Batboard

Has anyone been able to get a spectra tac voting system to vote properly using fiber cable? No mater how many times I set line levels, I can not get consistently good signal quality voting with my hand helds. It seems to perfer the receivers on fiber lines to the ones on copper even though the signal to noise might be better on the receiver with the copper line.

Joe

This is a known issue with Spectra-Tac.

Spectra-Tac will only vote properly if the circuits connecting the individual receivers to the comparator have comparable [preferably identical] frequency response characteristics.

A common situation is a receiver is co-located with the comparator that almost never votes. This is because the local pair of wires passes high frequency noise content that leased lines do not. The Spectra-Tac SQM "sees" that extra noise and decides that other receivers are better when they aren't really.

This problem can affect any Spectra-Tac system where different types of connectivity are utilized.

Motorola provided a solution to the problem in the form of a voice frequency filter module [roofing filter]. These could be factory ordered as option C366AD-SP or the modules could be ordered as QRN8498A. Each module takes up a single slot in a comparator chassis. The incoming wireline is connected to the filter module and the module then connects to the SQM input.

You can actually still order this part!

Item Number. QRN8498C
MODE AUDIO FLTR SPECTRA TAC-OUT
Unit of Measure: 1 EA

Thanks for the replay. I have the roofing filter, it doesn`t seem to help. My thought is that the phone company`s equipment is processing the receive audio in such a way as to interfer with the voters ability to properly determine signal to noise quality.

Joe

The other way to handle this is to send the local receiver out to one of the fiber hops and then back again.

...so in essence forcing all the local input to have the same noise characteristics?

I've encountered this before, although the circumstances were a bit different. In my case, the transport mechanism was analog microwave. There was another difference... There were three receivers, two contained in Motorola Micor upright's, and the third was a Micor satellite receiver. The Spectra Tac favored the stand alone receiver, even though the microwave circuits were otherwise identical, and even though it was in an area that was seldom used, so the system consistently delivered scratchy noisy audio to the dispatch center.

I eventually ended up injecting a low level tone into that SQM, along with the receiver audio to give the comparator something to chew on before it made it's decisions. That receiver ended up voting only when it was actually better than the others.

I suspect in your case, it's not so much the fiber that's the issue, but maybe the digital transport medium as opposed to copper. Is the copper line just an analog leased line, or is it a copper T1? Once the channel bank converts the audio to bits, the bits don't normally care how they're transported, so long as they get reassembled in the proper order at the other end. There could be other issues at play - circuits that use adaptive rate sampling (i.e. via routers or ADPCM 4W cards) can add quantizing noise that an analog circuit doesn't have to contend with. They may work fine with phones, but fax machines and modems don't like 'em. Maybe voting comparators don't either.

To solve the problem of a receiver co-located with a comparator and "throwing the vote", I've set up a digroup on a channel bank to loop back, and passed the receiver audio through the channel bank to give it identical characteristics.

Another thing to consider is the possibility of latency through the fiber. An analog phone line has a fixed and predictable propagation delay. Fiber circuits can rought all over the place, may pass through several switches along the way, and the increased (and variable) delay can be substantial. Part of the issue could well be that the two SQM's in the voter are trying to compare the audio equivalent of apples and bananas - it may not even be the same shape, let alone the same thing that's arriving at the comparator.

So what is the answer, use rf links? Would the Raytheon SNV-12 be better able to cope?

Some history:

SpectraTac (and even to a greater degree, JPS for FM) vote based on the level of high frequency (audio range, AF) noise in the range just below 3000Hz. Lots of folks call this white noise. Essentially we make the assumption that as an incoming radio signal becomes poor it produces noise in the range just below 3000Hz.

Typical voice grade telephone lines that are copper often have attenuation or high frequency roll off. This is often in the range just below 3000Hz.

So what can happen is that the voter picks the the telephone line audio, even though the recovered audio is very poor, because it sees less white noise in it's decison making process. This is often called down voting.

Several things can be done to help solve this. First, set the squelch such that the receivers only opens to deliver reasonable sounding audio onto the line. Second, set the squelch logic to require Squelch AND PL to open and to stay open.

Then, if you're using the JPS voter, follow "line equalization" setting proceedure and "bias" the SVM that uses the telephone line against voting by some amount.

For SpectraTac, Motorola used to make a line equalization kit. It was installed on the "good" lines (in your case the one supported by Fiber) to make them have as poor of quality as the telephone copper. It, for the most part, consisted of a capacitor across the line to intentionally cause high frequency attenuation (roll off) equal to the roll off being experineced in the copper based circuit.

In the past I've installed a non-polarized electrolytic capacity across the "good" line to make it look as bad as the telephone line and I may have used a small resistor in front of it to create a simple RC low pass filter. It took some playing around with the values (I had a box full of different values) to get the right results but it did work.

Regards, Bob

Thanks for the info. The equalization module you mention, is that the same as the filter module used for local receivers? I have one for the local receiver, but never thought of using one with the fiber lines.

Joe

Joe,
I'm not familiar with the fiber filter module you mention. Can you send me a web link to the specifications?
Bob

What I was refering to is the QRN8498A roofing filter module. Also I am looking for a copy of Motorola PSB-653 and SRN-1068.

Joe

For what it is worth:

One of the systems I attend to uses a JPS Voter and 8 receivers. Six receivers come back via all fiber. One is local to voter, and comes back on 8 feet of copper. The last receiver is 15 miles away and comes back by an exotic mix of T1, then 1800' copper, then fiber.

So far all have voted without issue.

This is my fifth or sixth JPS installation. Tweaking the JPS requires a bit of attention to detail, in part because there are more adjustments that can be made than on the SpectraTac, and in part simply because you approach it a tad differently. That said, so far all of the voters have functioned just fine.

One caveat: because they use vocoders, both fiber muxes and the JPS voter are both far less tolerant of "hot" audio than the SpectraTac. A line that, for some reason, is coming in +10 dBm or so may or may not be noticed on a SpectraTac system, but it will be noticed on the JPS and the fiber. Our standard, enforced with care, is -13 for 1000 KHz tone at 60% deviation and -10 for idle tone. We then adjust the SVM cards per the JPS manual, which is quite different from what one is used to.

Joe,
I've not seen that particular filter before so I can't be of much help. I'm guessing that if it's needed for the local line it's job is to create more attenuation on the local line at the higher frequencies. So, you should be able to apply it to the fiber lines on the local/voter end too. Good Luck!
Regards, Bob