Cutting-edge diy projects
The highlight of my weekend was a JT65C EME QSO with Andrey 4X1AJ in Israel! That’s my personal longest EME DX. My 1296 MHz PA output was 600W, out of which I’m putting 400W at the feed of the 3.5 dish. In Andrey’s words, I was “booming” on his side (-17dB).
Very pleased with my current setup! THX Andrey for the QSO! 73
Before I package my 600W 1296MHz Power Amplifier into a single enclosure to be placed closer to the antenna, I thought it would be a good idea to take some pictures and draw a block diagram in case anyone wants to build something similar.
My original idea was to build a 600W PA using W6PQL boards. However, things have become very busy at work, so realistically building the PA would have taken a very long time. As a compromise, I decided to buy a 300W pallet and its companion controller from VHFDesign, not really knowing how that would go.
VHFDesign’s 300W 1969MHz Pallet with Heat Sink and Controller. Image Credit: VHFDesign
I was very pleasantly surprised by the VHFDesign unit. I purchased it directly from their website using PayPal. The unit arrived quite quickly from Ukraine. The instructions are clear, so once I got around to it, installation went easily. Their 2-cascade LNA has also been a pleasure to use, but I’ll leave that review for a future post.
After working at 300W for a while, I decided it was time for an upgrade, so I purchased a second unit with the intention of combining their outputs using ring hybrid couplers. The final block diagram is shown in this figure. The splayed-out view of the modules is shown in the following figure:
With an output of 326W for each VHFDesign pallet, each (suitable for CW and SSB), The PA produces a combined output of 600W. IMPORTANT NOTE: See my post about phase equalization! I’ll optimize cabling and phase matching once I assemble everything into a single enclosure. For now, it all works nicely splayed-out indoors.
Last weekend I found that I was loosing quite a bit of power when combining the outputs of two 300W VHFDesign 1296MHz power amplifiers through a hybrid-ring coupler (rat race). The power was going into the isolated port of the output hybrid-ring coupler, so that pointed out to a phase mismatch between the two amps. I don’t have an adjustable phase shifter handy (bought one off eBay yesterday), so I decided to try a simple change in cable length using adapters. Just one NF-to-NM “port saver” adapter was enough to equalize the phases at the amplifiers’ inputs such that the power appearing at the output coupler’s isolated port is now just 4W, which boosted my output to the antenna from 520W to 600W!
Today was a nice, sunny day, with temperatures in the 40s, so I decided to measure the power drop along the coax from my 1296MHz power amp to the feed. The PA is at the closest indoors site to the dish.
I am measuring 520W going into the 35 feet or so of LMR900-DB that takes the signal to a junction box mounted at the antenna mast. There, I measure 420W remaining, which go into 27 feet of LMR600-UF to the feed (mast + rotor loop + cable to reach feed on 12 ft dish). After that much cable, barely 320W are left to go into the septum feed.
The numbers check out with the expected cable attenuation, so I understand that I’m simply experiencing normal cable loss. However, loosing 200W to heat the ground bothers me, so I’m looking for ways of increasing the power that reaches the septum feed. I’m first considering ways of keeping the amp indoors, but I’m afraid that I’ll ultimately have to put it all at the mast and use a hanging cable loop straight to the feed to minimize losses.
My dish’s Az/El are controlled using a Green Heron RT-21azel that Jeff kindly modified to accept input directly from the US Digital T7 RS232 absolute inclinometer on the dish’s mount. As such, I always have certainty about the dish’s elevation. However, I don’t have any absolute reference for azimuth, so my preference is to rely on the location of geostationary satellites for calibration.
To do so, I have a Chaparral Corotor II with C/Ku LNBs mounted 7 degrees to the side of the main feed. An issue occurred however when I had to mount a 18″ diameter scalar ring on the 1296 MHz septum feed to improve illumination of the dish, thus blocking the Chaparral’s aperture. Cutting a hole on the scalar ring solved the problem. This did not reduce the efficiency of the scalar ring (I still have 11dB of Sun noise over cold sky on the 3.5m dish). Jason (KC2TDS) will cut a second hole on the other side of the feed for the DSN X-band feedhorn that he built.
Of course, this arrangement invalidates the Chaparral’s scalar ring, but the signals from the satellites are strong enough on both C and Ku to be received at my QTH without the need for optimizing the feed pattern.
Now, to be able to receive linearly-polarized satellites along the full geosynchronous arc (the Clarke Belt), the skew angle needs to be adjustable. This is why I use a Corotor – because the polarization angle can be finely controlled through a servo motor on the Corotor II (the blue box in the back part of the feed) that rotates the polarization-selection element inside the Corotor’s feed.
The servo in the Corotor II works just like a regular hobby servo, which allows me to control it from the shack using a simple PWM servo tester.
My experience using the BIG-RAS rotator has been excellent since I got the Green Heron Controller (instead of the SPID Controller). I set it to ignore pulses when it’s not being commanded to move, and haven’t needed any resetting whatsoever after high winds. I have a few of the geosynchronous satellites as presets on PSTRotor, and always end up spot-on when checking positioning accuracy.
This weekend I was able to combine the power of two 320W pallets using ring hybrid couplers (rat races) for an output of around 600W. The strength of my own CW echoes improved dramatically. The setup above still needs cleanup, and I may end up placing it at the base of the antenna to reduce cable losses. On the meantime however, my system is ready for CW QSOs.
Conditions were good today, and there were quite a few stations calling CQ off the Moon. Unfortunately, I realized how rusty my code has become, and cheating by using decoding software does not help at these signal levels, so I decided to hold off on answering until I get my code back up to speed.
I also started playing with JT65. I can decode straight from both the IC-9700 and the AOR AR5000/SDR14. Next weekend I’ll build a circuit so that the computer can trigger the PA sequencer to be able to TX JT65C.
