Succesful EME tests on 3.4 GHz 27th. Aug. 1997/MF ============================== Compilation of e-mail messages to the moon-net@vm.sltawu.edu EME ---------------------------------------------------------------- mailing list. ------------- Success, but room for improvement... ------------------------------------ Our 3.4 GHz EME test on the Sipoo EME Racing Team 6.4 m dish at OH2AXH was successful with the help of Pertti, Harri, OH2JMS and my wife Anne. Cold sky/ground noise level around 2.0 dB, lunar noise approx. 0.5 dB and sun/cold sky noise 10.0 dB. So there is definately room for improvement. The RX line includes two sections of semirigid cable, the tx/rx relay ( a good microwave relay ) and two waveguide transitions. The LNA needs to be modified for coaxial input with an SMA female connector. Own echoes were copiable about 50 % of the time whenever the dish ( manual pointing ) was on the moon with nice peaks occasionally. We need more power !!! Brief description ----------------- My rig is running about 8 Watts of RF into the antenna/polarizer system. The antenna return loss is around 16 dB and varies slighly with polarisation. The LNA is a modified/retuned Gardiner C-band LNB. The noise temperature of the LNA might be around 25 K. The system noise temperature I will calculate later based on the figures measured this AM. The extra path loss was 0.25 db this morning ( due to lunar distance ). In focus or not ? ----------------- Our center frequency is 3400.100 MHz. Unfortunately we cannot use 3456 MHz in Finland. Polarisation linear and fully remotely adjustable. The entire transverter setup is mounted at the feed of the dish. The focus point and scalar collar need fine tuning, but right now we are calling it a day. More EIRP ? ----------- I just briefly tested a C-band SSPA originating from Chris, KL7FB. The amp puts out 12 W solid CW in compression at 3.7 GHz, but needs retuning down to 3.4 GHz. The improvement is not sufficient to motivate us into installing the amp onto my EME rig. We need some- thing like 40 Watts or so. Maybe a 2C39 design will be the solution. I hope time permits some work on this before the year is out. I have TWTA for 3.4 GHz that will do 30 Watts easily, but weighs a ton. It uses the RW42 TWT. More detailed description ------------------------- I am using the basic DB6NT design as a driver for 3400 MHz. The LO is around 136 MHz. I have also sampled the 136 MHz oscillator and have it amplified with a Maxim MMIC ( comparable to Mini Circuits ) as a buffer. Both the LO and 144 MHz IF i/o are fed down from the dish feed to the rear of the dish or even into the shack via cabling. The LO freq. I can check against my VHF frequency counter for which I have built a secondary Rubidium standard reference being transmitted nationally over the national TV network ( the line, field and coulour burst frequencies are phase locked to a Rubidium standard. Also the digital stereo sound carrier data clock is locked too ). In this way I can accurately track the frequency of my LO. I have not ( yet ) been bothered with locking the whole microwave LO chain to this secondary reference. My secondary reference is also very pure spectrally, one of the main specs I set for the system ( I also use it for referencing my signal generators in the shack ). I used this same secondary reference for my 10 GHz EME rig back in 1995 at OH2AXH. Comparison noise measurements in the shack ------------------------------------------ By comparison measurements ( I don't have a calibrated noise source in my shack ) I concluded the noise temperature to be almost the same on 3.8 and 3.4 GHz, and can be set to the same figure ( 25 K typ. ) by just slightly increasing the series input inductance on the input P-HEMT + also tuning the output of the same stage with a flake of copper sheet. Comparison measurements made with feedhorn and tr/rx switch system incorporated included to achieve best noise matching. The noise source is a regular incandescent lamp ( ! ) in front of the horn at a suitable distance and being switched on and off or alternatively any gas discharge tube or light that is set at a further distance so as not to overload the RX ( this also ensures better matching for the LNA as the feedhorn sees "free space" better ). I need to make an accurate noise temperature for the system soon. I intend to use a coaxial load I have that I am confident will withstand the boiling temperature of nitrogen. I will use nitrogen for the cold reference and melting ice for the hot reference ( or boiling water ). By measuring the Y factor I can determine the system noise temperature at the coaxial input. If I can make a reasonable load "wall" into a styrox box around 1 meter squared and fill the box with nitrogen, I could measure the noise tempereature at the feedhorn input. I am able to measure coax losses accurately though, so this may not be necessary, unless I begin to doubt correct operation of the horn. Noise level detection --------------------- The ideal way to measure the Y factor would be by measuring the IF noise level selectively, but the 136 MHz LO leaks badly into the 144 MHz IF, so selectively measuring noise power ain't so trivial. At least not comfortable. So I simply measure the AF noise level at the speaker output. This would require a linear detector, but I cannot switch off the AGC. So what I do is I measure a certain level, set this to zero dB reference, change the noise source and by means of a step attenuator at IF reset the audio level to the reference. The reading of the step attenuator equals the Y factor. I will revert to measuring the noise level at IF frequencies ( proper RF ) soon though dues to easier integration. The TVRO surplus hardware ------------------------- I originally started collecting this C-band stuff for a Dicke switch radiometer for 9 cm, but as Finnish hams got this new band last spring I though what the hell... The feed horn is of the "Chapparal" variety and seems to be doing a good job. It was originally a dual band ( C/Ku ) horn, but the Ku stuff was stripped out. The servo is left in and I have built a calibrated NE555 oscillator to drive the polarisation angle from the "shack". This is a variable PWM circuit with a calibrated 10-turn pot. The SSPA is definately a difficult problem. I have a 30 W TWTA I don't feel too happy to install at the feed. It's lots of power with zilch drive requirements, but all the hassle with high voltages on coax and the shear weight of the damn thing ( it is a PPM focused 1970's design using the RW42 from Siemens ). More power ---------- The DB6NT design uses two MGF0904 GaAs-FET's in parallel with a third one as the driver, and they do 7 - 8 Watts quite easily. This is the PA I barely heard my own echoes with ( remember the scalar choke is not in the correct location and the hold feed may be out of focus ). I now will install the commercial SSPA onto the rig. This is specified at 10 W at 3.8 to 4.2 GHz and it did 12 W ( yeah - my power measurement setup is in good calibration ) on this band and it now also gives the same output with the same drive ( and even less ) on 3400 MHz after retuning. Out of focus and overillumination: ---------------------------------- One improvement is made on the 3.4 GHz EME setup: the scalar ring was too far back ( this was set empirically due to lack of better information ) and the feed was consequently overilluminating the dish. The 6.4 m dish has an f/D of 0.30 and the used scalar setting was optimal for around 0.23 or so. The scalar ring setting and the focal point checks will be carried out and set for optimal sky/ground and moon/ground noise when time permits.