I saw the ultimate technician's gadget yesterday...

[Elroy Jetson]

Anyone want to make a gift to me in the amount of 100,000 dollars or so?

Or 10 of you want to give 10K each?


My workplace (a two-way radio shop) sent me to a Tektronix info/sales seminar focusing on this monster:

http://www.tek.com/products/spectrum_an ... /rsa6100a/


This is a real time spectrum analyzer that operates on a totally different principle than a traditional swept frequency spectrum analyzer. Past the first IF stage, all processing including frequency measurement is done digitally, and it has an amazing array of ways to analyze and visualize the data, including some that are truly unique in the industry.

Regular SA's don't do well at capturing short duration signals. Pulses,
frequency hoppers, transients...none of the above are easy to capture
or analyze with them. The most I can tell you about your spread
spectrum 5.6 GHz digital cordless phone is that it does operate on a
handful of discrete channels in the 5.6 GHz band, if I sweep the band
with the peak hold function and keep sweeping. Sooner or later it'll
detect the phone's pulses but that's about all I can get out of it in the way
of useful data. Frequency, amplitude, and nothing else.

This Tektronix beast will capture the whole thing, demodulate it, and send
the demodulated digital I/Q data out the rear panel ports for recording or
decoding by external processors, in real time.

It makes over 48,000 full spectrum (within the selected span) measurements per second. It'll capture any signal lasting more than
about 2 microseconds. And it can deliver multiple view correlated data
on a single pulse that it captures, without having to receive a repeat of
that one pulse.

It'll do a lot more, too.


I enjoy using test equipment as much as I enjoy messing around with radios. To me, it's just another aspect of the same electronics hobby.

Do I NEED this monster? No. But yes, I want one.

Since it's operating on a Windows XP base, it may go obsolete pretty
darned quick for such an expensive beast. I waited 20 years or so
between the time I saw Tektronix's new 492/494 series spectrum
analyzers to get one for myself. It took that long for the price to drop to
my level. I think that maybe I won't have to wait so long for this
new machine to get to that level due to its dependency on Microsoft
and their known tendency to stop supporting their operating systems
in not so very many years. I think Win '98 is no longer being supported, so that suggests that this machine will be unsupported by
Microsoft in 10 years...and it may be hard to find a working, compatible
replacement hard drive for it as well. A lot can change in 10 years.

Personally I think it was a mistake for Tektronix to base the operation
of such an expensive piece of hardware on an operating system with
a guaranteed short lifespan, compared to how long the hardware might last.


That doesn't mean I don't want one!


Elroy

[tvsjr]

And I thought my little NetTek was pimpin:
http://www.tek.com/site/ps/0,,2E-17534-INTRO_EN,00.html

[cbus]

100k ? who apart from govt agency's have that kind of $$ to drop on gear like that ?

What has got me interested is the functions below. As i read it, it could e used to capture a encrypted transmission and analyse it ?? am I wrong here ?.

Fast analysis of complex interactions

The surveillance systems you build increase your need to work with a real-time analysis tool. The Tektronix RSA6100A Series captures signals that other analyzers miss with the unique frequency mask trigger, which enables you to capture complex signals into memory for complete analysis in all domains.
Capture once and analyze many times

Tektronix RSA6100A Series gives you the ability to trigger on a transient RF signal, seamlessly capture it into memory, and perform time-correlated, multi-domain analysis without having to recapture the signal. This makes it possible to reliably detect and characterize RF signals that change over time.
Flexible signal capture

Unlike any swept spectrum analyzer, the Tektronix RSA6100A Series offers real-time IQ output capability, which allows you to use digital receivers and RAID arrays for signal capture and offline analysis. The output is available up to 110 MHz bandwidth, with a 300 MHz sample rate, and is delivered in the LVDS format.
Signal storage and security advantages

Tektronix recognizes your need to secure all test setups and measurement data. Tektronix RSA6100A Series analyzers offer a removable hard drive or DVD-RW as options so you can safely port your data to wherever you need.

[mr.syntrx]

100k ? who apart from govt agency's have that kind of $$ to drop on gear like that ?

What has got me interested is the functions below. As i read it, it could e used to capture a encrypted transmission and analyse it ?? am I wrong here ?.

Yes, but it doesn't offer anything substanial in that regard. You can capture an encrypted signal on a traditional spectrum analyser, or for that matter, basically any receiver.

Regardless, you can observe an encrypted signal all day long, but it's still encrypted.

What it DOES make easier is capturing frequency hopping (AJ/LPI) traffic, which may or may not be encrypted.

Software receivers are not a technology new to EW or SIGINT, but this a pioneering beast in test equipment.

[Elroy Jetson]

As for who buys these things...100K isn't all that much in the budget of any decent sized tech company. During the demo, the man who was running the show mentioned that customers would be anything from major consumer electronics makers like Sony, Samsung, or Motorola, various auto manufacturers (many RF devices can be found in some modern cars, even including tire pressure sensors in the wheels), and virtually every company that's developing or making spread spectrum or frequency hopping equipment. Here in central Florida, it's probable that Harris Corp. will buy
not one but several units, and then there's Rockwell-Collins, Northrop-Grumman, Lockheed-Martin, Boeing, several Harris spin-off companies, and who knows what all else?

If a lottery comes my way, I'll buy one. Sure. And it will be the centerpiece of what could arguably be one of the most complete RF listening
posts in North America, once I built it at my new home site! There'd be
so many antenna installations on the property, you'd think you were
looking at a NORAD facility!

These analyzers will be surplus in 10 years or less unless they make a
big commitment to developing replacement computer hardware and software for it as the field progresses.




Elroy

[Wowbagger]

I've always been very uneasy about basing an instrument on any version of Windows, for a variety of reasons:

1) Long term support. It takes years to design a complicated instrument. It takes more years of selling that instrument to make back your non-recurring engineering costs of developing that instrument. If mid-way through that period Microsoft decides that version of Windows upon which you based your instrument is no longer available, you now *have* to spend even more NRE to move to whatever they have decided is the latest "WOW".

2) Low-level interfacing. We (IFR) had looked at making an instrument based upon WinNT at one point. Just getting to where we could hook into the ACPI/APM stack enough to power the system off when you hit the power button required detailed access to the sources - which costs big bucks.

And I'll tell you this: if you look under the hood of Windows at how drivers work, you will need therapy. I know - I've been there, I've done that, and I still wake up screaming at night. Drivers under other operating systems are MUCH simpler - and simpler means less bug prone.

3) (lack of) granularity and modularity - until very recently, if you used Windows (NOT Windows CE), you did not have much modularity - if Microsoft decided you were going to have IE, or Media Player, or Minesweeper - you were going to have them in your instrument, taking up disk space and acting as possible points of failure, even if they had zero relevance to your instrument's role.

While you had some modularity if you went with CE, you then lost what I consider one of the main reasons to use Windows, namely applications support. You can no longer say to a user "You want to use your $FOOBAR 7000 printer on our box? Great - just install the drivers!" or "You want to put your radio service software on there too? Knock yourself out!"

Now, WinXP embedded gives you *some* modularity, but not nearly as much as I'd like to see.

4) Security (or the lack thereof). You either have to let the instrument constantly update Windows (with the attendant risk of it breaking) or your instrument becomes a security risk if it is on the client's network. I'm not going to belabor this point - we all know the issues here.

5) CPU support - the X86 is not always the best choice for an embedded system, but if you go with Windows (once again, NOT CE), you are stuck with it.

And IF you go with CE, then you can use whatever microprocessors Microsoft chooses to support - and that can change. Look at what happened with the MIPS processor family: several PDA vendors were using MIPS based chips running CE. One day Microsoft said "Nope. We don't support MIPS anymore. Move to ARM." And all those companies had
a) units in the field they could no longer update
b) A sudden and unplanned NRE expense to change their code and their production line.

Now, there are even more reasons to be uneasy: With Microsoft saying that all 64 bit drivers SHALL be signed by them or they won't work, and in order to be signed Microsoft has to decide you cannot use your driver to violate their idea of DRM, this makes moving to 64 bit Windows very problematic.

"Why on earth would you want to use a 64 bit operating system in a test instrument?"

Good question: consider looking at a 40 MHz wide signal. By Nyquist's law, you have to sample that signal at LEAST 80 million times a second. Actually, you really need to sample it at least 160 million times to get a good fidelity. If you want 80 dB dynamic range, you need to be sampling AT LEAST 10 bits per sample - usually you will run at least 16 bits.

So there's 320 million bytes per second of data. If you are doing protocol work you may want to sample ten seconds of signal for analysis. That's 3.2 Gbytes of memory. Working with that kind of a data buffer is a PITA in 32 bits - you are constantly swapping memory into and out of disk, and so you can kiss goodbye any hope of real time operation.

With a 64 bit processor, you just do it - allocate a 4G buffer, DMA the data in, chew on it, done.

That's one of the reasons we (Aeroflex Wichita) prefer to use Linux as our OS - we can strip it down, we can hook into it, and we can move to damn near any CPU we see fit.

Now, if you want a really drool-worthy toy to play with, look at our PXI measurement solutions - THERE'S some real power.

[Elroy Jetson]

I looked at the PXI measurement tools, and that's a very different category of equipment. It doesn't really compare directly. It's very specialized in a different direction.

And, the RSA series analyzers are "portable", though it takes one strong
man or two average men to lug it around, while the PXI equipment is
clearly intended to be installed in one location and STAY there.


If I had the money, I'd get both.


Elroy

[Microwave Mike]

I had a personal demo at the Irving, Texas hamfest from a Tek Ham there. That is one nice RTA Spectrum analyzer. It was set up to 6 Ghz along with most options on. That set up runs around 120K.
We looked at digital UHF TV signals, and P-25 Signals, and a few ISM 2.4 GHz computer links. The built in RF preamp made finding signals easy.
The frame runs XP on, I think, a 3.6 GHz processor. Great color LCD.

mm