GTR8000 Simulcast alignment

[dag]

I am putting in a new simulcast system using GTR8000 repeaters and equipment supplied by Simulcast Solutions. When I attempted to do a mod comp measurement using my Aeroflex 3920 opt 210 to supply the audio, with the radio keyed the PA failed, completely. No output. This happened on two radios.

I applied the 3920's Tracking Generator output to the TX audio input on the GTR's system connector. PTT was accomplished by simply grounding the PTT input on the system connector. The transmitter antenna port was connected to the combiner input, a brand new combiner which I checked with my wattmeter prior to doing any other testing. A dummy load on the antenna port of the combiner. The 3920's TX audio level was set to -20dbm.

Anybody have any suggestion on what may have been wrong?

[Wowbagger]

When you say the PA failed, was this a permanent, "magic smoke" type of failure, or was it more like it shut down temporarily?

And when you say you connected the tracking generator output to the mike - do you mean you connected the RF gen output to the mike, or the audio output to the mike?

[Bill_G]

I'm not familiar with the GTR8K series, but mod comp is usually done from a radio alignment screen within RSS to set transmit linearity on each radio. You align each radio to match the others are close as possible so that a given input produces the same output.

[dag]

Magic smoke release. Permanent.

Output from the Fctn Gen/Demod connector.

The mod comp adjustments we have done involve a sweeping audio signal fed into the transmitter audio input to provide a visual reference of the response curve as viewed from the demodulated signal. Pictured on page 8-18 of the 3900 operation manual.

[xmo]

With the GTR you service and configure the station using CSS - assuming you can get past the security BS and get logged on to the station.

Exactly which pins were you connecting to on the GTR?

Also, are these the same connections the system will be using - in other words - how are you distributing the simulcast audio and PTT from the central point to the GTR's and how are you providing delay optimization?

[Wowbagger]

I'm not familiar with the GTR8K series, but mod comp is usually done from a radio alignment screen within RSS to set transmit linearity on each radio. You align each radio to match the others are close as possible so that a given input produces the same output.

True, but to make the decision that each radio "matches" you compare the audio spectrum of each radio with one radio you've decided is going to be your reference. You get that spectrum by driving a stimulus signal into each radio to get the audio spectrum - either a tone sweep or a sin(x)/x pulse, which is usually generated by what you are using to measure the audio spectrum of the modulated signal - in this case the 3920.

[Bill_G]

I'm not familiar with the GTR8K series, but mod comp is usually done from a radio alignment screen within RSS to set transmit linearity on each radio. You align each radio to match the others are close as possible so that a given input produces the same output.

True, but to make the decision that each radio "matches" you compare the audio spectrum of each radio with one radio you've decided is going to be your reference. You get that spectrum by driving a stimulus signal into each radio to get the audio spectrum - either a tone sweep or a sin(x)/x pulse, which is usually generated by what you are using to measure the audio spectrum of the modulated signal - in this case the 3920.

Again, I'm not familiar with the GTR series, but all high tier product in the past generated it's own waveforms in the mod comp test, and the tech set the flatness of the slope by observing the service monitor while adjusting a slider. The result should be a vertical rise with a flat down slope. If you set that correctly, you are almost guaranteed faithful transmit envelopes. Then it is just a matter of setting your levels. The Quantar and all it's variants (including Nukes) were so well made, if you got the mod comp right, you could set the input levels blind, and still have deviation within a few hundred millivolts station after station. They were amazing. I would love to work on GTR's at least once, but that probably won't happen before I retire.

[dag]

Another day in the field yesterday on the second site. I was able to determine the cause. TX audio is fed to the GENTX inputs on the system connector. With the channel bank still connected grounding either of the two GENTX inputs causes the radio to key and also causes TX power control problems. Twice during my troubleshooting the radio keyed with maximum power, pegging my 150w meter, from a 100 radio. That level of power for an extended period of time will cause the transmitter to smoke.

Short answer, I have to break the ground between my service monitor and the GENTX input, either by isolation transformer with a wide and low frequency response, or put capacitors in series. On my way to site three today.

[Bill_G]

And there is no caution about that in the service lit? hoo boy!

[xmo]

It is surprising that connecting an unbalanced device to the GTR causes a PA malfunction, but it probably indicates a transformerless, balanced input, solid-state circuit.

That said, we wouldn't normally hook any test equipment onto the base station at the remote site site anyway. We perform optimization from the central point where the simulcast audio originates using an Agilent 35670A Dynamic Signal Analyzer in accordance with Motorola's simulcast system optimization procedures.

Here are some quotes from Motorola:

"Modulation Compensation is extremely critical to system optimization in order to achieve best delivered audio quality in overlap areas. The goal is to achieve less then a 0.3-dB difference in amplitude response between sites, the shape of the low frequency end of the Frequency response curve will be measured utilizing a dynamic signal analyzer.
...
Due to the simulcast equipment configuration this additional fine-tuning step is easiest to perform as part of the Fine Amplitude Adjustment procedure using the Dynamic Signal Analyzer and the optimization receiver.
...
When using the ... Dynamic Signal Analyzer it is possible to set the gain, check the modulation compensation adjustment, and do the Final System Level Check, all in one operation."

----------------------------------------------------

In addition to modulation compensation, overall frequency response, and absolute gain, simulcast depends on proper phase which the above procedure does not verify, since it is assumed that the delay optimization is automatic within the DSM-II (although we did verify proper phase one time to insure that there was nothing "inverted").

For our paging simulcast optimization which uses Convex delay modules, we perform the "two-receiver" amplitude and phase verification using the DSA. This entails using two analyzer channels and trace math along with an instrument setup that is rather challenging.

[dag]

I am familiar with Motorola's simulcast optimization.

A dynamic signal analyzer, or sending the audio sweep from the prime site to the remote sites is not possible on the channel banks we are using on this system. These are Harris NetXpress channel banks, the audio circuits are filtered to 300-3500hz. Subaudible tones are completely filtered out. I can't even force the PL tone through, so even trying to send anything as low as 20hz is out of the question. Pl is generated and inserted at each remote site, referenced and generated from the Spectracom GPS units.