Quadrature Sampling Exciter Experiment

Discussion in 'Homebrew and Kit Projects' started by N5HXR, May 3, 2021 at 4:38 AM.

ad: L-HROutlet
ad: l-rl
ad: L-MFJ
ad: abrind-2
ad: Subscribe
ad: Left-2
ad: Left-3
  1. N5HXR

    N5HXR Ham Member QRZ Page

    Today, I put together a quadrature sampling exciter (I guess that's what they're called? A Tayloe detector in reverse?) -- it's mostly a clone of the Ensemble Softrock RXTX, though a few changes for what I have on hand. It's just two pairs of inverting unity gain amplifiers into an SN74CBT3253 multiplexer. I attached my little signal generator to it, with two signals 90-degrees apart at 2KHz, and put a 40m low pass filter in front of it all.

    [​IMG]

    My oscilloscope's FFT function varies a lot, though if I turn on acquisition averaging I get a pretty stable result shown below. LO set for 7.050 MHz, so the LSB signal is at 7.048 and the unwanted sideband signal would be at 7.052. Is this an appropriate way to make measurements given that I don't have a fancy spectrum analyzer?

    [​IMG]

    I can definitely hear the difference between the two sidebands with my FT-818 nearby. I found that if I increase the amplitude at the signal generator, eventually the peaks that look like intermodulation spike up, and everything goes pretty bad :).

    Is something like this suitable for SSB transmission in a homebrew transceiver? What should I be thinking about if I pursue a QSE like this instead of the other common designs?

    Thanks for any thoughts you might have!
     
    KE5OFJ likes this.
  2. K7JEM

    K7JEM Ham Member QRZ Page

    It looks like a "phasing SSB exciter" (but modernized and improved). Here is a link to old time application:

    http://www.ce-multiphase.com/alignment-phasing.pdf
     
    N5HXR likes this.
  3. N5HXR

    N5HXR Ham Member QRZ Page

    That's a neat document! Lots of related content, though some of those schematics are... daunting! It'll take me some time to digest it. Thanks!
     
  4. K7JEM

    K7JEM Ham Member QRZ Page

    I was looking at the schematic, and shouldn't the RF output be coming from pin 9 and not pin 7 of the 74CBT3253?
     
  5. N5HXR

    N5HXR Ham Member QRZ Page

    Good eye -- that's the mistake I had to fix after assembly. There's always one...

    20210502_222638.jpg

    OK, in this case there are two -- an end mill broke cutting the board and took out a divot in the board on the far right. Minor damage that was easy to work around.
     
    Last edited: May 3, 2021 at 3:52 PM
    KE5OFJ likes this.
  6. N5HXR

    N5HXR Ham Member QRZ Page

    After sleeping for awhile, I realized maybe the FFT output is so wobbly because the output level is so low. The scope doesn't do so well with microvolts... so maybe I need to put a little amplifier on the output of this thing to get a closer look at how it's working. I guess turning on averaging generally increases the dynamic range, so I suppose what I'm seeing is reasonably accurate (as far as measuring with the "wrong tool" goes...). But maybe adding some gain will give me a better picture.
     
  7. K7JEM

    K7JEM Ham Member QRZ Page

    That's probably a good idea. Also remember that those OP-amps really don't like being loaded to 50 ohms, that may be part of the issue.
     
    N5HXR likes this.
  8. N5HXR

    N5HXR Ham Member QRZ Page

    Good point about the loading... but if I'm not terminating it into a load at the moment, then it's just terminated by the 1M input impedance of the scope, right?
     
  9. K7JEM

    K7JEM Ham Member QRZ Page

    Right. But when you tie it to a 50 ohm load, things will probably change a lot. You probably need some sort of high impedance to low impedance matching network, or a buffer amp with a few K ohm in, and 50 ohms out.

    The spectrum analysis looks pretty good to me, the carrier and opposite sideband are reduced by more than 50dB, which should be enough.

    Is the SI5351 providing 7MHz quadrature signals? I think that is what's happening.

    Also, I think you could jumper out capacitors C8,9,10, 11 and remove L2. The op amps should supply 2.5v to the input of the multiplexer, and since one input is always on, you don't need any bias at the output, it will always be at 2.5v.
     
    Last edited: May 4, 2021 at 1:19 AM
    N5HXR likes this.
  10. N5HXR

    N5HXR Ham Member QRZ Page

    I'm using the Si5351 with two outputs in quadrature to switch the multiplexer, yeah. The AF IQ input is currently just my signal generator. But someone got me playing with the Iowa Hills filter design software, and I was experimenting with the Hilbert filters in GNURadio, and hope to translate that to the microcontroller for a future test.

    Thanks for those thoughts -- I was thinking similarly, but went conservatively with the same arrangement as the Ensemble SoftRock TX chain. I can test that easily enough with mods to this board. Thanks for the suggestion.
     

Share This Page