Connect with us

how to learn low level RF design

Discussion in 'Electronic Design' started by [email protected], Dec 9, 2008.

Scroll to continue with content
  1. Guest

    I work as an EE, I don't have a degree, but I do have a working
    knowledge of analog and digital electronics and have worked on a very
    wide variety of circuits.

    I have always wanted to learn low level RF "black art" circuit design,
    but its just too difficult on my own, and believe me I have tried.

    Whats the best kind of job or environment to get started in this? A
    "furnace" to be forged in?
     
  2. No Spam

    No Spam Guest

    Assuming your talking about the more modern and harder to understand RF
    in the microwave range.....

    Try ham radio. There are allot of v/uhf books around and getting a tech
    lic (in the USA) which will allow you on that band is only 25 SIMPLE
    questions. Look for people who are hams that do microwave contesting and
    "buddy" around with them. Honestly, they will be honored to help.

    If your talking HF radio below 100Mhz, your not looking/working hard
    enough :)
     
  3. EE123

    EE123 Guest

    RF is difficult. You really should have an MSEE to be proficient at
    it.
    THe big problem is that RF extends well into the GHz range.
    Even if you stay in the upper 100's of MHz, it is fairly easy to get
    tripped up
    That would be police radios for example.

    Post a question and I will try to answer it!

    Dave
     
  4. Joerg

    Joerg Guest


    Designing an HF receiver that can listen to a teeny signal 20kHz from a
    station that makes a fluorescent lamp glow _is_ hard work :)
     
  5. Guest

    NICE. :). What was the station?

    I am talking about making a transmitter and receiver capable of
    sending and receiving voice with 3khz bandwidth 1 mile line of sight,
    so I suppose that puts me sub 100MHz, and also sub-par as far as my
    effort to teach my self it according to one of you. :)

    I figure if I could do that, from scratch, using discrete components,
    that I would be able to accomplish all of my RF circuit goals for
    life, which are basically farting around for fun.

    Yeah you are right, I "tried" to do this on my own about 8 years ago
    before I even started working as an EE. I've learned quite a bit since
    then and I bet I could learn it on my own now. But I am never going to
    underestimate the difficulty of successful, non-accidental success of
    RF circuitry design. If you can do it, you are pretty much in the
    highest rung of the EE ladder, IMHO.

    The tricky part here is that I must be able to design the circuit from
    scratch to have certain specific parameters, and not just monkey copy
    something out of the ARRL book. Although that is an excellent
    reference.

    Heres a question for you:

    Whats the hardest part about doing what I mentioned above? Opinions?
     
  6. Guest

    Thanks for the excellent reply Joel! Don't worry, I am not affected by
    discouragers. If I was, I would probably never post here, since there
    is always someone who replies "dont do it that way, dont do it at all,
    etc.."
     
  7. Rich Grise

    Rich Grise Guest

    Yeah, pretty much. ;-)

    Cheers!
    Rich
     
  8. Joerg

    Joerg Guest

    A multi-multi ham contest. Meaning the dudes in the next tent were
    blasting along on an antenna about 200ft from the one I was using.

    Radio Tirana was another story. They blasted commie propaganda into
    Europe from Albania, probably using up all the electricity there while
    the population was starving. The usual :-(


    But when one reaches 70 or 80 and the bones get shaky one quickly falls
    off that rung again. Got a few decades, I guess ;-)


    Can't see your whole original post (Can you use some better domain that
    google.com?). But I don't see anything hard with a one-mile line of
    sight audio com link. Even well above 100MHz.

    And then the neighbor comes hollering that his super-cheap TV is falling
    over backwards.
     
  9. Guest

    For less then 100 mhz, I'd go with

    Experimental Methods in RF Design
    -- by Wes Hayward, W7ZOI, Rick Campbell, KK7B, and Bob Larkin, W7PUA

    Serious EE in that book as well as a lot of fun. My boss with 45 years
    in EE but no interest in ham radio stole my copy and wont give it
    back. For those sceptics out there, its not your normal ARRL/TAB
    cookiecutter cookbook. But its not Artech House either.

    Older over 100 mhz stuff,

    The RSGB microwave handbook. The 60s-70-80s one, not the new
    international one.

    Not published any more, but still out there new.

    Then get on the microwave mailing reflector and listen for a while.

    Steve
     
  10. Phil Allison

    Phil Allison Guest


    ** You will not likely get a job doing something you have no ability to do.

    Suggest you start off building a "crystal set", these are more interesting
    than you might think.

    Built my first when I was 7 or 8 years old.

    Then a one transistor radio, a one valve set tc.

    Get yourself a nice RF generator too - essential really.

    A frequency counter and spectrum analyser are also needed for working with
    transmitters.

    Also, and importantly - do not attempt to re-invent what others did long
    ago, instead study what has been made and sold successfully and learn from
    it.



    ...... Phil
     
  11. Joerg

    Joerg Guest

    Yes, but if you have repaired the umpteenth piece of electronics they
    all know who dunnit ;-)
     
  12. Jasen Betts

    Jasen Betts Guest

    starting, persisting, and finishing.

    some suggestions,

    1: pick a frequency

    2: If both ends are fixed locations, find a good design for directional
    antennas.
     
  13. Guest

    A good RF Engineer needs to understand the following topics (Each
    requiring a fair amount of study and work to understand):

    1. Matching concepts. (Why is matching concepts so critical to RF -
    must understand - it is a constant theme in RF)

    1a. Matching with lumped elemenets
    1b. Matching with distibutive elements

    1b1. Understanding how distributive (Transmission lines)
    elements work
    (First step - a quarter wave short is an open :)

    2. Noise, Noise bandwidths,
    (as an example, why can you send a signal around the world
    in
    morse code with 1 watt, and you need a KW for voice?)

    3. Amplifier system concepts , Noise figure, IM products, cascading
    noise
    figures

    4. Feedback control theory. Concepts of stability, Loop bandwidth and
    loop
    response as a function of loop bandwidth.
    You need to understand this to do Phase Lock Loops, and ALC
    circuits.

    5. Filtering concepts and filter designs. Need to understand Zverev
    book of filter
    tables and how to use it. Need to undersatand concepts of group
    delay and
    filter trade offs

    6. Need to understand , inside and out, how to use a network analyzer
    and a spectrum analyzer. Absolutelty must get access to these two
    instruments and really , understnd them. (Must play around with
    distributive elements on network analyzer)

    7. Perhaps most importand is grounding concepts. I would say that in
    my experience, 50% of all problems I have ever had ultimately boil
    down to a grounding problem. Even to this day, and I know this, I
    invariably chase down other issues before I chase grounding, and sure
    enough, its a grounding problem.


    These are your first steps, the circuit design concepts are not so
    hard, but you must be willing to really really bang your head against
    the wall to get your circuits to work. You must have tenacity.
     
  14. Joerg

    Joerg Guest

    And always keep a good analog scope. Always. If the OP doesn't yet have
    one I'd recommend the Tektronix 2465. Several of my clients followed
    that recommendation and got them for around $500 off Ebay. "Wow, it's
    like someone turned on the light!" was a common comment. Start every
    measurement job with the analog scope first and use a DSO only when the
    analog one really, really can't do the job. Like on data lines or very
    low frequency erratic noise.

    Sssssht! Don't take away my business base ... ;-)
    You forgot one for the guys building analog power circuitry:

    8. Always know where the next fire extinguisher is.
     
  15. Guest

    Good gear for RF is not expensive these days, but perhaps the OP would
    like to start cheap:
    Some useful toys:
    :
    The Poor Man's Spectrum Analysis kit:

    http://www.science-workshop.com/

    Ok, so its a slightly modified TV tuner, but for up to say 850 Mhz,
    its a decent vision of whats going on, and you learn about birdies and
    mixing and splatter and good RF construction practices. You build it
    once and then rebuild it in die cast boxes to clean up the birdies.

    You need a decent oscilloscope , say 10 mhz, solid state
    deflection , CRT , minimum with it.

    Then you read this paper and clean up the IF response and extend the
    range

    http://www.arrl.org/qex/Henkel.pdf

    I know there are better Spec An kits out there, but t 0-110 mhz
    doesn't get you much these days.

    I will also agree that a decent scope is the best starting point.

    A older varacter TV tuner (external pll) with the lid off is a
    wonderful place to start, even if your only other gear is a 99$
    optoelectronics frequency counter.

    I once asked Dr Wenzel to calibrate a diode for me, this paper was the
    result:

    http://www.techlib.com/files/detect.pdf

    Then one of these:
    http://www.aade.com/lcmeter.htm

    Then one of these:

    http://www.aade.com/dfd4.htm

    if you have no budget left after the 1N914 diode, the scope, and the
    spec an from the dead VCR, you go here:


    http://web.telia.com/~u85920178/

    Then some mini-circuits mmics and vcos.

    The pro EEs here might laugh at me for this, but starting at this
    level with the above stuff in college got me to 10 Ghz SSB phase
    locked to a GPS reference.

    If you have budget, a PTS160 from ebay is also a great bench tool

    Steve Roberts
     
  16. Mark1

    Mark1 Guest

  17. Guest

    Thanks, I needed that.

    One more good ham radio resource, but written with Attitude:

    Green Bay Packet Radio:

    http://www.qsl.net/n9zia/

    Steve
     
  18. Guest

    My experience is a crappy digital scope is better than a great digital
    scope when you are looking at analog signals in an RF system.

    An analog scope, as you point out, is often better than a digital
    scope.
    And # 8- 2: Now, go get yourself a decent digital scope for power
    supply design. Need to catch the one time transcients.

    Brent
     
  19. Guest

    I'm excited!! Thanks for all the hints and tips folks! I am sure there
    are many people like me who want to learn this and now all these
    guidelines will be immortalized.

    Im getting that Experimental methods in RF design book for sure,and I
    will look at the other ones too. I found a library with it and I will
    try to get the others too. Hope they dont mind if I keep it all for a
    while :)
     
  20. Tim Williams

    Tim Williams Guest

    Here's my understanding of it, I think, more or less possibly... wrote
    this a little while ago, here's the quote:

    Wideband stuff is pretty easy: resistors are, for the most part, well
    behaved, and with resistance squashing the more unpleasant reactances
    in the circuit, you're free to push the boundaries. In wideband
    circuits, that means reducing those resistances until the reactances
    just start to bite back, then tweaking the circuit (or layout, or...)
    until the waveform simply looks good.

    But in RF, you're intentionally tempting those parasitics with juicy
    LCs, lumped constants that you so wish to be ideal. And they can be
    pretty nice, with high Qs for high selectivity. But without that
    resistance, parasitics like lead inductance are free to party. You
    can try building an amplifier for one frequency, but if it works
    better as an oscillator at any other frequency, its amplification is
    pretty well doomed.

    So I think it's exactly the fault of putting in a tuned circuit (most
    likely the capacitor in particular) which creates all those horrible
    VHF-UHF+ oscillations that so often spoil RF work. I think it's
    valuable to have experience in wideband as well as tuned circuits,
    especially where stability and power converge. When you're building a
    wideband amp, it has to carry a lot of current, because it's class A
    and it's fast. RF power outputs carry a lot of current because
    they're outputs, and they're fast because that's the point, but what's
    more, they're fast well above and below the one frequency you need
    them, so they are both similar to, easier than and harder than a
    wideband circuit!

    Now, I haven't had much experience with tuned amplifiers, but I think
    I've gotten enough of a feel for these things that this might actually
    be right.

    But I still say RF is black magic. ;-)

    Tim
     
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

-