There are a few items I’d wish to clarify earlier in the string, and make a recommendation and offer.
There was an overshoot issue on our early digital i/o’s. Specifically, prior to 2000, the Digit digital i/o had an aggressive 15khz roll-off filter for pilot protection, and early Omnia’s had a harmonic content from clipping that passed through their very wide band output. Omnia’s peak management, when used with the sharp cutoff filters in the Digit’s AES3 input, resulted in peak overmodulation, again, an issue with their peak control.
At that time, 3/16/2000, Harris opened up the filter in the DSP, U10, a socketed IC in the Digital i/o, Digit part number 917-2335-036, DigitCD part number 917-2335-042, to 17.5 khz, providing a wider input window with part number 917-2335-050, to make it compatible with Omnia audio processors. This information has been available in the Harris service department by simply calling 12172228200 or via email at
[email protected].
Since that date, the bandwidth limiting of the audio input to 15 khz needs to be accomplished in the audio processor, to protect the stereo pilot at 19 khz .
The performance of the DigitCD stereo generator is typical of a Digital Signal Processor and does use a single upconversion of the direct digital modulator to the FM carrier frequency, i.e. -94 dB mono/composite SNR, -83 dB STEREO SNR, -65 dB audio separation, -60 dB AM sync noise floor, and 0.005% stereo THD, setting the global standard for many years as the most popular exciter ever built. However, while the stereo generator has excellent performance, the clipper should not be used to increase loudness, and only is an absolute modulation limiter, as the tradeoff between loudness and IMD is poor. Such compression is better achieved through use of an external stereo generator, composite output, and composite clipper; which makes the analog composite input version of the i/o module a better tradeoff, as it has higher performance than the digital i/o in analog backup mode. As for the subject of use of 32 khz AES audio input, the maximum frequency audio output is half the sample rate, or in other words, 32 khz AES is adequate for audio frequencies up to 16 khz, which is adequate for analog FM audio, which must be limited to 15 khz.
In the case of using a higher AES audio bitrate, for example single AES sourcing of both an HD and analog exciter, the Digit/DigitCD digital i/o down converts to 32 khz.
In HD Radio compatible exciters such as Flexstar, the HD Radio AES3 sampling rate is separate and handled by the Importer/Exporter and has nothing to do with the AES3 sampling rate for analog stereo.
As for Flexstar exciters, this is a direct to channel FM exciter, with yet even higher audio performance, for example Stereo SNR at -85 dB. Flexstar exciters have an audio low pass filter that is selectable via the front panel between bypass, 15khz, and 17khz. Both provide excellent pilot protection, however the 17 khz filter is wider. The Flexstar also has a digital composite limiter with adjustable level and attack time.
When the Flexstar was first introduced, the default setting was bypass, and there were a couple reports of pilot stability and or interference, resolved with proper filter settings.
My advice, if you decide to purchase a Flexstar, go with the up charge and order one with the optional low level 10 mw output. This will allow you to drive a second cabinet from this exciter, should you later have a power increase and add another amplifier cabinet or implement HD Radio.
In my personal opinion, having factory and field tested and commissioned nearly 200-300 Digit exciters alone, prior to direct to channel exciters today, at the hayday of the Digit/DigitCD, head to head with the Continental exciter, Continental’s performance was only just below Digit but comparable (so was the price at that time), DigitCD set the industry standard for performance for many years, the Flexstar sounds significantly better than even the DigitCD. We offer a free demo, should you wish to try it first.
