Teardown - Military sigint radar signature analyzer

[TerraHertz]

The sad news on Friday of MH17, and the question of who shot it down, reminded me of an old bit of ex-military gear I've had in storage for many years. It's kind of relevant, so here's a partial teardown of the unit.

It's a SCP-2160 Pulse Analyzer, made by SciComm  (Scientific Communications) sometime in the mid to late 1980s.
Back then it was probably a highly classified bit of gear. I've been in contact with someone who was involved in designing it, and he tells me the security in SciComm was so tight that he never got to keep a manual or any other documents about it. Which was really disappointing, since I don't have a manual but want one very much.

The unit's purpose is to analyze the characteristics of radar signals detected in a military theater, and characterize the nature of the various signals in order to identify what kind of equipment is transmitting them.

All radar systems work by generating a pulse of RF energy, transmitting it in a desired direction, and listening for echoes of the pulse. With civilian radars (ie air traffic control, weather, etc) the design aim is usually to make a nice clean RF pulse, at a fixed frequency and repetition rate. For military radars the aim is to make a system that is as immune to jamming as possible, so just about everything about the transmitted pulse will vary as much as possible, and rapidly. The frequency will vary from pulse to pulse, and during the pulse. The repetition rate will vary, and so on.

Nevertheless, there tend to be detectable fingerprints in radar transmissions. Things like unintended frequency drift during the pulse, and the profile of the pulse's power envelope will vary between systems. They even vary enough between units of the same model, enough that battlefield electronic warfare monitoring can often identify the radars used by specific units of the enemy, and track their movements. Obviously very useful information during a conflict. Which is why in any combat zone such as Ukraine presently, there will be sigint/EW units operating 24/7, gathering all the intelligence they can. So there are certainly multiple groups who do know for sure who fired that missile; likely even down to the specific launcher truck.

How I came to have this instrument was amusing. I'd attended a Wyong ham meeting (years ago) and at the end of the day when the car boot sales people were packing up to go home I spotted one seller lugging a couple of large rackmount things over to a dumpster and heaving them in. I asked him why he was throwing them away, and he said no one wanted them and he couldn't be bothered taking them home. They were a pair, but had no interconnecting cables. One was some kind of rackmount printer, but the rubber paper feed roller had depolymerized and turned into a river of vile sticky goo dripping out of the slot in the front panel.

The mating unit looked kind of cool, though neither he or I knew what it was exactly. Apart from heavy. I figured it might contain some nice parts, so rescued it from the dumpster. I left the gooey printer, which in hindsight was a mistake.

When I got it home, it powered up and responded to controls. Only then, looking at the labels on the controls and what they implied (and the stickers) did I realize what it was. It's an electronic warfare radar signature analyzer.

But alas, no manual to be found anywhere. Also it's supposed to be mated with a receiver unit, and there's no way that's happening. I don't even know what kind of signal it wants to see on the 'RF in' connector. So it got buried in the storage area and forgotten about. Digging it out today I didn't expect it would still work. But it does seem to - at least the controls respond, and even without a source signal the screen can show something sensible.

Maybe someone here knows something about it?
Also it has one of those obscure round-body military three pin mains connectors on the back, which I don't know where to find a socket. Alligator clips for mains connection isn't a permanent solution.


Fiddling with the controls, it seems to display a real time spectral histogram, with the X-axis being frequency and the Y being amplitude. For repetitive radar pulses, the result will be a bunch of dots on screen, with smearing and positional jitter as the pulse characteristics change. In the real world there'd be multiple different signal sources, so a way is needed to select groups (from one source.)

The rectangle on screen is a selector box, that can be sized and moved about the screen with the 'QUAL LIMITS' controls.  This is to allow selection of pulse groups. Some other controls are obviously for input attenuation, threshold of acceptance of pulse bursts, and a range of pulse durations of interest.
After that I'm lost.

Something really nice to see - reading and saving the EPROMs to hard disk, the read checksums match the values printed on the labels. After nearly 30 years, the EPROMs are still good! I'm amazed.
Also it's great to find there are extender cards included for the two different card types. Now if only I had a manual...

There are 20 boards, plus front panel, CRT unit and power supply. I only included a few of the boards here. They seem to mostly use parts typical for the '80s - a mix of 74LS, 74H, some ECL, and assorted analog. The ones that intrigue me are the huge TRW chips. They are mounted in ZIF sockets! Also the CPUs (2 of them) are a number I don't recognize - 65SC02, which presumably is a 6502 variety. Yay, the CPU I first learned assembly on.

[SeanB]

That looks familiar.......... DENEL managed to make one that would fit into a plane, and which would work to detect where a RADAR guided missile ( or a SAM site RADAR) was coming from. Had a small CRT display in the cockpit and a great big box inside the avionics bay. We had a simulator that emulated a lot of missile systems and SAM sites, including a whole lot of NATO, USA and USSR units. That could be fun at times. You kind of had to tell the pilots what you were going to use so that they would not cream themselves when the panel lit up, to test it would detect more than the designated test signal.

Your receiver would detect the various RADAR signals direct and gave a baseband out of around 70MHz that would be fed down to the display unit. Basically a very frequency agile satellite LNB and first IF section of a TVRO unit. It was scanned rapidly around to get the largest signals by the microcontroller then it would demodulate the IF to give an AM video signal for the display. The TRW chip would be easy to decap to see what it is, likely a very custom ADC and a lot of logic as well to do video digitisation for the unit.

BTW those sockets are available from RS. Look at the circular locking bayonet connectors, then get the matching pins and sockets, then get the crimp tooling for them, the insert/extract tooling, the go/nogo gauges and then look at the price and rip it out, or solder wires to the pins and sleeve them. Only plug and socket where the tooling was cheaper than the plug and socket, especially when you got to the higher pin count connections ( 200 plus pins was the biggest I ever used), and the more esoteric polarising arrangements ( which were $%##* impossible to order, you just got the regular one and could only use the pins and sockets unless you found a good shell in the graveyard), and the mixed signal type connectors, and the tooling was not cheap either. We had some that failed with monotonous regularity, which was not a surprise seeing as the box in question was mounted about 2cm away from the afterburner shield, and had about 3cm of vibration with full power. Pins went in 3mm thick and broke when they were about 0.5mm thick. Pin sockets fared not much better.

[zapta]

That's a fascinating story.

They seem to mostly use parts typical for the '80s - a mix of 74LS, 74H, some ECL, and assorted analog.

Not the 54xxx military versions? I thought that this was what was used in that era for military equipment.

How did you power it? It's not 110V 60Hz I presume.

BTW, I share your curiosity about who shot MH17. The suspects pool is very small. ;-)

Edit: it can still be classified as munition and subject to ITAR restrictions.

[SeanB]

54 series would only be used on board the plane, the test equipment will use regular 74 series, as that is very reliable, and if it fails it is designed to be repaired in any case. A lot of avionics actually used 74 series logic in places, where the 54 series was either not available yet or would never be done. It is simple to take a regular 74 series batch and run the 54 series qualification testing on it, and this both weeds out the weaker parts and improves the probability of the remainder working for the expected lifetime. You miss out on the 54 series packaging in ceramic in some cases, but the plastic 74 series are very reliable in any case, especially in an environment where they are protected by either a conformal coating or by being hermetically sealed. A lot of 74 series in any case was also available in cerdip or side brazed packaging for automotive use in any case, which is essentially 54 series less the 100% burn in and QC that the mil spec are subjected to. Same die, same plant, same packaging, same in process checks and just not temperature cycled.

[jeremy]

I'm extremely curious as to how the "emulator" units were made. Did they hide in the bushes near the active site with a spectrum analyser, or did they recover the information by procuring the documentation?

[Alexei.Polkhanov]

Do you have any kind of pulse gen/function gen in your lab? Try to send it into RF input of the unit to see what it will show.
Looks like this is the unit that meant to be used in a LAB to tune radar transmitters. Radar output connected to a directional RF coupler and fed into RF input at the back of of this unit.

I was looking at more modern version of unit like this once thinking maybe I can make use of it in my toy project - which was the time of flight 3D laser scanner. Resolution was too low making it  useless for modern applications. Today you can do almost anything this units do and extra with SDR transceiver like USRP2 and GNU Radio.

[TerraHertz]

Do you have any kind of pulse gen/function gen in your lab? Try to send it into RF input of the unit to see what it will show.

Yes, I could rig up an RF generator through a pulse modulator controlled via another arbitrary waveform generator. But there are too many other more demanding projects, and a lot of the controls on this unit really are not clear. Fiddling with it I did manage to work out a few more features. But it's going back into storage, for if I ever get a manual. Simply not worth wasting time on otherwise.
I got some pretty cool looking displays, that I think I understand. But still guessing, so what's the point?

Looks like this is the unit that meant to be used in a LAB to tune radar transmitters. Radar output connected to a directional RF coupler and fed into RF input at the back of of this unit.

No, it's definitely for theater EW/sigint. Otherwise there'd be no need of all the facilities for selecting specific RF pulse characteristic groups from among many others. At least half the controls are for that. It works like:
1. Display everything going on.
2. Manually select a subset of the pulses, qualified by freq, power and pulse width min and max limits.
3. Within that set, pick an individual pulse number (can scan through them.)
4. Analyze characteristics of the pulse. Pulse width, delta F (MHz), amplitude, amplitude variation (A2-A1), duration between manually selected points in the pulse (T2-T1), and things called "F MOP" and "A MOP" which I don't understand yet. Frequency and Amplitude something... MOP, hmm.... stands for what? Mighty Obscure Parameter?

I've been told there are lists kept of these signature values for all different types of military radars, and so you can look up the list to find what you are hearing in operation.  Or use gear like this to characterize new ones for the list.

I was looking at more modern version of unit like this once thinking maybe I can make use of it in my toy project - which was the time of flight 3D laser scanner. Resolution was too low making it  useless for modern applications. Today you can do almost anything this units do and extra with SDR transceiver like USRP2 and GNU Radio.

Not doubt true. At some point I want to play with an SDR and a really big parabolic dish I have. (I finally sold the smaller dishes.) But not for a while yet - too many other projects in the air.

I will be changing the mains connector on this to an IEC before it goes back into storage though.

[zapta]

I will be changing the mains connector on this to an IEC before it goes back into storage though.

What power voltage/frequency does it take?

[nctnico]

It says that on the picture of the rear 

[TerraHertz]

A few more shots while I had it apart to replace the mains plug.
Turns out the display is quite a nice stand-alone X-Y-Z unit. It even has its own mains input IEC, though it isn't used in this machine. Instead an add-on plug provides regulated DC rails direct from the main system supply.
Strangely there doesn't seem to be any model number for the X-Y-Z display, though there is a company name: Data Check, San Diego CA.  For a moment I had hopes of finding schematics for this, but nope. Not so far.

There's also a nice SMA'd band pass filter (the tubular metal thing in the pic), but I can't find data on this either. It's:
 FSY Microwave Inc. BC160-50-6SS   67195   SN 111

The mains plug change is a bit of a rough hack, since the rear panel could not be fully detached, making it awkward to work on. But it does the job.

Last pic is another display format. Movable cursor lines are T1, T2 and pulse number selector. Still without bothering to provide any input signal - it's just seeing noise down at the bottom of its sensitivity range.
No, I don't really know what it means.

[TexDoug]

I have most of the original schematics and test data as I was involved in the design of the SCP-2160 Radar Pulse Analyzer. I even have some of the tested components.

[coppercone2]

I thought it was funny, when I looked up the delay line, it showed up as lobster tail. I guess they think the center tapped inductor equivalent circuit looks like a lobster tail. (its not an exact translation but google has a mind of its own)

What is the story behind the spiradel company name?

And I have some of those SMA-MW filters. Why is it in the chassis without a bracket? They are sort of heavy. It seems like a vibration hazard, unless he removed it from a clamp. I figured it would be mounted in a pipe clamp.

I like your design, it makes me think all the fabrication stuff I figured out how to do is not useless. It can get pretty depressing looking at pure CNC milled electronically designed equipment tear downs all the time. (modern single PCB spectrum analyzer the size of a laptop).

How did the assembly line look like for that part? Were the technicians highly paid?

And, was this equipment designed to be put on a avionics aircraft (AWACS?)? Or some kind of command vehicle? helicopter? Or in a radar trailer? (curious about the mechanical requirements, and what that kind of construction is rated to. It looks like a IFR portable SA rated for use inside of planes).

[cdev]

I was just fooling around with the SDR software Linrad and it has "Radar" features.  Ive never investigated it. But there it is. Maybe its explained somewhere on its author's extensive web site. His software has a wealth of anaytical features which I have just barely scratched the surface of. As  it has radar mode, perhaps you could learn a bit about how radar works from Linrad's radar features. (If you have compatible hardware) It can work with a sound card SDR.  It probably handles more different SDR hardware than any other piece of SDR software.

its at https://sm5bsz.com