Batboard

Does anyone know where I can find the detailed theory of operation of the Spectra-Tac Comparator?

Thanks,
Paul

If you go to http://www.jps.com, you can download the manual for the JPS voting system. This is a different piece of equipment, but the manual isn't a bad tutorial on the theory of how voting systems work.

Very (perhaps too much so) briefly:

In a voting system you connnect multiple receivers to the comparator, which is then connected to the transmitter. The comparator isolates a narrow AF band from the recovered unfiltered audio (usually above 4K and below 6K) and analyzes it for sound. Since this band is above the range where you would find human voice, and since the better the SNR, the less noise you would find in this slice range, the comparator concludes that the receiver whose noise analysis gives the lowest value is the best. It then keys the transmitter and routes that audio (now filtered) to be transmitted. This process is repeated very rapidly so long as any receiver is unsquelched.

Usually, the receivers are remote from the comparator and connected by leased phone lines. If the comparator has no way of detecting a logic COR or squelch signal from the receiver, then (i) it cannot distinguish a malfunctioning receiver or phone line from a clean one, and (ii) it may tend to vote on the malfunctioning receiver (since zero modulation in any range would meet its specification of lowest noise). Indeed, in such circumstances, the malfunction could cause the comparator to think that a receiver had unsquelched and it would key up the repeater, which would then hang.

To handle this, receivers used for voting systems (either based on own circuits or by an external interface) send a continuous tone (usually 1950 Hz) over the line when the receiver is unsquelched. This is called the "idle tone." The absence of the idle tone plus the absence of voice is specified as a faulted receiver, and the comparator on detecting this will exclude that receiver from the list of voting candidates until the condition is reset (either by the resumption of tone or by the technician).

In some systems, this supervision is raised to an additional level by causing the receiver (or its interface) to superimpose a different tone (called the "pilot tone") on any audio transmitted over the line when the receiver is unsquelched. In such systems, the comparator will not consider any receiver unsquelched until it detects both the absence of the idle tone and the presence of the pilot tone. The pilot tone is notched out of the audio that is passed to the transmitter.

As I said, a bit simplified. Someone else can take on transmitter steering, FBICR, voter locking for data, etc..

That helps alot. Thanks

Audio levels are also a big thing to perfect. The most ideal system will have perfectly equal audio level from all the recievers. The result is you can switch between recievers in mid sentence and never notice. Systems with poor audio balancing always show it with voting. You hear mid sentence changes in volume and pitch as it jump from one reciever to another.

Sometimes when I am bored at work, I will manually vote different sites during someones transmision to see what the other recievers are hearing.

Paul, I might note that the Motorola Spectratac automatically compensates for the varying phone line levels by monitoring the status tone from the receivers. That is why you don't need to constantly tinker with level settings to compensate for different levels coming in on the lines. The SQM sets the AGC level for each conversation based on the last measured status tone level. The manual is a very good source of information - 68P81039E50-A. The worst systems I've ever fixed were from someone trying to balance the audio levels coming into the comparator. That's just been my experience. I think they distorted the audio by overdriving or underdriving it from the source receiver. Good luck with it!

If you are using plain copper dry pairs then you'll need to set the line equalization or you could have an inbalance in receiver voting (having a receiver vote a slightly noisy signal while the signal is full quieting into another receiver). This is caused by an erroueous signal to noise ratio due to high frequency rolloff on the phone lines.


George