Question on Northrop's 1960-70s RF LO work

r3mu511

I really should change my personal text
Joined
18 June 2016
Messages
26
Reaction score
6
Hello all, just joined the forum and had a question on Northrop's 1960s/70s RF related work.

I was reading up on diffraction and in the textbook "Fundamentals of the Physical Theory of Diffraction" by Ufimtsev, the foreword was written by Kenneth Mitzner (who worked on the B-2 for Northrop), in which Mitzner said (in the quote "PTD" refers to Ufimtsev's physical theory of diffraction):

At Northrop, where I worked on the B-2 project, we were so enthusiastic about PTD that a co-worker and I sometimes broke into choruses of "Go, Ufimtsev" to the tune of "On,Wisconsin." At both Lockheed and Northrop we referred to PTD as "industrial-strength" diffraction theory...

So it seems that for the B-2, Northrop was using Ufimtsev's methods to numerically approximate RF scattering. What's interesting is that in Rebecca Grant's book "B-2 Spirit of Innovation", she described about how, prior to the B-2 development, Northrop went about designing their XST aircraft (ie. the competition that led up to the F-117) without the use of Ufimtsev's work, nor the benefit of the ECHO-1 software of Denys Overholser (the mathematician at Lockheed).

So I was wondering what was Northrop using at that time (pre-B-2 devlopment) for their numerical approximation of RF scattering in their LO design work? Grant had written that Northrop relied on their extensive experience in empirical testing of RCS of actual USAF aircraft in their design work, but she doesn't mention whether Northrop did or did not have a computational method to approximate RF scattering as an adjunct to their modeling and range testing.

During this timeframe (1960s-70s) there was an alternative method for numerical RF scattering approximation based on the work of Joseph Keller on the geometric theory of diffraction (GTD) which he developed in the early 1950s, but I haven't been able to find any traces of a possible link between Keller's work and Northrop's pre-B-2 LO design efforts.

So does anyone have any idea if Northrop did or did not use any computational method for RF scattering in their pre-B-2 work?
 
In the early 1970s Northrop had GENSCAT, roughly equivalent to ECHO-1 if not a little better, and were able to calculate the RCS of an F-4C Phantom II under government contract with reasonable accuracy. However, it didn't help you design a stealth aircraft, simply to calculate the RCS of a shape - the shape itself had to come from the mind of the designer.

Interestingly, the N-327 RCS test model was just as faceted as Lockheed's initial "Hopeless Diamond", and I believe the major difference between Lockheed and Northrop's XST proposals was that the Northrop aerodynamics guys insisted on sacrificing a bit of stealth to make a flyable airplane while Lockheed rigidly stuck to their faceted design principle.
 
^Thanks very much, just what I needed :D

I looked up GENSCAT and what I've found so far states it also used Ufimtsev's PTD for the edge diffraction contributions to scatter.

Since the version of PTD at that time did not have a numerical approximation method for the field contribution due to surface diffractions (ie. contributions of creeping waves) of curved surfaces (as opposed to edge diffractions of flat surfaces which was already in that version of PTD), then if GENSCAT was using that version of PTD (like ECHO-1) this could possibly explain why Northrop's aircraft also exhibited a flat-faced, faceted look.
 
This is interesting. Mr Mitzer's Incremental length diffraction coefficients report extending Ufimtsev's PTD work from 1973 - this was incorporated into GENSCAT by 1974.

http://www.dtic.mil/cgi-bin/GetTRDoc?Location=U2&doc=GetTRDoc.pdf&AD=AD0918861

This document reports a portion of the work performed by Northrop Corporation, Aircraft Division, Hawthorne, California, under USAF Contract F33615-70-C-1820, "Calculation of Radar Cross Section", during the period from 1 December 1970 to 1 June 1973. The work was sponsored by the Electronic Warfare Division, Air Force Avionics Laboratory under Project 7633, Task 13, with Dr. Charles H. Krueger, AFAL/WRP as technical monitor.

This document was prepared by Dr. K. M. Mitzner of the Electronic Systems Research and Technology Group at Northrop. Mr. S. Stanley Locus was Principal Investigator on the contract.

This document has been assigned NOR 73-104 by Northrop for internal control purposes and was submitted by the author in July 1973.

This Technical Report has been reviewed and is approved for publication.
 
http://oai.dtic.mil/oai/oai?verb=getRecord&metadataPrefix=html&identifier=AD0913626

Not available for download but...

Accession Number : AD0913626

Title : Calculation of Radar Cross Section Aircraft Geometry Methods.

Descriptive Note : Technical rept. Jul 70-Sep 73,

Corporate Author : NORTHROP CORP HAWTHORNE CA AIRCRAFT DIV

Personal Author(s) : Heath, Hugh C.

Report Date : SEP 1973

Pagination or Media Count : 114

Abstract : Geometry description methods have been developed to transform design data of an aircraft configuration into a computer model. The computer model is used as a basis for radar cross section calculations covering the frequency range of 500 to 20,000 MHz. Accomplishments include: development of a quadric surface fitting program, complete analysis of quadric surfaces including intersections with lines, location of scattering centers, calculation of Gaussian curvature and surface zoning. (Author)

Descriptors : *RADAR CROSS SECTIONS, *AIRCRAFT, GEOMETRY, AIRFRAMES, CURVE FITTING, TRANSFORMATIONS(MATHEMATICS), COMPUTER PROGRAMS, SCATTERING, SURFACES.

Subject Categories : Aircraft
Computer Programming and Software
Active & Passive Radar Detection & Equipment

Distribution Statement : APPROVED FOR PUBLIC RELEASE
 
Yup, Mitzner's ILDC was an expansion of PTD. Ufimtsev described ILDC as another method to interpret the directivity of EM edge waves (he discusses more on this in his textbook "Fundamentals of PTD"). ILDC was also expanded later by Roberto Tiberio (ref: http://www.dii.unimo.it/wiki/images/5/59/R3.pdf), but I haven't spent much time studying the expansions of PTD as I spent more time studying the later expansions of Keller's GTD leading up to the uniform theory of diffraction (UTD). I just find UTD more "intuitive" than the later forms of PTD, it just seems more "elegant", but hey to each his own preference.

---

Regarding Gaussian curvature, this is a measure of the curvature of a surface and is not identical to a computational method for surface diffraction rays scattered from a curved surface (ie. creeping waves) with varying incidence of irradiation. We'ld have to see the content of that report you linked (the one which isn't available for download in the link) to see if it contains a numerical approximation method to compute surface diffractions for a curved surface irradiated by a plane wave with arbitrary incidence. If it did then at that time they would have had a method to numerically approximate scattering for a curved surface and we might have seen a less flat-faced, faceted result. But if they did not have that computational approximation then they would have to rely on what they did have (ie. methods for flat-face wedge scattering). Thanks for the link, I now have another historical paper to try and track down :)

You'll notice both in the version of PTD and in ILDC of that time period, they do not contain numerical methods for scattering from a curved surface irradiated by a plane wave of arbitrary incidence. At best, the original version of PTD contains computations for rotated paraboloids with plane wave incidence along the axis of symmetry. While Mitzner's 1974 ILDC paper similarly lacks a method for curved surfaces with arbitrary irradiation incidence.

So at least historically, the papers of that time appear to show they could numerically approximate scattering (with varying irradiation incidence) for flat faces but not for curved faces.

(But that's just for that time period, modern expansions like UTD - as well as modern PTD, and other expansions like MEC/equivalent currents - cover curves as well as flat surfaces, with arbitrary incidence of radiation).
 
Northrop's N327 LO pole model.
 

Attachments

  • Northrop N327.jpg
    Northrop N327.jpg
    70.2 KB · Views: 289
Oshiro, F. K. and K. M. Mitzner (1967) "Digital computer solution of three- dimensional scattering problems"
Mitzner, K. M. (1967), An integral equation approach to scattering from a body of finite conductivity, Radio Science 2, 1459-1470.
 
Thanks for the references, much appreciated. :) I do not have access to Oshiro's original 1965 SDT (source distribution technique) paper but I do have textbooks which discuss his SDT method.

In both "Electromagnetic Modeling of Composite Metallic and Dielectric Structures" (Kolundzija, Djordjevic) and "Analysis of Metallic Antennas and Scatterers" (Popovic, Kolundzija) books, the SDT is treated as an early form of the MOM (method of moments) technique, and like the early forms of numerical solutions for scattering did not yet include edge diffraction contributions.

So if Mitzner was using this early form of Oshiro's SDT at that time, this might be why Mitzner had to wait till he got Ufimtsev's PDT method in order to finally be able to include edge diffraction in the GENSCAT program at Northrop.
 
Rereading the various sources the key point appears to be that in 1975 Northrop knew Ufimtsev and had included his work into GENSCAT prior to XST, and therefore had as good RCS modelling software for the XST as Lockheed did for faceted shapes, but they did not stick to facets with their design.

Northrop instead started moving towards incorporating curves ("2nd generation stealth") even though that meant they couldn't (yet) calculate the RCS but had to design by intuition and an understanding of radar basic principles and later measure it in model form. They formulated the design principles later used on the B-2 and YF-23 with Tacit Blue, while still lacking the ability to predict RCS from curves.

Lockheed in contrast stuck rigidly to shapes that could be calculated, and aerodynamics and stability be damned, which helped them win XST.
They even got a bit fixated on facets - "if we can't predict the RCS we can't use curves" - which cost them on the ATB.
 
Regarding Northrop's work post-XST on the B-2, it's curious that they did not utilize a method to numerically approximate scattering for curved surfaces by that time, as the work of people like Kouyoumjian on applying UTD for surface diffracted rays from convex surfaces was already published by 1980.

Perhaps the timing of the publishing of these later expansions of GTD came too late in the dev't cycle of Northrop's work on the B-2.
 

Similar threads

Back
Top Bottom