I have been attempting to communicate via SPI with an HMC987LP5E and an HMC1034LP6GE, using an FPGA. So far I have just been trying to read the version register of each, and nothing comes back from the device. The MISO (SDO) pin just stays high all the time, no matter what I try to do.
The HMC987 seems as though it should be an especially simple project, since you don't have to worry about getting it to select the right SPI operation mode (it only has one). I have set up my FPGA so that I can adjust when I strobe SEN with respect to when I send the data, and have tried a variety of different alignments. Still no response. I have verified that SEN stays low for at least 10 ns before I assert it. PMODE-SEL is grounded on the board, and CEN is pulled high.
So I have a couple of questions. 1) Is the SPI clock alone sufficient to access the registers, or is it essential that these devices be provided with their other clocks before SPI access will work? E.g. does the HMC987 need something driving into CLKP and CLKN? Does the HMC1034 need a clock supplied to XREFP? 2) The HMC987 datasheet mentions that a pull-down resistor on the SDO line is recommended for best results when reading. How essential is this?
Any other thoughts on common gotchas that might match these symptoms are, of course, appreciated.