Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Crystal Osc 30MHz Design or Purchase

  1. May 14, 2013 #1
    Hello experts…i am new in the RF field and i need some help…
    So i have a double balanced mixer (ZAD-6+ of Mini Circuits) that need +7dBm Local OsC power and 30MHz Loca.osc frequency. At the time i use signal generator for l.o

    How can i make a crystal oscillator with these parameters? Can anyone tell me simple how to start? I can buy also some ready crystal osc ?

    Many thanks
  2. jcsd
  3. May 14, 2013 #2


    User Avatar
    Science Advisor

    Yes, you can buy crystal oscillators on EBay.

    You need the ones that have 4 or 5 connections. The 2 wire ones are crystals which you would need to build a circuit for. This needs some test equipment, so the ready-made ones are easier.

    These oscillators need 3 volts or 5 volts and this needs to be accurate. You can get regulators to supply these voltages.

    Also, these are computer oscillators so they are not very accurate in frequency.
    So only get them if an error of up to 1 kHz doesn't matter.

    The output may be a square wave, too, but a simple filter should turn this into a sinewave if this matters.
  4. May 15, 2013 #3
    Hi and many thanks for you reply! Can you please recommend me a crystal oscillator from Ebay for example to understand? Also:

    1) i want to feed a mixer so us i understand that i need a clear sine. Or maybe i can use and square wave?

    2) Also the mixer has coaxial port input SMA 50 ohm what about the oscillator's output?

    3)about the power?

    Last edited: May 15, 2013
  5. May 15, 2013 #4


    User Avatar
    Science Advisor

    I saw this one on Ebay:
    Works on 5 volts or 3 volts. 20 ppm so they could be up to 600 Hz off frequency at 30 MHz
    You get 10 of them so you can pick the most accurate if you like. Not bad for $15

    Output is TTL, so you would need a buffer to drive 50 ohms. So, maybe an emitter follower.

    Square wave drive is OK for diode mixers as it gets clipped anyway by the diodes.

    Just a caution, though, the output is similar to the supply voltage and the diode mixer is expecting about half a volt RMS or about 2 volts peak to peak.
    So you need to attenuate the oscillator signal.

    Diode mixers are lossy, though, and they can lose a lot more than the data sheet says if they are not set up properly. I have used them, but avoid them now.
  6. May 15, 2013 #5
    Make sure the phase noise of the oscillator you buy is low enough, since the mixer will up-convert it to RF.
  7. May 16, 2013 #6

    The 30Mhz filter is a crystal filter!Also the fc=10khz of the siganl maybe will go higher to 40khz.

    @vk6kro: To be honest, i need a 30.01MHz or 30.010 KHz crystal and i don't find it easy.Where i must search?

    Yes the square will be ok.Also i must make a emitter folower (is it difficult? ). About the attenuator with some voltage division i think i will be ok.

    @carlgrace: This is a big question. The datasheets of the mixers don't write about the frequency stability.i think 1khz max maybe is good
    Last edited: May 16, 2013
  8. May 16, 2013 #7


    User Avatar
    Science Advisor

    That looks like a very ambitious project.
    Do you have the test equipment to build it?

    You can get crystals made to order. You might have to build an oscillator for it and adjust the frequency.
    I checked Google and there are plenty of makers of crystals. Very hard to get prices online but they all give email quotes.

    You can also get DDS signal generators for about $50 on Ebay which you can set to any frequency up to 50 MHz or so to a precision of 1 Hz. The accuracy depends on an internal crystal which can be a bit inaccurate. So you may have to set the frequency on a LCD with an offset.
    If you do this, look for the LCD display in the picture. Cheaper DDS devices require you to do the programming.
  9. May 16, 2013 #8
    It is a hard task..and there is no more knowlegde to help me..can i have some of the crystal (oscillators right?) makers? I will send them e-mail...for the 30.010 Khz it is high accuracy and i don't believe i can find so much accurate and stable crystal. Maybe i use a third mixer to upconvert from 10khz to an IF higher and then to 30mhz..i am confused because first i must check what crystal and devices i can buy and then i implement them...

    About the DDS i didn't know that..i will check this.seems interesting...!
  10. May 18, 2013 #9
  11. May 19, 2013 #10


    User Avatar
    Science Advisor
    Gold Member
    2017 Award

    A 30.000MHz crystal will be good. you will be able to tune the oscillator over a reasonable range, at least ~ 0.1 to 0.2MHz anyway

    you cant use a 30 kHz crystal there's a huge difference between 30MHz and 30kHz

  12. May 19, 2013 #11


    User Avatar
    Science Advisor

    Oscillator modules can't be pulled at all unless they have a voltage control pin, and home made crystal oscillators for 30 MHz can only be pulled 1 KHz or so.
    10 KHz would not be possible and 200 KHz would be even less possible.

    That DDS module sounds good, Fovakis. There are cheaper ones, about $50, but this one is better with special sockets for square wave output and it includes a power supply.

    The 30.01 MHz output will be a sinewave and you may have to amplify it. I have seen MMIC chips used for this. These are wideband amplifiers which can give you increased output without any need for adjustment.

    There is a possible problem if the DDS unit uses a 30 MHz internal crystal. This may cause a spurious signal in your output. You could ask before you buy if this is the case or not.

    I have some doubts about that circuit. Can you really mix 30.01 MHz with 10 KHz and filter out the 30.02 MHz signal that is generated as well as the desired 30 MHz signal?
    I don't know if you can or not, but it seems that the filter would have to be very good to do this.
  13. May 19, 2013 #12


    User Avatar
    Science Advisor
    Gold Member
    2017 Award

    agreed, a specific module isnt going to be pulled and getting one made for 30.01MHz isnt going to be easy
    but a discrete component crystal osc, say a Colpitts type can easily be
    which is what I had in mind :)
    in the days of working on the old Tait commercial, xtal controlled, tranceivers, they could easily be tuned over a 100 kHz.

    this one here ...

    part of a transceiver that I worked on regularly over recent years. Up to 50kHz at 35MHz operating freq wasnt too difficult :)

    But I also agree with you DDS is a good way to go these days


    Attached Files:

    • osc.GIF
      File size:
      2.4 KB
  14. May 19, 2013 #13
    It was my mistake...i need 30,01 MHz or 30.010 KHz !!!
  15. May 19, 2013 #14
    Hi friends and good morning from Greece. Here is 12 the morning!

    I prefer to take a DDS because there is not a lot of time to make on my own and maybe i need to experiment with some frequencies not to have a fixed one. That crystal oscillator Colpitts sound perfect by the way!wow!

    About the DDS:

    1)Sine Wave Vpp: 0~5V adjustable (1k resistance load)
    MAX VOLTAGE IS: 5Vp-p! i want 5-7dBm so it covers me right? Why you say to amplify that?

    2)Sweep minimum frequency definition: >=1Hz

    So this device can give me also for example 10Hz ! Excellent accuracy!right?

    3)SMA COnnector is 50 ohm?

    4) i didn't understand that with the 30MHz crystall that might have internal.explain please?>

    5) yes...there is a huge problem with the mix of 10khz and 30.01MHz... i cannot find filter that suppresed the carrier tha LSB the 3IP etc...maybe i will use a 90 Phase Shift after the DSP to "create" I and Q components and then a I&Q-Modulator ... i don't know if this can succeed
    Last edited: May 19, 2013
  16. May 19, 2013 #15


    User Avatar
    Science Advisor

    DDS signal generators sometimes use AD9851 chips which use a 30 MHz crystal oscillator then multiply this by 6 internally, then use this to derive all the output frequencies.

    So having a powerful 30 MHz oscillator nearby could be a problem if you are using 30.01 MHz output.

    The output may be 5 volts with a 1000 ohm load, but it may be only a few mV with a 50 ohm load.

    Look for data about MMICs. A typical one is the MAR-6.

    Yes, I think you will have a filter problem.
  17. May 19, 2013 #16
    I will ask them if they use 30MHz crystal oscillator...

    Ok the filter is a problem so...About the 90 phase shift after the dsp's output and then I&QModulator , what is your opinion about this?
  18. May 19, 2013 #17
    Ok...i see some 3dB splitters 90 degrees but i cannot find for low frequency like 10khz...
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook