RF Test Gear

Most of the RF work I do is in the 902-928MHz ISM band.  I have lots of RF test gear, but concluded that I really only need a few pieces and found that great bargains can be had by purchasing and re-purposing surplus cellular test equipment on eBay.  Some cellular gear includes general purpose RF test capability and this has allowed me to equip my lab at a tiny fraction of the cost of comparable new general purpose equipment:

  • Aug 2023
    I bought a Siglent SSA3021X+: which is a modern, compact 2.1GHz spectrum analyzer+TG that is easily hacked to become an SVA1032X (the same hardware platform running different software). The hack expands the BW to 3.2GHz, adds VNA (S11, S21) capability, and enables all options. The SpecAn covers 9kHz to 3.2GHz, VNA=100kHz to 3.2GHz, 0.7dBm amplitude accuracy, 1ppm reference, -98dBc/Hz PN, -161dBm/Hz DANL, and generally decent specs with a modern user interface. Costs around $1650 from TEquipment (whom I highly recommend).
  • Oct 2020:
    small gear that I can use at my desk often gets more use than fancy lab gear that takes up too much space.  I bought a TPI-1005 RF USB-powered signal generator/analyzer and love it.  What distinguishes this from the toy signal generators widely found from China is that it is calibrated and accurate.  What distinguishes it from lab gear is harmonic suppression (or lack thereof).  However, it is an incredibly handy tool, compact, and reasonably inexpensive.  You can generate signals on frequency with precise amplitude, measure signal strength, and sweep antennas and filters.
    Oddly, the device comes with two (rarely used) push-buttons without caps.  The caps can be ordered very inexpensively from Digikey, Mouser, etc.: Wurth 714651010100, 714656010100.
  • Feb 2016:
    An Anritsu MT8222A (see datasheet).  Is meant for cellular base station service, but it includes an even richer set of features than my prior favorite cellular test tool, the Agilent E7495.  Most importantly, the MT8222A includes VNA capability!  It also provides a spectrum analyzer that covers 100kHz through 7.1GHz with decent specs (PN = -100dBc/Hz, 1.25dB amplitude accuracy without PSN50, 1ppm reference). It provides a broad set of 1 and 2-port swept gain/loss analysis tools for cable and antenna analysis from 10MHz-4GHz.  It is significantly smaller and much lighter than the HP/Agilent E7495A and the battery works!  At some point I need to try it with Anritsu Handheld Software Tools or Master Software Tools. I use it with an Anritsu PSN50 power sensor (50MHz-6GHz) for high accuracy (0.16dB) measurements (see datasheet).  Unfortunately it has no CW/AM/FM/FSK signal generation capability although obviously the hardware is capable, but probably not in a leveled way.
  • Anritsu MS8609A – 13GHz RF Spectrum Analyzer
    A lab-grade 10GHz spectrum analyzer is needed to measure harmonic compliance prior to FCC testing.  I use an Anritsu MS8609A (13+GHz) which is not quite as good as the best HP gear, but is fairly modern and more than good enough for my needs; it includes a bonus power meter.  Mine has a rubidium frequency standard too so I can slave my other gear when high frequency accuracy is needed.  See the datasheet.
  • Marconi/Aeroflex/IFR 2025 – 2.5GHz RF signal generator
    Generates pretty clean RF signals with analog modulation.  Not HP quality, but still plenty good and more than meets my needs.  It is a pleasure to use. See the datasheet and  manual.

Hobby-grade RF Tools
Hobby-grade RF tools have gotten a *LOT* better over the last 5 years. I had earlier versions of these and they were marginally useful…more toy than tool, but their quality has improved steadily and these RF essentials (SA, VNA) are now useful as tools and an amazing value for a little over $100 each; they make RF design much more accessible to hobbyists.

  • Tiny SA Ultra (100kHz – 5.3GHz) spectrum analyzer (V 0.4.5.1)
    Yep, a 5GHz spectrum analyzer with bonus signal generator for under $150. Crazy!
    Specs are getting decent too: PN is now -92dBc/Hz, 2dB amplitude accuracy, leveled signal generator provides CW, and also AM and FM modulation. See the wiki. Some important limitations: 200Hz min RBW, the SigGen cannot serve as a tracking generator for the SA. See links here. and the developer’s youtube videos here which include honest assessments of the limitations and how to work around them. This video describes the ultra and its limitations. It’s most useful when you limit the signal level into it using a fixed attenuator.
  • NanoVNA SAA-2N (V2_2) 50kHz-3GHz – surprisingly accurate; also almost out of the toy category. Due to its size and battery operation, it is often what I grab to do a quick sweep of an antenna. The manual, firmware, VNA Qt PC software, etc. are here. $110 for a VNA…

  • I’ve also added a separate page on the frequency counters I use, however I rarely use counters these days. Most of what I need to measure is under 100MHz and for convenience, I just use the counter built into the SDG2042X signal generator which is already on my desk; when slaved to an external 10MHz DOCXO standard, it is quite accurate.

BG7TBL USB RF Signal Generator

Some time ago, I purchased a super-cheap RF signal generator on eBay from fly_xy.  It was the same as this item.  This is yet another version of the the popular BG7TBL signal generator and “simple spectrum analyzer”.   It cost $65 and covers 138MHz through 4.4G; for another $20, you can get one that goes down to 35MHz.  It’s certainly not lab gear, but for the price and size, it’s still decent; the main challenges are the software and the resolution bandwidth.

USB RF Signal Generator

Harmonics are not attenuated; with the generator configured to output a 915Mhz CW signal, the fundamental lands at 914.993Mhz @ -5dBm signal (7ppm error), pretty clean to -40 to -50dBm, acceptable to -70dBm.  However the harmonics are ugly: 3rd = -13dBm, 4th = -24dBm, 5th = -36dBm, 6th = -48dBm, etc.

Inside are:

  • AD8307 500MHz demodulating log amp
  • AD4350 RF synthesizer
  • IAM 81008 Mixer
  • ATMega processor
  • FTDI FT232RL USB to serial interface
  • AMS1117 super cheap LDO

A review (translated from Polish) that measures harmonics and frequency accuracy and such:  http://sztormik.com/Radio/Wpisy/2014/10/1_BG7TBL_measurements_files/bg7tbl_lo_meas.pdf

It would be *much* more useful with selectable RBW filters; as it is, I can’t resolve much detail finer than 200kHz, so an FM signal with 25kHz deviation will look the same as an FM signal with 100kHz deviation.  See the pictures below for some sweeps of the 2m band.  The NWT4000 or NWT4000-2 might be more capable (but they are sufficiently more expensive that it would be better to buy a real piece of lab gear like a used R&S CMU200 – even though it is admittedly much larger).

Sweep of 2m band

2m band

Close-up of signal in 2m band

The software it uses is open-source LinNWT / WinNWT which supports many similar devices.  The English language documentation is not great and because the native language is German, there’s not much English language support.  A German magazine article is here it looks like it would be very useful if translated into English.  The author (Andreas) is friendly and helpful and responds to email.

To run WinNWT in English on windows, use the command (in the Target field of the shortcut properties):

“C:\Program Files (x86)\AFU\WinNWT4\winnwt4.exe” app_en.qm

To build/run the software on linux:

  • Download latest .tgz source here
  • Install tools if needed: sudo apt-get install gcc qt4-qmake libqt4-dev
  • tar xzvf linnwt_X_Y_Z.tar.gz
  • cd linnwt_X_Y_Z
  • qmake-qt4
  • make
  • If you want to make it available system wide: sudo cp linnwt /usr/local/bin/linnwt
  • sudo linnwt app_en.qm

To use the software, I configure Settings->Options->StartFrequency=80000000, StopFrequency=100000000, DDS clock=400000000, Interface=/dev/ttyUSB0 on linux or COMx on windows, max.Sweep=200000000, frequency multiply=10.  Then in the Sweepmode tab, to monitor the ISM band, configure StartFreq 902000000, StopFreq=928000000, Samples=1000 and press Continuous to sweep continuously; press Stop to stop sweeping.  This results in 26kHz steps.

I also bought the BG7TBL “tracking” (i.e. broadband) noise source in the hope of being able to do some basic antenna return loss/swr measurements; it does generate broadband noise at around -40dBm, but so far I’ve had no luck using it with the “simple spectrum analyzer” and a return loss bridge to sweep an antenna.
noise_source

Links:
https://www.facebook.com/media/set/?set=oa.809669625718111&type=1
http://ea4eoz.blogspot.com/2014/07/modifying-bg7tbl-noise-source.html
http://ea4eoz.blogspot.com/2014/07/installing-bg7tbl-noise-source-as-poors.html
http://www.mikrocontroller.net/topic/336482 (English Translation)
http://cas.web.cern.ch/cas/Germany2009/Lectures/PDF-Web/Caspers.pdf
http://bg7tbl.taobao.com/  (English Translation)
https://vma-satellite.blogspot.com/2019/04/new-simple-spectrum-analyzer-device.html (newer D6 version)
https://www.rudiswiki.de/wiki9/SpectrumAnalyzer_LTDZ#Links (newer LTDZ version)
http://alloza.eu/david/WordPress3/?page_id=478 (SNASharp software)

Rigol Oscilloscope

Update 2023: This post was from 2014; I sold the Rigol scope many years ago. It was great in its day, but I’m sure it’s discontinued and there are certainly better choices now. One of the best value bench scopes today is probably the Siglent SDS 1104X-E; in no small part because, like the Rigol, it is easily hacked. Also like the Rigol scope, the Siglent fan is too loud and I’m probably going to have to address that soon (the consensus seems to be that this is best accomplished by lowering the fan voltage to 9-10v)

Original 2014 post: The Rigol DS1052E is a fantastic oscilloscope and an amazing value at $329.  It is every bit as good as my favorite Tek scopes (maybe even better) at a fraction of the price.  I have previously owned/used budget scopes (Owon, Wittig, etc.) as well as Tek and HP scopes and this is hands-down my favorite.

DS1052E

There is a software-only hack to convert it from 50MHz to 100MHz bandwidth…I’ve done the hack and can confirm it works.

The only problem with the scope is that the fan is 36.5 dBA (very) loud.  I finally got around to replacing it with a 22dBA Xilence 60mm case fan (less than $10 shipped on eBay) and it is now delightfully quiet.  The Xilence fan spec says it moves 30CFM (more than the 21CFM of the original); it is audible, but inoffensively so: mostly noise from the air moving through the plastic case, not from the fan; it is dramatically better than the original fan.  Note: opening the scope case voids the warranty; the process is well documented and requires a T-10 Torx screwdriver (Craftsman 41473) which is also useful for opening Tivos.  I cut the connectors off both fans and used the connector from the old fan with the new fan.

Warning: if you open your scope, please make sure it is unplugged and observe proper AC safety procedures.

Future: