[Discuss-gnuradio] Interesting piece of test equipment

[Ward Ramsdell]

A few months ago I picked up a used Agilent E4406A VSA to play around with and I thought I'd share a few of my observations.  In short: I think it's a very good piece of equipment for the money, and I'd love to see it develop a broader following in the experimenter community.

The E4406A was initially intended for use as a transmitter tester, covering the 7 MHz to 4 GHz range (5 Hz-4 GHz with the B7C I/Q input option).  I'd seen them all over eBay and the used test equipment houses for relatively cheap and, encouraged by an Agilent forum post (link below), when they got cheap enough I bit the bullet.  I think the market is hampered by the fact that it's viewed (partly, rightly so) as a very specialized piece of gear, and supply is definitely outstripping demand.  They can be often be found for sub-$1000.

The number one thing that attracted me to the instrument is that, as it's intended to be a VSA, the LO is not swept and has pretty good phase noise characteristics...certainly better than anything else in the price range.  This obviously works to its detriment as a general-purpose spectrum analyzer, but for narrow spans (<=10 MHz) it's great.  I've had good luck retrieving the spectrum trace in Python (using PyVisa) and piecing together multiple shots for a broader span...the instrument has an Ethernet interface and tunes very quickly, so grabbing a good swath (100s of MHz) doesn't take much time at all.  If it was all I had for spectrum analysis, I could certainly make it work.  John Miles, KE5FX, has done a lot of work to support it with his tools, link below.

There's good potential for pulling out the I/Q data for offline (or even near-real-time) analysis.  Agilent's older 89601A VSA software supports this box, and I've successfully made some very good DQPSK measurements using the trial version.  I haven't gone as far as sniffing the Ethernet to see what's being sent across, but I'd imagine they're just using the standard calls which are well documented in the Programmer's Manual.  Getting Python to talk to the instrument was trivial, so I think there's some really interesting work that could be done here.

One reason I mention this on an SDR forum is that there's potential to use the instrument as a tuned front-end for an IF digitizer.  There's provision to take the 321.4 MHz IF out, and although my unit didn't come with the IF option, it took less than an hour to figure out how to wire the connector on the RF deck to a back panel connector.  Happy to provide guidance here if anyone's interested.  Quick measurements show the IF bandwidth to be fairly flat to 100 MHz or so, though I haven't done full testing and there may be some gotchas in various modes or at certain frequencies.  It's obviously wider than 10 MHz.

One thing to note: since this is a transmitter tester, there's no preamp and the noise figure is almost guaranteed to be lacking. In fact, there's a beefy 7 dB pad right at the front panel input connector, just in case there was any doubt as to its intended application.  That's certainly not an insurmountable limitation.

All in all I think this instrument is definitely worth a look.  It's not a replacement for a general-purpose spectrum analyzer, but has other attributes that might make it very valuable to the radio/SDR experimenter.

--Ward

http://forums.tm.agilent.com/community/viewtopic.php?t=3857&f=50 (the forum post that sparked my interest)

http://forums.tm.agilent.com/community/viewtopic.php?f=31&t=25805&p=75890 (further info)

http://tech.groups.yahoo.com/group/hp_agilent_equipment/message/36988 (John's take, and links to his software)