APQ-81: radar for the Missileer

[overscan (PaulMM)]

Some details of the APQ-81 radar from the Missileer

Targets tracked simultaneously: 16
Frequency: C band
Average power: 2 kW (to be increased to 5kW later)
PRFs: 65kc/s, 11/10ths of 65kc/s, 11/9ths of 65kc/s
Aerial diameter: 60 in
Beamwidth: 2.5 deg
Range gates: 10
Scan limits: +-60deg in azimuth, +48 to -40deg elevation.
Frame time: 2 sec
0.3sec (2 x 0.15 sec) turnaround time, of which each half is used to transmit guidance commands to missile.
Aerial search: 30 deg x 5deg
Scan pattern: 2 bar scan
time on target: 60-70msec
Number of filters: 10 x 2 x 30
Maximum unambiguous range: 112nm
Target range rate: 215 - 2900 knots
Target altitude: 0-100,000ft
Range, 90% probability on 5 sq m target: 72nm

IBM 704 computer
Reflex klystron oscillator with 4 cavity klystron amplifier.

An early method of range while search was attempted. The radar switched between 3 prfs and then attempted to detect which range gate the signal is received in (output from doppler bank). However signals could overlap two range gates at certain distances. THey considered either measuring the number of times threshold was exceeded in each range gate, then assuming that the range gate where the threshold was exceeded most often was the correct one, or alternatively choosing one and then comparing the range information for consistency with previous range information.

Westinghouse weren't completely happy with the method of range measurement and were considering a linear frequency modulation to the transmitter instead [FM ranging].

 

[Tuomasn]

Re: APG-81: radar for the Missileer

I think you mean APQ-81; AN/APG-81 should be the radar for the F-35 Lightning II (which I don't think uses klystrons ).

 

[Archibald]

... which raise an interesting question : why some radars are APG- when others are APQ- ? ???

 

[Tuomasn]

It seems that at least since the early 1970s fighter radars have been designated AN/APG, while AN/APQ has been used for various other radar types (attack aircraft and bomber radars, terrain following radars etc.) I think the age of the AN-system shows here, as specialized "Fire Control Radars" (what APG means) are hardly made anymore and most radars are at least somewhat "multipurpose".

 

[TinWing]

Is there ever been a reference to French interest in the Missileer's radar? The 1960 MN TALP requirement was very similar to Missileer, albeit in a lighter platform suitable for the Clemenceau class - although I have never found any references to the radar or missiles meant for the subsonic TALP?

 

[overscan (PaulMM)]

Developing New Radar Techniques
Fegely:

It’s 1955 by now and I became supervisor in the radar section.

Hochheiser:

So by this time had you solved the—?

Fegely:

Antenna problems? Ok for today’s systems but not for the future.

Hochheiser:

Okay. So now we’re off from BOMARC onto - - .

Fegely:

Developing techniques that will be needed for the next generation Navy, Air Force or Army radar.

Hochheiser:

Ah.

Fegely:

You’re doing good if one out of five ideas will be key to winning the next radar.

Hochheiser:

[Interposing] So it’s not that closely tied to any current product?

Fegely:

No. We did assist solving some production problems. But our task was to bring in the next one. Design Engineering Dept. handled production problems.

Hochheiser:

Okay. And about when is this?

Fegely:

It took about two years to convince the Navy we were the ones to develop their long range track-while-scan C-band pulse Doppler radar, the APQ-81 which started in 1961.

Hochheiser:

Okay.

Fegely:

The key techniques for multiple target track-while-scan were changing the prf several times during the antennae beam dwell time, a very low side lobe antenna, digital processing to establish track files, and multiple range gated channels to reduce clutter. [Data from Pete D'Anna: After having been used for the BOMARC seekers, the Pulse Doppler technology in] theAPQ-81’s [pioneering pulse doppler, multiple target, track while scan system] was the [principal] base[line] for our future fire control systems and the driver for the AWACS radar. [The APQ-81was originally used as the radar fire control system for the Bendix long range air to air Eagle Missile System. Through all of these early programs it was] Johnnie Pearson’s persistence [that] got the hardware and test done on time [as was his later skillful management of the AWACS fly-off at Boeing which won this key major program for Westinghouse. Initially a young engineer by the name of Bill Jones explored the application of the AN/APQ-81’s technology to resolving the basic AWACS detection and tracking problem. When] Bill Skillman [became available to the program, his diligent pursuit of the many ramifications for fully developing the full potential of Pulse Doppler techniques earned him the well deserved title as the “Architect” of the AWACS radar system. In his following work on the AWACS program, he was the] technical Guru [assigned to] “Keep AWACS Sold” to the Washington crowd. [It t]ook years to get him back on new programs. Can’t forget Charlie Calhoun’s Chinese Remainder Theorem or Gwen Hays' tenacity to get it done. On the management side, Pete Waterman (NRL) wanted Westinghouse to use a new (to Westinghouse) Program Manager Concept. After a bit of in-house deliberation knowing Pete would not give up,Wx agreed and I became the first Wx Program Manager. I obviously got my guidance from Pete. We did get a follow on program from the Navy–FWCS (Future Weapons Control System). It was flight tested in an A3D.

Oral-History:Wayne Fegely

ethw.org

 

[Archibald]

Some details of the APQ-81 radar from the Missileer

Targets tracked simultaneously: 16
Frequency: C band
Average power: 2 kW (to be increased to 5kW later)
PRFs: 65kc/s, 11/10ths of 65kc/s, 11/9ths of 65kc/s
Aerial diameter: 60 in
Beamwidth: 2.5 deg
Range gates: 10
Scan limits: +-60deg in azimuth, +48 to -40deg elevation.
Frame time: 2 sec
0.3sec (2 x 0.15 sec) turnaround time, of which each half is used to transmit guidance commands to missile.
Aerial search: 30 deg x 5deg
Scan pattern: 2 bar scan
time on target: 60-70msec
Number of filters: 10 x 2 x 30
Maximum unambiguous range: 112nm
Target range rate: 215 - 2900 knots
Target altitude: 0-100,000ft
Range, 90% probability on 5 sq m target: 72nm

IBM 704 computer
Reflex klystron oscillator with 4 cavity klystron amplifier.

An early method of range while search was attempted. The radar switched between 3 prfs and then attempted to detect which range gate the signal is received in (output from doppler bank). However signals could overlap two range gates at certain distances. THey considered either measuring the number of times threshold was exceeded in each range gate, then assuming that the range gate where the threshold was exceeded most often was the correct one, or alternatively choosing one and then comparing the range information for consistency with previous range information.

Westinghouse weren't completely happy with the method of range measurement and were considering a linear frequency modulation to the transmitter instead [FM ranging].

Was it ever build, and tested, on a Skywarrior ? or did McNamara axe fell before this could happen ?

 

[Maury Markowitz]

IBM 704 computer

No... definitely not. The 704 was a very large tube-based machine:

IBM 704 - Wikipedia

en.wikipedia.org

I could believe that it was developed on, or simulated on, a 704, but it's difficult to believe a F6D could carry even a small part of that system.

Given the project start date at the end of 1958, transistorized computers were already fairly widely used, especially in aerospace.

 

[RetiredAFGuy]

IBM 704 computer

No... definitely not. The 704 was a very large tube-based machine:

IBM 704 - Wikipedia

en.wikipedia.org

I could believe that it was developed on, or simulated on, a 704, but it's difficult to believe a F6D could carry even a small part of that system.

Given the project start date at the end of 1958, transistorized computers were already fairly widely used, especially in aerospace.

That may have been a reference to a Raytheon 704. Those were 1950s/early 1960s "miniature" computers. All solid state, but no integrated circuitry, core memory and a tape drive. I would think any aircraft application would be a nightmare for vibration related problems.

The one I worked with in USAF tech school was packaged about 20 inches x 20 inches x 40 inches and looked like it was ruggedized. I didn't get in depth with it, my introduction to it was only for a one week programming course

 

[Basil]

Some details of the APQ-81 radar from the Missileer

Targets tracked simultaneously: 16
Frequency: C band
Average power: 2 kW (to be increased to 5kW later)
PRFs: 65kc/s, 11/10ths of 65kc/s, 11/9ths of 65kc/s
Aerial diameter: 60 in
Beamwidth: 2.5 deg
Range gates: 10
Scan limits: +-60deg in azimuth, +48 to -40deg elevation.
Frame time: 2 sec
0.3sec (2 x 0.15 sec) turnaround time, of which each half is used to transmit guidance commands to missile.
Aerial search: 30 deg x 5deg
Scan pattern: 2 bar scan
time on target: 60-70msec
Number of filters: 10 x 2 x 30
Maximum unambiguous range: 112nm
Target range rate: 215 - 2900 knots
Target altitude: 0-100,000ft
Range, 90% probability on 5 sq m target: 72nm

IBM 704 computer
Reflex klystron oscillator with 4 cavity klystron amplifier.

An early method of range while search was attempted. The radar switched between 3 prfs and then attempted to detect which range gate the signal is received in (output from doppler bank). However signals could overlap two range gates at certain distances. THey considered either measuring the number of times threshold was exceeded in each range gate, then assuming that the range gate where the threshold was exceeded most often was the correct one, or alternatively choosing one and then comparing the range information for consistency with previous range information.

Westinghouse weren't completely happy with the method of range measurement and were considering a linear frequency modulation to the transmitter instead [FM ranging].

As a general question to the experts of AI radars here - I always wondered about the mentioned peak power of fighter aircraft radars. Earlier AI radars like the SCR-720 had a peak power output in the region of more than 100 kW (see for example https://www.ibiblio.org/hyperwar/USN/ref/NightFighterRadars/index.html) and for AI radars of the 1950s even higher values are mentioned with 200 kW and more (I am not talking about average power).
Today's AI radar outputs are described with normally less than or about 10 kW of peak power. I am aware that radar designers today try to minimize the output power because of stealth but regarding the huge difference - is power output today measured / specified in the same way as in earlier decades?

 

[P-STICKNEY]

Some details of the APQ-81 radar from the Missileer

Targets tracked simultaneously: 16
Frequency: C band
Average power: 2 kW (to be increased to 5kW later)
PRFs: 65kc/s, 11/10ths of 65kc/s, 11/9ths of 65kc/s
Aerial diameter: 60 in
Beamwidth: 2.5 deg
Range gates: 10
Scan limits: +-60deg in azimuth, +48 to -40deg elevation.
Frame time: 2 sec
0.3sec (2 x 0.15 sec) turnaround time, of which each half is used to transmit guidance commands to missile.
Aerial search: 30 deg x 5deg
Scan pattern: 2 bar scan
time on target: 60-70msec
Number of filters: 10 x 2 x 30
Maximum unambiguous range: 112nm
Target range rate: 215 - 2900 knots
Target altitude: 0-100,000ft
Range, 90% probability on 5 sq m target: 72nm

IBM 704 computer
Reflex klystron oscillator with 4 cavity klystron amplifier.

An early method of range while search was attempted. The radar switched between 3 prfs and then attempted to detect which range gate the signal is received in (output from doppler bank). However signals could overlap two range gates at certain distances. THey considered either measuring the number of times threshold was exceeded in each range gate, then assuming that the range gate where the threshold was exceeded most often was the correct one, or alternatively choosing one and then comparing the range information for consistency with previous range information.

Westinghouse weren't completely happy with the method of range measurement and were considering a linear frequency modulation to the transmitter instead [FM ranging].

As a general question to the experts of AI radars here - I always wondered about the mentioned peak power of fighter aircraft radars. Earlier AI radars like the SCR-720 had a peak power output in the region of more than 100 kW (see for example https://www.ibiblio.org/hyperwar/USN/ref/NightFighterRadars/index.html) and for AI radars of the 1950s even higher values are mentioned with 200 kW and more (I am not talking about average power).
Today's AI radar outputs are described with normally less than or about 10 kW of peak power. I am aware that radar designers today try to minimize the output power because of stealth but regarding the huge difference - is power output today measured / specified in the same way as in earlier decades?

The lower Peak Power numbers of Pulse Doppler Radars vs Pulse Radars is due to their much higher Pulse Repetition Frequencies. (PRF, pulseds/second). As you can see from the data above, the Pulse Doppler APQ-81 is sending out on the order of 65,000 pulses/second, while the contemporary Pulse APG-37 (The basis for most USAF Interceptor Fire Control Systems) searched at a PRF of about 410 pulses/sec, and tracked at a PRF of 900-920 pulses/sec. What's listed above is Average Power, which is the product of Peak Pulse Power, PRF, and Pulse Width. The modulator of the transmitter stores up the energy it's getting from its power supply, and blasts it out when triggered. So, while peak power may be, say, 250 KW, if it's only transmitting .5% of the time, its Average Power is about 1.25 KW.
With the same input power to the Pulse Doppler radar, the peak power per pulse is less, but with several orders of magnitude more pulses, the Average Power is the same or higher.

 

[overscan (PaulMM)]

I was going to reply in the same vein. If you wanted the peak power of a pulse-doppler radar to be the same as an old-school pulse radar it would need to be attached to a power station to supply enough power. Its all about duty cycle.

Peak- and average power - Radartutorial

www.radartutorial.eu

 

[Basil]

Now I understand. Thanks to both of you.

 

[Archibald]

More APQ-81 documents. Seems Westinghouse at least built one complete radar and tested it on the ground. No idea if it ever flew.

 

[Maury Markowitz]

That may have been a reference to a Raytheon 704

1970 actually. So not that one.

Today's AI radar outputs are described with normally less than or about 10 kW of peak power

I was curious about this a while ago, I was writing mostly about WWII radars but took a slide into the Marconi Martello. Only 200 kW, compared to the Type 85 it replaced, which had a dozen 8 MW klystrons!

So basically it's all due to pulse compression. Instead of sending out a short burst of very high power on a single frequency, you send out a much longer pulse with lower peak power and shift the frequency during the pulse. Now of course, normally this would mean you couldn't tell where the target was in range, it's somewhere in that pulse length times the speed of light.

The reason you can do this is because on the receiver end you have a crystal (now done entirely in software of course) that slows some frequencies more than others. This causes the original spread-out pulse to form back into a short high-power one. That gets you the range accuracy you need.

Total energy can be the same, in theory, although that's no longer required because the receivers are so much quieter now that you can pick up lower energy returns.

 

[Basil]

1970 actually. So not that one.


I was curious about this a while ago, I was writing mostly about WWII radars but took a slide into the Marconi Martello. Only 200 kW, compared to the Type 85 it replaced, which had a dozen 8 MW klystrons!

So basically it's all due to pulse compression. Instead of sending out a short burst of very high power on a single frequency, you send out a much longer pulse with lower peak power and shift the frequency during the pulse. Now of course, normally this would mean you couldn't tell where the target was in range, it's somewhere in that pulse length times the speed of light.

The reason you can do this is because on the receiver end you have a crystal (now done entirely in software of course) that slows some frequencies more than others. This causes the original spread-out pulse to form back into a short high-power one. That gets you the range accuracy you need.

Total energy can be the same, in theory, although that's no longer required because the receivers are so much quieter now that you can pick up lower energy returns.

Thanks, interesting. Just curious regarding the last sentence - can it be assumed that the total energy output of a modern aesa radar like the AN/APG-77 in the F-22 exceeds that of the 1950s fighter radars with their (1950s) nominally more than 250kw pulse power?

 

[Maury Markowitz]

Thanks, interesting. Just curious regarding the last sentence - can it be assumed that the total energy output of a modern aesa radar like the AN/APG-77 in the F-22 exceeds that of the 1950s fighter radars with their (1950s) nominally more than 250kw pulse power?

IIRC the total average output from the APG-77 is something like 20 kW. This is not much different than, say APG-66, but is rather different than the APG-67's 350 W. The wartime SCR-720/AI X was about 170 W average.

 

[Basil]

IIRC the total average output from the APG-77 is something like 20 kW. This is not much different than, say APG-66, but is rather different than the APG-67's 350 W. The wartime SCR-720/AI X was about 170 W average.

... but the SCR-720 had a pulse power of about 70 kw. You are sure the APG-77 has an "average" output of 20 kw? According to Wikipedia 20 kw is the peak power.

 

[overscan (PaulMM)]

Old pulse radars had high peak power output but spent most of their time listening not transmitting (low duty cycle). High PRF pulse doppler radars transmit up to 50% of the time (high duty cycle), so the average power is much higher even though peak power is lower.

Modern radars like APG-77 will have all kinds of available cycles, but the goal is minimum output power for the required detection range.

 

[stealthflanker]

20 KW average power for fighter radar tho is i would say beyond the cooling capacity that can be provided by a fighter jets for radar. that value must be the Peak.

While average power is less depending on duty cycle, which in turn related to the PRF and Pulsewidth of the transmitted waveform.