Oscillator design (garbage band)

Discussion in 'General Electronics Chat' started by etl, Mar 28, 2006.

  1. etl

    Thread Starter New Member

    Mar 28, 2006
    2
    0
    Hello
    I will implement a oscillator for a wireless medical application. The frequency will be around 8-9MHz, and the signals should be as close to sine as possilble.
    I have never designed a oscillator before, so I am quite a newbie. After looking around a little, I found out that I could implement a crystal together with a CMOS inverter in a Pierce-oscillator, instead of building the oscillator with discrete components. This would save me a lot of time and testing.

    Does anyone have any suggestions or comments to this and suggestions to parts to use, I would be grateful.
     
  2. thingmaker3

    Retired Moderator

    May 16, 2005
    5,072
    6
    One might also try a crystal resonator at 2x desired f, a flip-flop to get 50% duty cycle at f, and an active filter to approach sine wave.
     
  3. Mazaag

    Senior Member

    Oct 23, 2004
    255
    0
    thingmaker3 could you please ellaborate on that? I mean , I have a crystal aswell as a ceramic resonator, but I too have no clue how I could use it to design an oscillator (sine wave) .

    thanks
     
  4. Papabravo

    Expert

    Feb 24, 2006
    9,905
    1,723
    The oscillator is only the beginning of your problems. Although you have specified the frequency of the oscillator you have said nothing about the rest of the application. Depending on the modulation method, power requirements, operational distance, and so forth you may have substantial difficulty in getting things to work.

    For example what size antenna will you need at 8-9 MHz. to effectively radiate any appreciable power. If you were planning on a random length of wire you shold be prepared for a rather large return loss. A large return loss can damage your output stage; not a good thing. On the other hand if you have an efficient radiator you can expect your signal to propagate several thousand miles with only five watts or so. It all depends on the sunspot activity. A half wavelength at this frequency is about 18 meters. A quarterwave whip antenna would be about 9 meters. Clearly impractical.

    I think there is much about your proposal of which you are unaware.
     
  5. etl

    Thread Starter New Member

    Mar 28, 2006
    2
    0
    Thank you for your reply.
    Both digitalised data(from an implanted sensor) and power-transmission will be done by an inductive link of two flatshaped coils. By using two tuned resonancefilters on either side, the transmission of power will be optimised when the filters are tuned to the same frequency as the sine-signal. (The coupling-factor of the two coils will also influence the power transfer efficiency). The operational distance is within a few centimetres (you hold your application to your head just like a mobile-phone)
    The modulation method is based on socalled Load-Shift Keying, which means that a change in the load resistance in the implanted circuit is reflected in the voltage on the outer. The critical point in the inner circuit is the power consumption, but this circuit is designed to consume very little power and the test results look good.
    I therefore, as far as I can see, just need a reference sinesignal for the outer circuit.
     
  6. Papabravo

    Expert

    Feb 24, 2006
    9,905
    1,723
    I read some papers on Load Shift Keying and I have some observations.

    1. Any crystal oscillator that you build will have a power output measured in milliwats, because crystals are delicate, and will fracture at high power levels. Many practioners use a Class E power amplifier following the oscillator, to drive the outer coil. Further reading reveals that I would have trouble designing such a class E power amplifier.

    2. Let's do a back of the envelope power budget. If the inner electronics runs at 3.3 VDC and requires 10 mA of current that's still 33 milliwats, which given some loss in the voltage regulator means more like 50 milliwatts from the coil, and maybe 60 to 75 milliwatts from the power transmitter, so you will need an amplifier on the output of the reference oscillator.

    3. I understand how the switch on the implant side creates a load transient, but I'm less clear on how that can be demodulated.

    4. Most of the authors provide little more than a block diagram so it's hard to devine the inner workings of their circuits, but it looks kind of interesting.

    5. Nobody has a canned circuit for the 8-9MHz. range.
     
  7. thingmaker3

    Retired Moderator

    May 16, 2005
    5,072
    6
    Nor do I. That's why I suggsted following with an active filter. :D
     
  8. CoulombMagician

    Active Member

    Jan 10, 2006
    37
    0
    Considering that crystal oscillators are less than $1 USD each your time may be better spent on the other unique parts of your product. Just a thought.
     
  9. n9xv

    Senior Member

    Jan 18, 2005
    329
    1
    Whats meant by "garbage Band"?
     
  10. Papabravo

    Expert

    Feb 24, 2006
    9,905
    1,723
    I take it to mean that he thinks there are no meaningful transmissions in that band, and there will be nobody there to complain about interference. I've tried to warn folks about this kind of thinking before. This time I'll just wait for the FCC enforcement letter to show up. If you are foolish enough to blow off this first letter you'll be thrilled to meet Mr. Riley Hollingsworth.
     
  11. thingmaker3

    Retired Moderator

    May 16, 2005
    5,072
    6
    Back in the mid '80s, the FCC actually did create some deregulated bands - popularly known as "garbage" bands. The term is a misnomer, of course. I don't know if 8-9MHz is one of these or not. Microwave oven frequency (some specific GHz) is one of them. I think WiFi runs on "garbage" bands as well, but I am probably mistaken. The official term is "MSI" Band or "MIS" Band or some such. A little googling might dredge up the info for those interested enough to spend the time.
     
  12. Mazaag

    Senior Member

    Oct 23, 2004
    255
    0
    How does the FCC know that you're little amateur radio is trasmitting on an illegal frequency band? How do they track you down basically ?
     
  13. Mazaag

    Senior Member

    Oct 23, 2004
    255
    0
    How does the FCC know that you're little amateur radio is trasmitting on an illegal frequency band? How do they track you down basically ?
     
  14. Papabravo

    Expert

    Feb 24, 2006
    9,905
    1,723
    Interference complaints are number one, and the official observers are number two
     
  15. n9xv

    Senior Member

    Jan 18, 2005
    329
    1
    The only "garbage" band I'm aware of exist on 27-MHz CB band. There's plenty of important stuff going on in the 8 to 9-MHz range. The FAA utilizes some HF frequencies in that range for transoceanic flights. A major military tactical frequency is on 9.023-MHz.

    The FCC has countless monitoring stations in the HF arena for this very purpose. They use various techniques to "track" a signal. Triangulation, signal strength and trigonometry of signal reflection/refraction off the ionisphere etc.

    In the world of Amatuer Radio, a few hundered milliwatts of CW (morse code) in the HF range can propogate around the world. Even voice communication can be surprising at around 5-watts or so under the best of atmospheric conditions.

    As liscensed Amatuer Radio operators, we have privlages from 1.8-MHz up through 300-GHz. In the upper GHz range, the govenments attitude is - you build it, let us know about it so we can improve it.

    Kevin
     
  16. etl

    Thread Starter New Member

    Mar 28, 2006
    2
    0
     
Loading...