In an ongoing testing effort to get the highest speeds from VARA-FM, I have been examining the characteristics of the "9600 baud" 6-pin mini-din input on my Icom-7100. Although there is only one input pin, the treatment of the incoming signal can be switched to wideband via a menu setting.

My first experiment has been to try the Bessel zero method of determining deviation. My first experiment was to generate a sine wave at 1250 Hz to find the amplitude of the signal that would result in 3kHz deviation. The result was that the radio went into TX fault (flashing red) and stopped transmitting, due to overdriving the input signal. The carrier amplitude had not reach a null.

The next experiment was to set the amplitude of the signal to 500mVpp as specified in the manual and increase the frequency from 200Hz until I saw the Bessel Zero (minimum carrier amplitude). To my surprise, this occurred at only 227Hz, indicating a deviation of only 548 Hz (227 x 2.41).

The same experiment at 300mVpp saw the Bessel zero at about 250Hz.

I'm reasonably confident that my measurements are within 10%, although I've seen 60Hz sidebands on the carrier and audio sidebands about 30dB down. I have tested 5800Hz at the same amplitude and see the carrier, the J1 sidebands at +/- 5800Hz from the carrier, and very weak J2 sidebands at +/- 11,600 Hz, as I would have expected.

My question is that is this a reasonable result? Is the deviation really that low or should I be looking further for faults in the setup or method?

Setup: Leader AF generator. Audio voltage measured on Tek 'scope. Transmitting into dummy load next to Elecraft K3 with 2m transverter and no antenna. IF-out to SoftRock-IF plugged into Steinberg UR22C audio interface. Viewing signal with HDSDR.

73, Chris VE3NRT

  • $\begingroup$ I woke up this morning realizing that I made an error. There were multiple sidebands present, so the Bessel zero I observed in the carrier was not the first one that shows up with a modulation index of 2.41. I will rerun the experiment and gradually lower the frequency from 1250 Hz instead of increasing it from 200Hz. $\endgroup$
    – Deepstop
    Mar 20 at 12:45

1 Answer 1


The answer: It is an unreasonable result.

The Bessel zero I was reading, somewhat inaccurately as subsequent readings were about 192Hz, was the 3rd zero occurring at a modulation index of 8.65, not the first zero with an index of 2.41. To get a better reading I set the oscillator to 2kHz and brought the amplitude up to the maximum that the IC-7100 documentation permits. The carrier amplitude decreased but not to zero. Next the audio frequency was reduced until I had a good reading on the minimum carrier amplitude at 697 Hz. Further decreasing the frequency revealed a second Bessel zero at 333 Hz and a third at 192.3 Hz. From the [table in Wikipedia][1] which relates modulation index to the amplitude of the carrier and the sidebands, this yields:

  1. The first zero at 697 with a modulation index of 2.41 indicates the deviation to be 1680 Hz.
  2. The second zero at 333 with index 5.53 yields 1663Hz
  3. The third zero at 193.3 Hz with index 8.65 yields 1676 Hz.

As my audio oscillator frequency isn't all that accurate and I'm using an oscilloscope to measure both frequency and amplitude, I'm quite pleased with how close the 3 results are.

Based on my understanding that VARA-FM Wide has an audio bandwidth of 5800 Hz (that might mean the upper limit is 6KHz), applying Carson's rule puts the FM bandwidth at 15-15.6 kHz, which sounds about right.

Next adventure is to figure out what voltage VARA-FM is sending to the radio.

I hope this helps someone!

73, Chris [1]: https://en.wikipedia.org/wiki/Frequency_modulation


You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .