Draft 1.0: Jun 7, 2003 Steve Barwick, UCI
Mod 1.1 Jul 23, 2003
Schedule for Ice Attenuation Studies at Radio Wavelengths
Date 
Task 

JunJul 03 
Acquire Electronics for Attenuation Studies (UCI+UK+UCLA) (red= Have in house)1. TRX antenna  RICE dipole 2. Modify TRX for high frequency operation (~500 MHz) UK 3. 25W Power Amp (500MHz) Supplied by UCLA 4. Signal Generator Wavetek 2405 and HP8657Bmod3 (2 GHz) 5. Digital Pulse generator (HP) Supplied by UCLA 6. GPIB clock: Truetime 7. Cable: 7/8" Andrews Cable w/N connectors 
We have all necessary equipment in house and required modifications are completed. Testing scheduled for 7/28/03 in SPRESO Lab in Albuquerque. 
Aug 03 
Drive TRX equipment to SPRESO Lab in Albuquerque (UCI+UK) Deliver cable and winch to Port Hueneme, CA (UCI) 
Jul 28, 2003 Aug 15, 2003 
SepOct 03 
1. Write software reader for RICE data format (Dave B, UK) 2. Incorporate FFT analysis into realtime analysis 3. Incorporate correlation matrix analysis if time. 

Nov 03 
Develop SemiReal Time Analysis tools for use at Pole (Dave B., UK)  
Approx Dec 15 '03 
1. Begin first series of tests of TRX and RCV equipment at pole 2. Unbury open RICE hole and install TRX 3. Collect data and check linearity of power supplies. 4. Move TRX equipment to SPRESO hole as early as possible (Dave B, Steve B). 
Planned Tests
1. Power measurements at distance of 100m, both in situ TRX and surface TRX. Adjust signal strength via TRX power adjustment and RF attenuators prior to DAQ input.
2. Perform tests on one or perhaps two specially designed RICE dipole antenna: inice resonance frequencies approximately 200 MHz and 500 MHz.
3. Should we start to think about Birefringence tests? We know the polarization from a dipole antenna, and we can use a surface dipole to look for birefringence components. But this would ideally be done with a transmitter in ice with two polarizations, which makes this a difficult measurement for now. Perhaps we can propose to OPP SGER to study these effects.
Details
RICE DAQ
Oscilloscope readout  1 GSa/s (500 MHz BW), 16us depth at 1 Gsa/s
Data Recording  1 Hz
TRX
25W amp, 50W amp (assembled by UCLA)
7/8" Andrew Cable: Information can be found at http://www.andrew.com/catalog38/cat38viewer.aspx?PageNum=509.1
Antenna: Copper cylinder, 36.5 cm length, OD=4.8 cm, ID= 4.1 cm
VSWR plot vs frequency for RICE antenna. (in ice frequency should be divided by 1.8, the index of refraction at the relevant frequency)
.
PHYSICS CONSIDERATIONS
Order of magnitude estimates by Dave Saltzberg and Dave Besson show that we need 25W of power to receive a signal from a dipole antenna at a distance of 7km, and we should get some additional safety margin from the following considerations: (1) spare 50W amp if 25 W not sufficient, (2) 7/8" Andrews coax, (3) FFT of 1000 cycle pulse train.
Ratio of power between near hole (assumed to be a distance of 100m), and the far hole at SPRESO site (assumed distance is 7km), with 1km attenuation length and 1/r^2 fall reduction in power vs distance, yields a ratio of ~10^{6}. We plan to adjust relative gains by reducing power to antenna by a known amount (and assume radiated antenna power is strictly proportional to input power from the amplifier) and attenuation of the signals from the RICE receivers. If necessary, we can reduce the gain of the downhole amps, but then we are not running the RICE RCVs in a way that is calibrated.
Reflections off surface from a TRX buried at 300m could be a problem due to variable index of refraction, but again, timing may help to pull out surface reflection by looking for the final 1000 cycle pattern in waveform.
To minimize confusion from reflected signals, we plan to send a pulse train of about 5us in length. We do not anticipate that reflections off the bottom will pose much of a concern. The warmer ice near the bottom has much lower attenuation lengths; the distance to the bottom and back is 6km more ice creating further attenuation; and the pulse train will arrive much delayed relative to the direct signals. The reflections off the lower index ice near the surface may pose additional problems, but we can probably identify the direct pulse by looking for the last few cycles, and count forward. We would then use cross correlation techniques to help deconvolve the faster reflected signal from the direct signal. However, we have not estimated the power than will be reflected in this way.
Back of envelope calculations for RF power:(modified from email by D. Saltzberg)
Assume that the Rice system has a system temperature of about 400K and bandwidth
500 MHz. That means that rms noise into 50 Ohms has a system temperature of about 400K and bandwidth 500 MHz. Then the rms noise into 50 Ohms is
0.45 nV/sqrt(Hz) *sqrt(400K/290K) * sqrt(4e8 Hz) = 11 uV
For 6 sigma then the voltage received has to be 66 uV in the 50 Ohm load. (Averaging would lower this a lot of course). With a 1000 cycle wavetrain, we would get an improvement by a factor of 10 (or 7 uV).
This is equivalent to a threshold detectable power of V^2/R= 8.7e11 W, which accounting for antenna and transformer (balun) efficiency of 90% is 9.6e11 W into the antenna.
The Friis formula for antennaantenna transmission without attenuation is
P_r = P_t *A_r *A_t/(r^2 lambda^2)
where A, the effective area for a halfwave, dipole is 0.13 lambda^2 so
P_r = P_t * 0.13^2 lambda^2/r^2
200 MHz in the ice corresponds to a wavelength of lambda~(3e8/2)*(1/2e8) =~ 0.75 meters. r=7000 m
So P_t threshold is :
P_r*r^2/(0.13^2 * lambda^2)=
Assume that there is 6 (10) dB attenuation in 350m of cable at 200 (500) MHz, then only 2 (5) watts is necessary. The number of efoldings in the Efield attenuation then that you can tolerate is 0.5*ln(25W/2 W)= 1.3 (0.8)
(Where the first factor of 0.5 accounts for power being the square of Efield.)
So we are okay if the attenuation length is ~4 km, but not if ice is closer to 12km which is what Dave Seckel has been claiming. We can of course do much better by averaging, say an extra 10 sigma in voltage which is 100* in power. So this is what we have to do. We also are planning to bring a 50W amplifier down just in case it is needed. And we are purchasing the low loss Andrew 7/8" diameter cable with only half the attenuation of the 0.5" cable