The KA7OEI ATV repeater

Note- This repeater is off the air until further notice: The repeater's equipment is being reworked. The transmitter will be used on 1248 MHz for a NASA backhaul feed to the WB7FID ATV repeater.

The KA7OEI ATV repeater was located on the campus of the University of Utah, in the 'shack' of the University of Utah Amateur Radio Club room. This is in the extreme northeast corner of the valley (near the "U" on the mountain) and covers Salt Lake City and it surrounding communities. It has been in operation since July, 1992.

ID screen on KA7OEI/RPT

ID/Beacon screen on KA7OEI/RPT

This repeater is a cross-band repeater, with its input on 70cm and its output on 23cm. The receiver is remotely controllable, but it is usually parked on 439.25, the output of the once-was WB7FID repeater. This repeater transmits on 1.2480 GHz using FM.

Antennas (horizontally polarized):

The receive system has dual antennas: A Yagi, still pointed at the point-of-the-mountain where the WB7FID repeater was (in order to get the best possible signal) and a Corner Reflector with a wide, flat pattern, directed into the valley.

The antennas are manually switched by a coaxial relay. There is an MRF966-based preamplifier, followed by a 3-pole bandpass filter (with its passband covering 423 through 447 MHz) followed by another identical preamplifier. The receive signal sent to the receiver through 1/2 inch CATV coaxial (with 1/4 wave, 63 ohm sections at each end to provide the 50-75 ohm transformations) and into another 3-pole bandpass filter, identical to the first. The antenna is selected by changing the power supply voltage on the coaxial cable: One antenna is selected when the voltage is below 9.5 volts, and the other when it is above. No problem for a 5 volt preamplifier...

The transmit antenna is a 90 degree corner reflector with the dipole stationed to produce a very wide, flat pattern which, owing to the geography of the valley, provides complete coverage.

Receiver:

After the second bandpass filter, the received signal goes directly into a high-level DBM. The IF output of the DBM is diplexed, with the low split being sent to the video/audio demodulator, which was pulled from a defunct VCR. The Local Oscillator is synthesized in 1 MHz steps, and the receive frequencies of 426.25, 439.25, and 434.25 (close enough to 434.0 that it doesn't matter...) are selectable.

*Equipment rack at KA7OEI/RPT*

Controller:

Nothing fancy... A 70cm receiver is coupled to a DTMF decoder, a PL decoder, and an EPROM-based state machine. This combination allows remote frequency selection, controller resetting, antenna selection, and enabling/disabling the repeater.

The received video is passed to the video circuitry. This consists of a sync detector (for keying) and a recovery system for both sync and colorburst (for the video ID overlay). Before hitting the transmitter, the video is run through a video corrector to clean up both sync and colorburst.

Transmitter:

The transmitter is also homebrew. The video is lowpass filtered above at 5 MHz, pre-emphasized, and combined with a 5.8 MHz audio subcarrier. The transmit frequency is synthesized, and a lock-detect circuit disables the transmitter unless it is on frequency and fairly stable.

Two hybrid power modules take the power level from the 5 milliwatt level to 20 watts. The feedline is 3/4" CATV coaxial cable. This, too, has a 1/4 wave 63 ohm section on each end for the 50-75 ohm transformation. Of the 20 watts going into the coax, about 12 watts makes it to the antenna, yielding an EIRP of approximately 150 watts.

Operation:

When not repeating a signal, an computer generated ID/info screen (as seen above) is transmitted as a beacon signal (legal on bands above 450 MHz.) The peak deviation is 4 MHz, and the pre-emphasis curve applied is that for wider deviation satellite transmissions so that old satellite receivers may be used to properly demodulate the signal.

In the receiver chain, one may wonder about the receiver passband filtering allowing 423-447 MHz to get through. Doesn't that leave one open to intermod? It depends. GaAsFET preamplifiers have much higher intercept points than do most other types. Also, since the vast majority of any intermod that one gets occurs in the mixer stage (hence the total 6 poles of filtering...) it would make sense do the filtering, and make the mixer as resistant to intermod as possible. The use of a high-level DBM goes a long way toward this goal: It can take a lot of signal before nonlinearities creep in. As long as the total signal getting into the mixer is far enough below its intercept point, it will not produce problematic intermod. Because of this, this receiver can withstand not only a 12 watt repeater on site, but also a 10 watt 9600 baud packet node, both within the receiver antenna filter's passband without being clobbered.

When selecting the frequency of operation, it was decided to put the repeater as low in the band as possible. If Carson's rule is applied, you may note that the calculated bandwidth would indicate that the signal was too wide. Actual measurement, however, showed that the strongest out-of-band signal (under worst-case video input conditions) was, in fact, the second sideband of the audio subcarrier, and it was greater than 45 db down from the unmodulated carrier, cleaner than most 1241.25 MHz VSB signals currently on the air. A newer version of the transmitter will reduce this even more by bandwidth limiting at a lower IF before upconversion.

Single-board 23cm ATV downconverter

Single-board 23cm ATV downconverter

It would be pointless to have a transmitter transmitting with no-one able to receive it. To this end, two different, successful approaches are used: The first approach uses 70 MHz satellite receivers (still relatively cheap and plentiful) so a downconverter was needed. To this end, a single-board converter was designed and built (with LO, mixer, etched filters, etc.) The second approach is to use a block-type satellite receiver (which can tune the 23 cm amateur band directly) with pre-amplification.

To be sure, both of these approaches do require that the receive bandpass filter be narrowed from its original 27-36 MHz to something in the 12-16 MHz area if one is to obtain optimal noise performance. However, owing to the strong signals from the repeater, and the fact that there is no activity on the band below the mode L satellite frequencies, interference has not proven to be any sort of problem.

Any questions about the KA7OEI ATV repeater?
Why not email KA7OEI and make an ask of yourself?

Return to the Utah ATV home page...