEWICE/X indicate that the sizes of surface regions containing feathers are influenced by initial roughness element height. This suggests that a previous argument that feather region size is determined by boundary layer transition may be incorrect. Simulation results for two typical test cases give improved shape agreement over unmodified LEWICE.

Matthew T Velazquez↗
NASA NTRS1999

How an Electron Beam (Eventually) Penetrates Ceramic Cloth

In anticipation of the International Space Welding Experiment (ISWE) the effect of an electron beam was investigated on Nextel AF-62 ceramic cloth designed to withstand temperatures up to 1427 C. It was anticipated that the cloth would take a static charge that would repel the beam and remain undamaged. It was found that after some seconds the impinging beam penetrated penetrated the cloth. Further, the penetration time went up significantly both at longer and at closer standoff distances. A tentative explanation appears to fit the observed facts. The electrons in the beam generate positive ions by collisions with the contaminant gas molecules in the vacuum chamber. The positive ions transfer a small but significant fraction of the beam power to the cloth. Under the impingement of the positive ions the cloth heats up until sufficient outgassing occurs to initiate arcing. Once arcing occurs the full beam power impinges on the cloth and, almost instantaneously, burns a hole.

Nunes, A. C., Jr.↗
NASA NTRS1966

Storable Propellant Combustion Instability Program at Lewis Research Center

An experimental program at Lewis Research Center investigates acoustical-mode combustion instability in liquid propellant rockets. One phase of this program is concerned with nitrogen tetroxide and a 50-50 fuel blend of hydrazine-UDMH. Effects on combustor stability and performance by variations in injection velocities, impingement angle and distance, amd thrust per element of triplet injectors were studied in a 10.77-inch- diameter cylindrical combustor at chamber pressures of 100 and 300 psia. Nominal thrust levels were 6700 and 20 000 pounds. Stability rating was accomplished by the use of various size RDX explosive charges to generate tangential pressure disturbances. The injection velocity effect correlated by using a ratio V(sub o)/V(sub f). Stability correlation with V(sub o)/V(sub f) resulted in a sharply humped curve with maximum stability at a ratio of 1.2. Performance decreased with increase in V(sub o)/V(sub f) after a maximum at 0.7, which agreed fairly well with the "uniform mixture ratio distribution" criteria of the Jet Propulsion Laboratory. Variation of impingement angle from 38 deg to 120 deg indicated that performance correlated with the absolute velocity ratio V(sub o)/V(sub f), but stability correlation appeared to be best with a velocity ratio using the radial component of the fuel injection velocity V(sub o)/V(sub fr). Impingement distances from 0.5 to 1.0 inch appeared to have only slight effect on stability and performance. Investigation of injector thrust per element for V(sub o)/V(sub fr) ranging from 0.8 to 2.8 yielded convex maxima curves. Acoustic liners were also tested to improve a liner design computer program, to obtain design criteria, and to develop flight-type liners. A marginally stable and spontaneously unstable combustor were used to evaluate linear configurations. All liners were effective in the marginally stable configuration, and only liners with large theoretical absorptivity were successful in the spontaneously unstable combustor.

Tabata, William K.↗
NASA NTRS2000

An Optimization-Based Approach to Injector Element Design

An injector optimization methodology, method i, is used to investigate optimal design points for gaseous oxygen/gaseous hydrogen (GO2/GH2) injector elements. A swirl coaxial element and an unlike impinging element (a fuel-oxidizer-fuel triplet) are used to facilitate the study. The elements are optimized in terms of design variables such as fuel pressure drop, APf, oxidizer pressure drop, deltaP(sub f), combustor length, L(sub comb), and full cone swirl angle, theta, (for the swirl element) or impingement half-angle, alpha, (for the impinging element) at a given mixture ratio and chamber pressure. Dependent variables such as energy release efficiency, ERE, wall heat flux, Q(sub w), injector heat flux, Q(sub inj), relative combustor weight, W(sub rel), and relative injector cost, C(sub rel), are calculated and then correlated with the design variables. An empirical design methodology is used to generate these responses for both element types. Method i is then used to generate response surfaces for each dependent variable for both types of elements. Desirability functions based on dependent variable constraints are created and used to facilitate development of composite response surfaces representing the five dependent variables in terms of the input variables. Three examples illustrating the utility and flexibility of method i are discussed in detail for each element type. First, joint response surfaces are constructed by sequentially adding dependent variables. Optimum designs are identified after addition of each variable and the effect each variable has on the element design is illustrated. This stepwise demonstration also highlights the importance of including variables such as weight and cost early in the design process. Secondly, using the composite response surface that includes all five dependent variables, unequal weights are assigned to emphasize certain variables relative to others. Here, method i is used to enable objective trade studies on design issues such as component life and thrust to weight ratio. Finally, combining results from both elements to simulate a trade study, thrust-to-weight trends are illustrated and examined in detail.

Tucker, P. Kevin↗
NASA NTRS2000

High-order Two-way Artificial Boundary Conditions for Nonlinear Wave Propagation with Backscattering

When solving linear scattering problems, one typically first solves for the impinging wave in the absence of obstacles. Then, by linear superposition, the original problem is reduced to one that involves only the scattered waves driven by the values of the impinging field at the surface of the obstacles. In addition, when the original domain is unbounded, special artificial boundary conditions (ABCs) that would guarantee the reflectionless propagation of waves have to be set at the outer boundary of the finite computational domain. The situation becomes conceptually different when the propagation equation is nonlinear. In this case the impinging and scattered waves can no longer be separated, and the problem has to be solved in its entirety. In particular, the boundary on which the incoming field values are prescribed, should transmit the given incoming waves in one direction and simultaneously be transparent to all the outgoing waves that travel in the opposite direction. We call this type of boundary conditions two-way ABCs. In the paper, we construct the two-way ABCs for the nonlinear Helmholtz equation that models the laser beam propagation in a medium with nonlinear index of refraction. In this case, the forward propagation is accompanied by backscattering, i.e., generation of waves in the direction opposite to that of the incoming signal. Our two-way ABCs generate no reflection of the backscattered waves and at the same time impose the correct values of the incoming wave. The ABCs are obtained for a fourth-order accurate discretization to the Helmholtz operator; the fourth-order grid convergence is corroborated experimentally by solving linear model problems. We also present solutions in the nonlinear case using the two-way ABC which, unlike the traditional Dirichlet boundary condition, allows for direct calculation of the magnitude of backscattering.

Fibich, Gadi↗
NASA NTRS2002

Wake Closure Characteristics and Afterbody Heating on a Mars Sample Return Orbiter

Aeroheating wind-tunnel tests were conducted on a 0.028 scale model of an orbiter concept considered for a possible Mars sample return mission. The primary experimental objectives were to characterize hypersonic near wake closure and determine if shear layer impingement would occur on the proposed orbiter afterbody at incidence angles necessary for a Martian aerocapture maneuver. Global heat transfer mappings, surface streamline patterns, and shock shapes were obtained in the NASA Langley 20-Inch Mach 6 Air and CF4 Tunnels for post-normal shock Reynolds numbers (based on forebody diameter) ranging from 1,400 to 415,000, angles of attack ranging from -5 to 10 degrees at 0, 3, and 6 degree sideslip, and normal-shock density ratios of 5 and 12. Laminar, transitional, and turbulent shear layer impingement on the cylindrical afterbody was inferred from the measurements and resulted in a localized heating maximum that ranged from 40 to 75 percent of the reference forebody stagnation point heating. Comparison of laminar heating prediction to experimental measurement along the orbiter afterbody highlight grid alignment challenges associated with numerical simulation of three- dimensional separated wake flows. Predicted values of a continuum breakdown parameter revealed significant regions of non-continuum flow downstream of the flow separation at the MSRO shoulder and in the region of the reattachment shock on the afterbody. The presence of these regions suggest that the Navier-Stokes predictions at the laminar wind-tunnel condition may encounter errors in the numerical calculation of the wake shear layer development and impingement due to non-continuum effects.

Horvath, Thomas J.↗
NASA NTRS2004

Progress in NEXT Ion Optics Modeling

Results are presented from an ion optics simulation code applied to the NEXT ion thruster geometry. The error in the potential field solver of the code is characterized, and methods and requirements for reducing this error are given. Results from a study on electron backstreaming using the improved field solver are given and shown to compare much better to experimental results than previous studies. Results are also presented on a study of the beamlet behavior in the outer radial apertures of the NEXT thruster. The low beamlet currents in this region allow over-focusing of the beam, causing direct impingement of ions on the accelerator grid aperture wall. Different possibilities for reducing this direct impingement are analyzed, with the conclusion that, of the methods studied, decreasing the screen grid aperture diameter eliminates direct impingement most effectively.

Emhoff, Jerold W.↗