VOIP Voting

[Gregg]

I have a really good mind puzzler for everyone. We are runing the JPS SNV12 over our network VOIP with Cisco Routers. All the Base Radios are Motorola Quantars. The Channel that is voted is alway the worst possible site. At this point the system working well but because we have one of our channels set up for TX steering the problem is magnafied. If a unit calls dispatch and the the system chooses the farthest site that can hear his radio then the Transmitter that is assigned to that reciever will be the TX site. If there is one thing that is consistent is that it will pick the worst possible site to Vote. We have set levels, Completed line Equalization with 400 hz, 1k, 2500k tones and +5 dB to the Front of the SVM cards. The Network seems to be attenuate the line levels of the sites that field unit is the closest to but on the distant sites it appears there is no attenuation. You can change locations and go to another site and do a radio check and the the same symtom follows. Which ever site you are closest to the line level will be attenuated, and the farthest station will not. The JPS is suppose to vote on the least noise so even with this symptom you would think that 10 db lower and less noise should be no problem. You can listen to the Audio and force the Jps to voth the site you expect it should vote and is sounds great. It will vote the lower line level. Has anyone ran into this problem or even aware it is happening. I have our system pretty well optimised now and it sound really good but becuase we have had so many problems we have really gone to great lengths to see what the system is doing and have tired everything to get the system working the way it should. the only wild card that was left was the network. So we decided to put a line level meter on the channel voted and the channel we know it should be voting. We are using G711ulaw Codec on the network which should have no AGC. Any bright ideas?

[Dan562]

I've only seen the problem occur on /\/\'s Analog Spectra-Tac comparators but I suppose it can happen on any brand. The problem you've described is called Down Voting, where the comparator selects a different noisier receiver site rather than the closest Full Quieting site.

I've got a question: Is your JPS Comparator colocated with the Base Station / Repeater?

If the answer is YES then I've got two solutions:

Call JPS and request their latest software which is suppose to correct your Voting problem.

Or

Order the following /\/\ QRN8498B Roofing Filter from A.A.D., the National Parts Department. The QRN8498B Roofing (VF Low Pass) Filter for the repeater’s receiver audio because of the audio bandwidth of the wireline circuit is infinite including up to 6 KHz noise between the repeater’s audio output and the JPS Comparator's Input #1 Signal Quality Module. You'll be required to supply +13.6V DC and Ground to the Roofing Filter and wire up the repeater's Wire Line Output to the module's Input connections and wire the module's Output to the 1st JPS Signal Quality Module's Input Wire Line. The /\/\ Roofing Filter Module has unity gain.

Dan

[psapengineer]

Here's a few suggestions:

First make sure that you're not set to lock on the first card that sees signal.

What is the bandwidth of your VoIP converters? If they don't have a bandwidth past 3K they could be the source of your problem.

I'd also be tempted to drag a 'scope out and look at the audio going into the svm input. Does it have some superimposed noise?

We had a similar problem with just one site; we found that the T1 channel card at one end was defective and had significant roll off above 2 KHz. It caused the site's svm to vote even when the audio was really poor.

Timing can also be an issue; is there a propagation delay that is causing cards to vote based on incoming signals that are not close to being in phase?

I'd be tempted to run a frequency response test at a given frequency like 2600Hz. I'd inject into each receiver at the same level and measure at the SVM for each site. They outght to be the same.

Ugly Problem you have; good luck; let us know how it turns out.

[/\/\y 2 cents]

Your problem is with the Cisco routers, Codec fidelity, and QOS. I would like to help you with this but I will need more information on the packet side (iP) of your network and what the routers and sitting on. Most of the "noise" in your signal to noise comparitor is being rolled off in the A/D --> D/A conversion creating this effect of "choosing" the wrong site beause the comparitor only knows what is being fed into it. Let me know if you need any more help.

Steve

[chpalmer]

Not to bring up an old subject but has this been resolved? Im having a similar issue with mtr2000's that Im just starting to diangnose...