Connect with us

Simple spectrum analyzer for pre-compliance

Discussion in 'Electronic Design' started by Joerg, Sep 18, 2007.

Scroll to continue with content
  1. Joerg

    Joerg Guest

    Hello Folks,

    Clients often ask me what they'd need for pre-compliance. My default
    answer is the Hameg 5510, around $2k, probably more after the $ decline.
    Fits the bill, gets the job done, but it's hard to document your work.
    The Atten AT5011 costs half that but I don't know it, haven't used it.

    The problem with most of these: You can't easily dump the screen onto a
    PC, let alone memory stick. My new DSO does that very nicely but that
    company's analyzers are too expensive for this kind of job.

    Any other contenders? Ideally not a handheld but something that gets
    plopped onto the benchs and plugs into mains. And no, I am not asking
    again about USB analyzers as they don't seem to exist (I still think
    somone could become rich if they made one).
  2. Paul Mathews

    Paul Mathews Guest

    I use HP859x series of spectrum analyzers, particularly 8591EM if
    available. For an interface, I use Nat'l Instruments ENET, which
    converts the GPIB port to 10BaseT. The spectrum analyzer then has a
    LAN address, and relatively simple commands sent via the NI GPIB layer
    suffice to control the instrument and exchange data. Instead of using
    LabView, for example, I wrote a VB6 app to make it easier to store and
    compare traces, subtract baselines, etc. Not a straightforward
    approach, I'll admit, but it does the job. I'm quite fond of these
    old HP SAs.
    Paul Mathews
  3. Why not try photographing the screen with a digital camera?
    I've done this many times on gear that doesn't have output capability.
    Results have ranged from crappy but usable to excellent. Never as good
    as a direct screen dump to a USB stick, but good enough for most
    documentation purposes.
    A slow shutter speed is the go to avoid screen refresh problems etc. A
    mini tripod or Farnell catalog or two works a treat to keep it steady.

  4. Joerg

    Joerg Guest

    Looked on Ebay, none there :-(

    But I am not a fan of those GPIB garden hose connections. Best would be
    something simple. Small box plus USB cable, just like the new DSOs.
  5. Joerg

    Joerg Guest

    I am doing that right now. However, this gets old when you have to scale
    for a bumpy antenna gain curve.
  6. Joel Kolstad

    Joel Kolstad Guest

    I use HP859x series of spectrum analyzers, particularly 8591EM if
    Don't they have "instrument BASIC" built-in? I.e., you can write all sorts of
    data processing programs write on the instrument itself if you don't have a PC
    handy? Granted, these days that's unlikely, but in many ways it seems that
    those older instruments had more features than some of the new models!
  7. Check if your SA has a plotter output. There are PC based HPGL
    plotter emulators available:

    I've used a digital camera for grabbing scope traces. My black
    cardboard "scope camera" is ugly but functional.

    One device just wouldn't cooperate. I found where the X-Y signals to
    the scope section could be found and attached scope probes from a DSO
    (digital storage oscillosope) to grab the trace. Many scopes that
    have an external display or perhaps "HP storage normalizer" have
    external X-Y signals available on a connector.
  8. This mob are now saying that their analysers are "USB2"
    ready,(whatever that means). Includes free pc software

  9. Joerg

    Joerg Guest

    If you can get it from that dreaded HPIB connector to something more
    normal. There are ways but when on the road that gets old.

    Ok, call me lazy but I'd rather have a button that says "Store trace on
    USB stick".
  10. Joerg

    Joerg Guest

    I know that company a little (and their chip designer). They do not have
    any units suitable for the pre-compliance frequency range. Supposedly
    coming but right now you'd have to buy two units and that becomes
    expensive and cumbersome. And some clients need this capability now, not
    next year or so.
  11. OK, you're lazy. {;-)

    Did your friend in the investment game get sufficient answers from the dean
    of my department that runs the business department? I saw one answer that
    said he had to have a masters to teach that particular subject and saw no
    more correspondence.

    The only exception to that requirement for a masters is teaching
    Career/Technical Ed (what we used to call VocEd) such as welding, woodwork,
    electronics technology, and the like where a simple bachelor's in any field
    and X years working in the field. X is a function of subject and level.

  12. Joerg

    Joerg Guest

    Well, let's call it spoiled by a new DSO. Plugged it in, loaded their SW
    and bingo. Now I wish there was an easier method to get live pics in
    Excel. How is it than mankind always wants more?

    Haven't talked to him in a week, we usually see them during some of our
    dog walks. After your correspondence we had them over for a dinner and
    he was really thankful for your help.

    Why are they keeping such artificial barriers? In my case I'd have the
    masters but I bet they'd find a hair in the soup because it's from a
    foreign country, albeit formally blessed by the US goverment. One thing
    I really dislike about IEEE is that some of my dues go to their ABET
    "efforts". I am a firm believer in academia being able to decide on
    their own who is good and who isn't, no need for ever more bureaucratic
  13. Well, there are spectrum analyzers that have the tradition mess of BNC
    connectors on the back, which usually include something to drive a pen
    plotter. They're usually labelled X, Y, and pen lift. You can
    probably do something useful with that. XY pen plotters can still be
    found, or just use your DSO.
    You're lazy, but such laziness has been the inspiration behind some of
    my best ideas. Don't let it bother you (much).

    How about a compromise? Instead of a USB stick, one of the ever
    growing variety of memory cards that are being used in digital cameras
    will probably suffice. Build a cardboard or foam board hood and mount
    the camera at the correct location. Make it fairly long (about 0.8
    meters) so that you don't have a depth of field problem, where the
    focus is different in the display center and at the edges. Pre-focus
    and set the zoom. Paint the hood black inside. When the display has
    something that's worth capturing, push the shutter button. Clean up
    the picture with Photoshop or Irfanview.

    If dealing with the memory card is too much for you, run a USB cable
    to the camera, and capture the image directly on your PC. The only
    problem with this approach is that the "video" mode on most cheap
    digital cameras is 320x240 or 640x480 at best. That may not be enough
    for a detailed photo. However, I've seen spectrum analyzer photos on
    FCC type certification documents that look like they were frame
    grabbed from a video camera, so this method is probably suitable for
  14. Joerg

    Joerg Guest

    The pen lift thing would present a wee problem with the DSO. Other than
    that it could do it since mine and probably most others are capable of
    X-Y plots.

    That's how I do it now. Would be nice if I didn't have to ;-)

    But it's not so much me, it's for clients. Most young engineers are not
    very familiar with spectrum analysis, much less with vintage equipment.
    It would be great if I could tell them "Press this, then that and later
    the other button and shoot it over via email attachment". Then I could
    look at the plot and tell them "Place 220uH between D17 and C45, then do
    the spectrum again."
  15. Connect it to the Z axis input (trace inhibit) of your DSO.
    You don't have to. You can hire someone to do it for you.

    Perhaps hire an artist have have them sketch the display?
    You'll be amazed at how quickly they learn after you explain that
    their job, life, future, and salary depend on their ability to operate
    the device. No rush, take all night if you need to.
    I have a camcorder for such exercises. I've video taped how I setup a
    rather messy rack full of acronyms before it was shipped. On arrival,
    the techs followed the video religiously, including purchasing a trash
    can to match the one I accidentally left in the video. Immitation
    works amazingly well, even with little thinking involved.

    Incidentally, I intentionally delayed the shipment of the manuals so
    that they had to follow the video and could not get creative. The
    problems started immediately after the manuals arrived, when the
    tweaks, adjustments, and "calibration" began.
    You really are lazy. Pry yourself out of your overstuffed easy chair,
    get up, and go look at the spectrum analzyer display. Then tell the
    tech to make the changes. Think paperless office or save a tree. When
    you're happy, say "print it" and do the photographic ordeal process.
  16. Joerg

    Joerg Guest

    I am afraid it doesn't have that function. Unless it's thoroughly hidden
    Hmm, I might not even have to, my wife is pretty good at pencil
    drawings. Now I'd just have to cajole that pulse noise into sitting
    still for 10 minutes or so.
    There remains a wee problem though: I am a consultant and the engineers
    I am talking to are anywhere from 25 miles to 6000 miles from here. We
    do have a runway right in town and my office is next to it but I haven't
    found the matching LearJet on EBay yet.
  17. krw

    krw Guest

    I second the recommendation for the 809x analyzers. I bought one
    several years ago and sobbed when they sent it and the calibrated
    antennas out to scrap a year later. :-(
  18. Paul Mathews

    Paul Mathews Guest

    My clients typically pay about $1000/day or more for EMI test lab
    time. I've come into situations where they've already spend $50K or
    more on repeated trips to the labs, so they are sometimes ready to
    discuss spending $20K on some pre-compliance gear. The engineers
    ALWAYS need training, although I agree that it would be a little
    easier if the equipment was easier to use, and we can hope that we
    won't have to wait much longer. In the meantime, anyone can learn a
    lot by sniffing around with near field probes, particularly B-field
    probes, and the controls of most portable spectrum analyzers are
    really pretty simple: Start Freq, Stop Freq, BW, Sweep Time, Level/
    Attenuation are the main ones to know about. They need to understand
    how Sweep Time needs to be long enough to capture sufficient energy to
    measure at each frequency. A good broadband preamp is a necessity, but
    the Ramsey SA7 is available for $20 in a kit or about $70 assembled.
    You can pick up an older Anritsu, HP, Advantest, Marconi, Tektronix or
    other brand portable spectrum analyzer for under $2K with GPIB. It's
    not too difficult to get any of these to respond to remote commands.
    If anyone wants to use it as a starting point, I'll share my VB
    sourcecode for an app that controls the SA, uploads and downloads
    traces, overlays EMI compliance limit lines, compares traces before
    and after mods and setup changes, sets up the SA in various default
    configurations, etc, etc. I have HP959x and Anritsu MS420 versions. No
    support beyond commented sourcecode.
    Here's some articles that teach about building your own near field
    probes and LISN and how to use them: and downloads.htm

    EMI sniffer probes: buy from Bruce or use his design
    (see link on page) (see middle of page and get ready to be

    The link above gets you to more info than you'll ever need to build
    and use your own pre-compliance probes.

    Wideband Preamplifiers:

    Search ebay for HP 8447A/B/C/D and pay a few hundred $ or
    this looks interesting, too:

    other good sites:
    (first of a series of 7 excellent articles by Sanjaya Maniktala)
    (last of the 7 above) Testing Fund.pdf

    Paul Mathews
  19. John Larkin

    John Larkin Guest

    If a decent spectrum analyzer costs $12K, and it may last for 10
    years, that's 1200 a year. Less tax depreciation (or just expense it)
    that's maybe $800, under $70 a month. That's not even coffee+pastry
    money around here. It usually makes sense to buy good new test gear
    and spend your time designing products.

  20. Paul Mathews

    Paul Mathews Guest

    I agree, although many small companies have trouble coming up with
    capital expense money, particularly after they have blown big bucks on
    failed compliance efforts. Then, they bring in the expensive
    consultant, and it can be even harder to get the money. Most companies
    I've gotten into are set up in a way that encourages wasteful internal
    spending over judicious external 'off-budget' spending.
    Paul Mathews
Ask a Question
Want to reply to this thread or ask your own question?
You'll need to choose a username for the site, which only take a couple of moments (here). After that, you can post your question and our members will help you out.
Electronics Point Logo
Continue to site
Quote of the day