A Low-Cost Phase Cal Monitor and RFI Spectrum

A Low-Cost Phase Cal Monitor and RFI Spectrum Analyzer for VLBI 2010 Mark -5 / Mark-6 using Cheap, COTS Software Defined Radio (SDR) Hardware & Software Gleaned from the World of Amateur Radio Tom Clark NASA Goddard/NVI k 3 io@verizon. net Socorro NM Nov 1, 2011

In the Mark-3 world we monitor Phase Cal Signals in the final baseband (“video”) IF THE “OLD” WAY IF Amp LNA PCAL THE “NEW” WAY with $600 SDR and a small Netbook PC LO=8080 f=325. 99 10 k. Hz USB Only 10 k. Hz BPF & Scope HP 3528 FFT Anal 10, 1010, 2010…. k. Hz USB or 990, 1990, 2990…. k. Hz LSB with 50+ d. B Dynamic Range

In this case, we look at the Phase Cal signal at 8080 + 325. 99 – 5. 990(LSB) = 8400. 000 In this case, we see a span of ± 600 Hz around the Phase Cal “rail”. The Resolution Bandwidth (RBW) is 0. 85 Hz and the screen is showing a 40 d. B amplitude range.

In the new Mark-5 Digital Backend, all the analog Video Converter functions become Digital. And the equivalent of the Mark 3/4 VC’s USB/LSB BNC jacks exists only inside a Xilinx FPGA. LNA PCAL Microwave Fiber UDC = Up. Down Converter = First LO RDBE = Digita l Backend = 2 nd LO & Digital Baseband Filters & Formatter Mark-5 Disc Recorder

In the VLBI 2010 configuration, the RDBE (a. k. a. ROACH Board) performs all the (formerly analog) IF Back-end functions: 2 -Bit Samples for each 32 MHz PFB channel 0. 5 -2. 5 GHz IF IN Sampler & A/D Conv “”Formatter” ” Disk Array

Russ Mc. Whirter has recently provided me with an RDBE with evaluation v 1. 4 firmware 2 -Bit Samples for each 32 MHz PFB channel 0. 5 -2. 5 GHz IF IN Sampler & A/D Conv Added Firmware Sign-bit Monitoring Port “”Formatter” ” Isolation & Multiplexer Disk Array Baseband One-Bit Monitor Port

Using One-bit (sign) Samples

The Attributes desired for the Monitor SDR (in order of desirability) 1. Oscillators (RF Conversion and sampling) in the SDR should be “clean” to verify coherence. 2. Oscillators should be derived from Maser master clock. 3. If you switch between 2 frequencies A>B does it return to >A at the same phase? 4. Can the timing be synchronized with the house 1 PPS reference?

Some Commercial SDRs RF Space (http: //www. rfspace. com) in Atlanta GA makes several SDR’s, in particular: The $500 SDR-IQ (used in the previous example) covers the DC-30 MHz range with up to 192 k. Hz bandwidth. Interface is USB. This is competent small, cheap SDR that is a very useful piece of test equipment that fails in all 4 criteria The new $3700 SDR-IP looks like a perfect VLBI Phase Cal monitor. It’s internal clocks can be locked to the Maser for fully coherent system monitoring. It interfaces via Ethernet and UDP packet covering up to 2 MHz bandwidth in the 0 -34 MHz frequency range. I hope to be able to try one soon. The SDR-IP appears to meet all 4 criteria.

Some Commercial SDRs The $800 “Quicksilver” QS 1 R (http: //qs 1 r. wikispaces. com or http: //www. srl-llc. com/ or http: //groups. yahoo. com/group/qs 1 r/) by Software Radio Laboratory in Columbus OH shows much promise. It covers DC-62 MHz (or up to ~500 MHz when oversampled) with up to 2 MHz bandwidth. The QS 1 R interfaces to its PC by USB. The entire design and all its support software is “open” licensed. This unit meets Criteria 1, 2, & 3 and possibly can be modified to also meet 4.

Some Commercial SDRs The Italian Pegasus SDR is available for $1000 in the US (http: //www. universal-radio. com/catalog/commrxvr/0122. html) is quite similar to the QS 1 R but only covers 40 MHz. It is essentially a clone of the QS 1 R

Commercial SDR that may be relevant GNU Radio (http: //gnuradio. org/redmine/wiki/gnuradio) represents a major collaborative effort from a number of sources. All the GNU software is supported on the USRP hardware available from Matt Ettus (http: //www. ettus. com/). FYI – Ettus was recently acquired by National Instruments and I anticipate NI will be making a splash in the SDR world soon. See the gnuradio & ettus web sites for more details.

The Final Bottom Line 2 weeks ago I did a crude, partial test of Russ’s v 1. 4 firmware on a QS-1 R and an SDRIQ. Given the known limitations on the test setup, everything worked precisely as planned. Not tested were end-to-end coherence and the 32 -was multiplexer switch. Not implemented yet is a simple thru-path to operate the multiplexer switch.