Need to Measure Transmitter Spurs, w/o Spectrum Analyzer

Discussion in 'General Technical Questions and Answers' started by KK5JY, Oct 19, 2017.

ad: L-HROutlet
ad: l-rl
ad: Left-2
ad: abrind-2
ad: Subscribe
ad: Left-3
ad: L-MFJ
  1. KA9JLM

    KA9JLM Ham Member QRZ Page

    A cheap SRD can give you a comparison clue, But is far from calibrated.

    Please do not trash our bands, There is enough trash on them already. ;)

    Have Fun.
     
    KK5JY likes this.
  2. K6CLS

    K6CLS Ham Member QRZ Page

    Instead of feeding the SDR off the Tee, use a nice air gap directional coupler - any piece of wire along the coax.

    Or make a directional coupler, several turns of magnet wire around a ferrite, coax core goes through the ferrite.

    Enjoy
     
  3. WB2WIK

    WB2WIK Platinum Subscriber Platinum Subscriber QRZ Page

    Problem with any kind of coupling that isn't purely resistive is it will have its own frequency response.

    If the coupler "works great" at say, 3-30 MHz, and covering one decade with flat response is quite a trick, it may not pick up spurs at 137 kHz or harmonics at 300 MHz, at least not with flat enough response to provide any useful "-dBc" information.
     
  4. K6CLS

    K6CLS Ham Member QRZ Page

    Steve, good point, that's why we pay the money for good test equipment. The conceit here was $cheap, I thought.
     
  5. KK5JY

    KK5JY Ham Member QRZ Page

    That's what I was hoping to accomplish, if I can just get the SDR and transmitter decoupled from each other's near field (or whatever mode it is that is bringing energy from the transmitter into the SDR even when the SDR has nothing but a termination resistor across the antenna jack). If the fundamental shows at -20dB on the waterfall, then I need to get the harmonics below -63dB to at least be in the ballpark, or so went my thought process.
     
  6. WB2WIK

    WB2WIK Platinum Subscriber Platinum Subscriber QRZ Page

    Good thought process but sneak paths can render readings pretty meaningless.

    The sneak path likely has its own frequency response. You really want something dead-flat.

    If you install the SDR in a very tightly RF shielded enclosure with feedthrough filters on the USB lines (power and signal, it's only four pins) and a solid well-shielded RF input connector (TNC, N, SMA), assuming the SDR has sufficient dynamic range, you could probably get some pretty good results.

    When I was with both HP and Boonton Electronics back in the 70s (engineering teams) we developed both RF signal generators and spectrum analyzers and everything was double-shielded with pierced copper gasketing on all metal-to-metal interfaces, and LC RF shielded feedthough filters on all lines not intentionally carrying RF and considered "140 dB isolation" just about good enough.:p

    One reason the "lab" gear costs so much is all that machining, gasketing, filtering, decoupling, etc. Half the cost of any of these products was just that.
     
    WA7PRC and KK5JY like this.
  7. KK5JY

    KK5JY Ham Member QRZ Page

    The real shame is that 20y ago, I knew several people who had access to gear like that, and I could have borrowed it or visited them to use it without any challenge at all. Nowadays, I only know one guy who is in LMR, and I'm pretty sure that everybody at all the shops that he knows about would say, "spectrum ana-what?" :oops:

    Edit: it seems like the future of kit-building would kind of depend on the ability to do these kinds of measurements without having to invest in five or six figures worth of diagnostic gear. I guess I could just trust that the filter design is solid and has the frequency curves that the original articles say, but that doesn't seem very thorough. The old FCC rule for 5W transmitters (-30dBc) was loose enough that I might just light-er-up and see what happened. -43dBc seems to be a little tight for that kind of experimentation.
     
  8. SM0AOM

    SM0AOM Ham Member QRZ Page

    There is a "trick" around from the times when spectrum analysers were non-existent or prohibitively expensive, which
    consists of having an elliptic high-pass filter between the dummy load and a terminated RF millivoltmeter.

    The high-pass filter should suppress the fundamental by at least 60 dB, so the sum of the out-of-band spurious or harmonic components dominate above the suppressed fundamental. The ratio between the fundamental voltage before the filter and the voltage after the filter is an appoximation of the worst-case harmonic suppression. If you have, say, 50 V of fundamental carrier and measure 100 mV after the filter, the suppression will be at least 20*log(0.1/50) or 54 dB.

    73/
    Karl-Arne
    SM0AOM
     
    KO4LZ, WA7PRC and KK5JY like this.
  9. WB2WIK

    WB2WIK Platinum Subscriber Platinum Subscriber QRZ Page

    That would assume all spurious emissions are above the desired carrier frequency.

    In the case of rigs using transmitting mixers, spurs can be anywhere, including well below carrier frequency.
     
    WA7PRC likes this.
  10. SM0AOM

    SM0AOM Ham Member QRZ Page

    This is correct.

    The method predates the wide-spread use of heterodyning-based transmitters,
    and I first saw it in a 1950's German textbook about verifying the harmonic suppression in
    HF and low-band VHF mobile transmitters.

    A related method, that requires a good SWR meter and a low-pass filter, is to measure the reflection into a wideband dummy load
    with and without the low-pass filter, whose insertion SWR at the fundamental frequency with a "clean" signal should be known.

    If the reflection rises appreciably above this value, you have stronger parasitics, spuriouses or harmonics than permitted.

    73/
    Karl-Arne
    SM0AOM
     

Share This Page