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: L-MFJ
ad: abrind-2
ad: Subscribe
ad: Left-3
ad: Left-2
  1. K4KYV

    K4KYV Premium Subscriber Volunteer Moderator QRZ Page

    Looks like the lows are sharply cut off at 200~ and below. In the second plot, the lows appear to cut off sharply at 400~. It would have less space-shuttle/EAS sound if the lows extended down flat to around 100~ and then gradually rolled off.

    Here is the built-in frequency response of the pre-amp I use with my unamplified D-104 (scroll down mid-page on the attachment). The acoustical bump in the D-104 response adds to the presence rise built into the pre-amp. The low end is flat to below 40~, but the microphone frequency response rolls off at the extreme low-end.

    The low-frequency roll-off of the D-104 is less than that shown in the curve, when the mic works into a high impedance load on the order of 10 megohms or more. Mine works into 20 megohms; low-frequency rumble from distant aircraft registers on my monitor scope before I can actually hear it with my ears, but I have never had a signal report that I was 'bassy'.
     

    Attached Files:

  2. AC0OB

    AC0OB Platinum Subscriber Platinum Subscriber QRZ Page

    "...Since modern IF sections are so unforgiving of high-frequency shelf boost, let’s see just how we can get more highs without the shelf. The NRSC curve has a boost of 5 dB at 5 kHz so let’s improve on that with a boost of 9 dB at 4 kHz. But instead of the Orban shelving method let’s make it a bell curve boost and allow it to roll off back to 0 dB as it approaches 10 kHz. I did not alter the NRSC cutoff filter so as not to introduce any complications that would muddy the results.

    radioworld@futurenet.com with “Letter to the Editor” in the subject field..."

    Actually, the NRSC boost he speaks of is only 5 dB with respect to 1kHz. It is almost 7 dB with respect to 50 Hz.

    It appears he too wanted a bell curve as well but he kept the NRSC system ON during his experiment, which meant that there was a HF limiter engaged, necessary so that HF's don't over modulate the transmitter.
     

    Attached Files:

  3. AC0OB

    AC0OB Platinum Subscriber Platinum Subscriber QRZ Page

    Yes I saw that circuitry in ER some time ago and I thought the differential input circuit idea was a good one.

    I agree that for improved articulation a boost of from 3dB to 6 dB is needed.


    The curves I presented were not detailed enough to show the circuits' LF response.The curves were for AM BC service, not Amateur service.

    The time constants in the circuit proper allow audio reproduction from 20 Hz to about 15 kHz with sharp rolloff at 9.5 to 9.7 kHz, the rolloff most broadcast AGC/Limiters use.

    It's my view that an un-modified and properly loaded D-104 needs very little help in terms of boost and cut.

    The issue in most vintage Amateur transmitters is their audio response is not flat, and were not designed to be so.

    One has to be careful when using commercial audio processors to bandaid the audio response in an unmodified vintage Amateur transmitter. I prefer to first go into the vintage transmitter circuitry and flatten the response and then use an external AGC/Limter.

    For many of the older BC tube transmitters from the sixties, they were flat from 50 Hz to at least 12.5 kHz.


    Pheel
     
    Last edited: Jan 20, 2019
  4. AC0OB

    AC0OB Platinum Subscriber Platinum Subscriber QRZ Page

    OK, here's what I found:

    Inovonics 222 Pre-Emp ON (18uS TC), Limiter, ON.

    An audible improvement can be had with an extra 3 dB pre-emphasis starting at about 1.2 kHz with a 3 dB post emphasis at 12.5 kHz.

    Starting at 2.5 kHz with a 3dB extra pre-emphasis is starting a bit late. I suspect however, that if one had a 2.5 kHz starting pre-emphasis with an extra 7.5 dB or so of boost and a 6dB rolloff at 9.75 kHz, the window might be tighter and more noticeable.

    The "naked" response of this circuit is essentially flat from 30 Hz to over 50kHz without the filtering, of course.


    Pheel
     

    Attached Files:

    Last edited: Jan 23, 2019
  5. K5UJ

    K5UJ Ham Member QRZ Page

    When I ran a 222 I drove it to around 12 dB gain reduction. It had a lot of punch but it didn't sound all that great. -3 dB is probably about right. Beyond that, and sound quality starts to suffer. This isn't to dump on Inovonics; the box was intended to be an inexpensive way to have NRSC compliance for stations running older processors.

    The NRSC premphasis curve or standard is 75 microseconds. I need to have a guru explain to me how a response curve is expressed as a time constant. I've read about it but it doesn't sink in. I need to have someone dumb it down for me.
     
  6. N1BCG

    N1BCG Ham Member QRZ Page

    A parallel R/C circuit results in both a low frequency cut (but we call it a high end boost) and a time constant. The time constant is a convenient way to express the combination of R and C values that produce the curve. A loading resistance determines the circuit loss (high end boost) as it creates a frequency dependent voltage divider.

    For example, a 75uS time constant results from a parallel 75k R and a 0.001uf C (75,000 x 0.001). A useful presence boost for AM is 150uS (10k R and 0.01uf C) because the curve starts to head up earlier, yielding a greater boost in the presence frequencies than a 75uS curve.

    A time constant is much easier to express than to say "it has a boost of XdB at YkHz, and then a boost of Z at..."
     
  7. K5UJ

    K5UJ Ham Member QRZ Page

    True, however it is impossible for me to instantly do the mental math calculation with R and C, and translate that into "x dB at Y kc." All you broadcast engineers must be walking calculators and circuit analyzers to instantly leap from N microseconds to a frequency response curve.

    Quick, what's the response curve of 100 microseconds? No cheating, answer up right now.
     
  8. AA5CT

    AA5CT Ham Member QRZ Page

    At what looks like the -30 dB point!

    3 dB audio BW looks more like 2.6 kHz.
     
    Last edited: Jan 27, 2019
  9. N1BCG

    N1BCG Ham Member QRZ Page

    LOL, no... it's more of a generalization. You know that the shorter TC will yield a curve that begins higher in frequency. 75uS is really the only one we can visualize *, but mostly from FM b'cast.

    Do it the other way around. Use a cap and a potentiometer to come up with the desired curve then document the result in uS. A time constant makes it easier for others to replicate the results. Did I allude to the importance of the loading resistance? OK, good.

    * does not apply to Pheel
     
    Last edited: Jan 27, 2019
  10. AC0OB

    AC0OB Platinum Subscriber Platinum Subscriber QRZ Page

    For broadcast I usually use about 4-6 db of compression in the Inovonics 222.

    The Inovonics 222 uses an 18 uS preemphasis RC circuit consisting of a series .001 uF and 18k combo to get the "breakpoints" in the NRSC curve of 2122 and 8700 kHz. Of course, the Inovonics 222 also has HF limiting as well.


    Pheel
     

Share This Page

ad: cq2k-1