A Station Owner Tinkers With an AM's Response Curve Reviving Vintage Processors

Discussion in 'Amplitude Modulation' started by K5UJ, Jan 15, 2019.

ad: L-HROutlet
ad: l-rl
ad: abrind-2
ad: Subscribe
ad: L-MFJ
ad: Left-2
ad: Left-3
  1. W7TFO

    W7TFO Platinum Subscriber Platinum Subscriber QRZ Page

    I think Phil just luvs dat 222.

    73DG
     
  2. WZ5Q

    WZ5Q Ham Member QRZ Page

    Great Unit for hamateur use. I have 3 of them each modified for a different Low Pass Filter Cutoff Frequency, 3.6Kc, 5Kc, & 7.2Kc. Each of these are then selectable with a mixer/splitter to change the transmitted bandwidth on the fly. Once the elliptical filters and compensation circuit are re-tuned, these units work FB. Depending on how old the units are, a recap of the 'lytics helps a lot on the digital noise superimposed on the PWM Timing Pulse.
     
    Last edited: Jan 30, 2019
  3. K5UJ

    K5UJ Ham Member QRZ Page

  4. AC0OB

    AC0OB Platinum Subscriber Platinum Subscriber QRZ Page

    Here is the way I simulate and test pre- and post-emphasis circuits for verification and the circuit that sounded best:

    If anyone would like the MatLab code I will gladly send it.


    Pheel
     

    Attached Files:

  5. AC0OB

    AC0OB Platinum Subscriber Platinum Subscriber QRZ Page

    Great as backup units for stations as well when their $10k+ processors take a nose dive.

    This one had about 5 mV p-p of ripple on the +, - 15V and about 0.5 mV p-p of ripple on the +, - 6V.

    The only digital waveforms are created by the PWM system. The third order filters after the analog switches (CD4066) filter out the majority of the 152 kHz switching transients.


    Hint: For transparent AM BC audio turn off the Low Pass Function and let the 222 function as an NRSC/AGC/Limiter box.


    Pheel





     
  6. WZ5Q

    WZ5Q Ham Member QRZ Page

    I miss spoke, I meant to say diode noise in lieu of digital noise. The 6.2v zener's were spewing noise into the +/- 6v power rails and I had a lot of ripple on the +/- 15v rails. The result was seeing unacceptable noise on the 152Kc clock pulse from the multivibrator, on the clock ramp waveform, as well as the PWM Waveform from the AND Gates. Once all the 'lytics were changed, everything cleaned up nicely.
     
  7. AC0OB

    AC0OB Platinum Subscriber Platinum Subscriber QRZ Page

    I may go ahead and update the PS caps since they appear to be originals.

    One of the things that I question are the passive RC filters on the front end, one being 12.6kHz@-3 dB and the second
    being 4.05kHz@-3 dB. With the subsequent active filters up to fourth order, it seems redundant.

    Maybe they were trying for a composite -4.5dB@ 9.7kHz rolloff.


    Pheel
     
  8. WZ5Q

    WZ5Q Ham Member QRZ Page

    I calculate that passive 2nd order RC Filter at a fc of 17Kc which would be about right as I believe they are using this as an anti-aliasing filter.
     
  9. AA5CT

    AA5CT Ham Member QRZ Page

    Quick rules-of-thumb:

    75 us - wideband broadcast FM, pre-emphasis/de-emphasis breakpoint ~ roughly 2 kHz

    750 us - (NB) commercial 2-way FM*, pre-emphasis/de-emphasis breakpoint ~ roughly 200 Hz

    .

    .
    *EIA/TIA 603

    .
     
  10. AC0OB

    AC0OB Platinum Subscriber Platinum Subscriber QRZ Page

    It's a bit more complex than that for AM's 75 us curve since it is a modified or "truncated" 75 us curve.

    The first breakpoint occurs at the 2122 Hz zero and the second breakpoint is at the 8700 Hz pole. The breakpoints are the main "deflection" points in the NRSC curve and very suttle.

    In any processor that has NRSC capabilities one must protect from this preemphasized audio by using a sliding and variable "rollover" circuit (adaptive preemphasis) to limit (duck) that high frequency audio.


    Pheel
     

    Attached Files:

Share This Page

ad: M2Ant-1