Digitac Voter Line Levels

[fullyfuly]

I am working on a Digitac Voting system that seems to have an issue voting the "correct" site. The voter has 4 different sites and (1) site (site D -see below) seems to vote even in areas where I am 99% sure it should vote a different site (i.e. I could be standing a few hundred feet from a sites antenna and it will vote this other site). My first thought is line levels, so at the voter I measured all of the idle tones which were all -29 db (keep in mind I'm using an old lineman so this may be off slightly. When I had a coworker inject a 1000 Hz tone @ 2.5 deviation (it's a narrowband system) at each of the receiver sites they were all over the board (not good). Here are those levels:

Site A -16.5
Site B -13.5
Site C -12
Site D -18.5

What is confusing to me is that Site D has the lowest audio level overall so why would it vote this site more often, even when closer to other sites? I believe the difference from the Idle tone to the 2.5 deviation tone should be 13 db or so. Could this be the reason as site D is only about 10.5 db apart? I know all of the levels should have the same idle tone and 2.5 deviation tone at the vote (or that's how I understand it) and I will have to make adjustments to the MTR's in place. I'm just miffed at why the site with the lowest audio levels would get voted. Any insight or suggestions would be appreciated, so if you have experience with this unit I'm open ears. Thanks in advance.

[The Pager Geek]

Briefly, before I run out the door:

- It's voting based on signal to noise ratio. Therefore, if a line level is low, it can't "hear" the noise as well as the other sites. (Simply put.) If it sees less noise from Site D compared to the others, it votes it.
- Also, You need to make sure telco line characteristics are the same. Same freq response from 300-3000, otherwise you'll need to use the equalization in the receivers to make up the differences. Telco lines typically have a freq rolloff that decreases the high frequencies. Why is this important? "Noise" and static is high frequency. If the telco line attenuates it, the voter cannot make the best judgment.
-Lastly, this is why co-located receivers to the comparator almost never vote. There is no rolloff to the comparator, and it sees EVERYTHING all the time. There are kits called roofing filters for the spectratac that mimic the telco line rolloff I mentioned, but I'm unaware of anything for the digitac.

Off the top of my head - gotta fly

tpg

[Bill_G]

+1

Less noise means voted more often.

Now you get the joy of figuring out if it is truly a levels problem, or a line equalization problem, or both. Woohoo!

[fullyfuly]

Thank you for your input. So far it looks like it was not the telco lines because an ajustment to site D to -13 (with deviation) allowed the other receiver site to be voted appropriately. I appreciate the expertise and that you once again.

[The Pager Geek]

Thank you for your input. So far it looks like it was not the telco lines because an ajustment to site D to -13 (with deviation) allowed the other receiver site to be voted appropriately. I appreciate the expertise and that you once again.

I'm glad it was that simple. Tracking line characteristics are a REAL pain in the...

tpg

[GlennD]

Years ago we had a MSF5000 that always voted. We finally tracked down the problem. It turns out that the tech that programmed the radio forgot to disable the 2175hz filter.

This made the unit look better than the other seven receivers.

[fullyfuly]

Speaking of Telco Line issues. I did see in the MTR2000 software that there was adjustments in the RX Wireline to make compensations for the Telco Lines themselves, thus potentially minimizing telco line issues (notice I didn't say eliminating). I will post this information on Monday when I'm back at the shop or you can view it in the MTR2000 software help files. I've yet to use this feature, however the next time around we go through and make adjustmens to the receivers this will be included in the process.

As a follow up questions. How often do you think the systems levels should be monitored? This is a bit of a labor intensive job as you need a tech at the comparator and a tech at the receiver site, unless there is a better way.

[Bill_G]

Years ago we had a MSF5000 that always voted. We finally tracked down the problem. It turns out that the tech that programmed the radio forgot to disable the 2175hz filter.

This made the unit look better than the other seven receivers.

That's interesting. Don't recall running into that, but it's subtle enough it could bite you in the butt. Thanks.

[Bill_G]

Speaking of Telco Line issues. I did see in the MTR2000 software that there was adjustments in the RX Wireline to make compensations for the Telco Lines themselves, thus potentially minimizing telco line issues (notice I didn't say eliminating). I will post this information on Monday when I'm back at the shop or you can view it in the MTR2000 software help files. I've yet to use this feature, however the next time around we go through and make adjustmens to the receivers this will be included in the process.

As a follow up questions. How often do you think the systems levels should be monitored? This is a bit of a labor intensive job as you need a tech at the comparator and a tech at the receiver site, unless there is a better way.

Actually, I check basic levels at the comparator as part of the yearly system PM comparing them to the original "as builts" we generated, but I never check the line equalization unless I suspect a problem. We use 66 blocks to terminate signal wires making it easy to check with a TIM set. While that has proven mostly unnecessary with full T1 microwave backbones, it is still needed with mixed telco & microwave systems. As you found out, a difference of several db in rx level will cause the voter to favor one site over another despite the status tone AGC.

I don't recall what the spec is, but the line AGC is supposed to help overcome slight changes in levels based on the status tone so that "normal" changes in the levels received from telco do not cause service outages. The assumption is the 13db relationship of the rx audio remains constant relative to the status tone. That is, with a 0db 1khz tone at the rcvr and a 2175 status tone 13db down through a 10db telco loss, we can expect the 1khz at -10db and the 2175 at -23db if the line has a flat response. The comparator will do it's best to raise or lower the receive levels on the audio bus based on the last status tone of each line input. However, if the response is not flat, if the 13db relationship does not remain constant across the audio band, than voting is affected especially if there is a cut in the high freq content above 3khz where the comparator does it's work.

Then you have to perform equalization tests on each line, and either compensate for them by boosting or cutting at the rcvr line driver if such control exists, or have the telco fix the lines. The MTR does have some latitude. Certainly more than most equipment. Since we are only concerned with three octaves, simple bass / treble controls will work. But then comes the matter of making them, installing them, and testing them which is often beyond most shops. Quite often the problem is loss of high freq content which is usually an indication of wet splices. The overall quality will be mushy. Sealing current can only do so much and then it fails to keep a line working well under adverse conditions. So, it's often best to call in the telco to do their work first before you attempt to compensate. Once you've made adjustments, it's generally not necessary to readjust them for many years.

[FMROB]

Bill is right on with the above post. He helped me through setting up a four site system using astro tac rx'rs and a spectra tac voter, mixed microwave and phone line, with a co located rx'r... And it all works great, his advice is spot on. The only thing with the 12.5 systems is low TX audio, which is a pain to adjust. The more you crank up the overall TX audio to the transmitter the spoke audio sounds better, but reliable keying can become and issue, and in our case it does.

Conversly, with the astro tac rxr's the built in line conditioning works well. Even with the co located rx'r, it puts you throught a series of tone tests and measurements that I assume ad deigned to narrow down settings to adjust +/- line loss.

Thanks again Bill from months ago. Rob

We plan in the near future of replacing the spectra tac with a Digitac or JPS unit?? Jury is still out, and I will put that out over here during our decision making. The district has a brand new in the box Astro Tac 3000 unit that was sold to them by another vendor, but I am not going to use it with DIU's, modems etc. So if anyone has a brand new digitac unit, a swap could be made....

[wa6jbd]

Another potential issue with the Digitac's is the 4 AGC chips on the input cards. There are 4 identical inputs per card, and the AGC chips have a known issue that can cause problems like, improper voting, low levels, and radical level shifts during a transmission. The fix is to replace the original AGC chip with a small SMT carrier board that fits into the original socket. I ended up replacing every AGC chip on every digitac voter we have (number of chips in the 100's) and things have settled down quite nicely.